Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot,...

86
MYSTERIES M IND M IND OF THE OF THE Mind-Body Connections Happiness Depression Dreams Consciousness Memory Violence NEW AND UPDATED EXPLORATIONS OF HOW WE THINK, HOW WE BEHAVE AND WHAT WE FEEL SPECIAL ISSUE SPECIAL ISSUE $4.95 MYSTERIES Copyright 1997 Scientific American, Inc.

Transcript of Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot,...

Page 1: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

MYSTERIES MINDMIND

OF THEOF THE

Mind-BodyConnections

Happiness

Depression

Dreams

Consciousness

Memory

Violence

NEW AND UPDATED

EXPLORATIONS OF

HOW WE THINK,

HOW WE BEHAVE

AND WHAT WE FEEL

SPECIALISSUE

SPECIALISSUE

$4.95

MYSTERIES

Copyright 1997 Scientific American, Inc.

Page 2: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Mysteries of the Mind 3

Master detective Hercule Poirot,the hero of many an AgathaChristie novel, boasted re-

peatedly about the power of “the little graycells” in his head to solve the toughest mys-teries. For philosophers, writers and otherthinkers, however, those little gray cells havebeen the greatest mystery of all. How do acouple of pounds of spongy, electrically ac-tive tissue give rise to a psychological essence?How do we emerge from the neural thicket?

Empirical scientists may be relative new-comers to this investigation (unlike the

philosophers, they’ve been on the case foronly a few hundred years), but they havetaken long strides forward in that short time.In this special issue ofScientific American,some of the lead-

The Persistent Mystery of Our Selves

®

Established 1845

F R O M T H E E D I T O R S

Scientific American Mysteries of theMind is published by the staff of Scien-

tific American, with project manage-ment by:

John Rennie, EDITOR IN CHIEF

Michelle Press, MANAGING EDITOR

David Pope, PROJECT EDITOR

W. Wayt Gibbs, Kristin Leutwyler, STAFF WRITERS

ArtEdward Bell, ART DIRECTOR

Lisa Burnett, PRODUCTION EDITOR

Bridget Gerety, PHOTOGRAPHY EDITOR

CopyMaria-Christina Keller, COPY CHIEF

Molly K. Frances; Daniel C. Schlenoff; Terrance Dolan; Katherine Wong

AdministrationRob Gaines, EDITORIAL ADMINISTRATOR

Sonja Rosenzweig

ProductionRichard Sasso, ASSOCIATE PUBLISHER/

VICE PRESIDENT, PRODUCTION

William Sherman, DIRECTOR, PRODUCTION

Carol Albert, PRINT PRODUCTION MANAGER

Janet Cermak, MANUFACTURING MANAGER

Tanya DeSilva, PREPRESS MANAGER

Silvia Di Placido, QUALITY CONTROL MANAGER

Carol Hansen, COMPOSITION MANAGER

Madelyn Keyes, SYSTEMS MANAGER

Carl Cherebin, AD TRAFFIC; Norma Jones

CirculationLorraine Leib Terlecki, ASSOCIATE PUBLISHER/

CIRCULATION DIRECTOR

Katherine Robold, CIRCULATION MANAGER

Joanne Guralnick, CIRCULATION PROMOTION MANAGER

Rosa Davis, FULFILLMENT MANAGER

AdvertisingKate Dobson, ASSOCIATE PUBLISHER/ADVERTISING DIRECTOR

OFFICES: NEW YORK:Meryle Lowenthal, NEW YORK ADVERTISING MANAGER

Randy James; Thomas Potratz, Elizabeth Ryan; Timothy Whiting.

CHICAGO: 333 N. Michigan Ave., Suite 912, Chicago, IL 60601; Patrick Bachler, CHICAGO MANAGER

DETROIT: 3000 Town Center, Suite 1435, Southfield, MI 48075; Edward A. Bartley, DETROIT MANAGER

WEST COAST: 1554 S. Sepulveda Blvd., Suite 212, Los Angeles, CA 90025;

Lisa K. Carden, WEST COAST MANAGER; Tonia Wendt. 225 Bush St., Suite 1453,

San Francisco, CA 94104; Debra Silver. CANADA: Fenn Company, Inc. DALLAS: Griffith Group

Business AdministrationJoachim P. Rosler, PUBLISHER

Marie M. Beaumonte, GENERAL MANAGER

Alyson M. Lane, BUSINESS MANAGER

Constance Holmes, MANAGER, ADVERTISING ACCOUNTING

AND COORDINATION

Chairman and Chief Executive OfficerJohn J. Hanley

Corporate OfficersRobert L. Biewen, Frances Newburg, John J. Moeling, Jr.,

Joachim P. Rosler, VICE PRESIDENTS

Anthony C. Degutis, CHIEF FINANCIAL OFFICER

Program Development Electronic PublishingLinnéa C. Elliott, DIRECTOR Martin O. K. Paul, DIRECTOR

Scientific American, Inc.

415 Madison Avenue • New York, NY 10017-1111(212) 754-0550

PRINTED IN U.S.A.

To be consciousthat we are

perceiving orthinking is to beconscious of our

own existence. —Aristotle

I have a prodigiousquantity of mind; ittakes me as much as a week sometimes tomake it up.

—Mark Twain

ing researchersin neuroscience

and in psychologydiscuss how much isnow known about

the nature of consciousness, memory, emo-tions, creativity, dreams and other mentalphenomena. Their answers suggest thatsome of these mysteries may be largelysolved within our lifetimes—even if new onesare posed in the process.

But treat these articles as you would anygood detective story: don’t turn right to theend for the answers. Half the fun is in tracingthe deductions.

ROB

ERTO

OST

I

The scene of the crime

Memory is thecabinet of

imagination, thetreasury of reason,

the registry of conscience, and

the council chamberof thought. —St. Basil

The whole machinery of our intelligence, our general ideas andlaws, fixed and external objects, principles, persons,and gods, are somany symbolic, algebraic expressions. —George Santayana

JOHN RENNIE, Editor in [email protected]

Copyright 1997 Scientific American, Inc.

Page 3: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Stress makes the body more vulner-

able to some physical illnesses; im-

mune responses can contribute to

depression and fatigue. Even though

the brain and the immune system

differ in their functions and organi-

zation, they are interlinked at a sub-

tle biochemical level. This fact sug-

gests that drugs traditionally used to

treat neurological problems might

help against inflammatory maladies,

and vice versa.

4

8The Mind-Body

Interaction in DiseaseEsther M. Sternberg and Philip W. Gold

18

The Problemof Consciousness

Francis Crick and Christof Koch

Neuroscientists are on the trail ofhow the physical brain gives rise tothe psychological experience ofmind. The key may be synchronousfiring among sets of related neurons,generating coherence and meaningout of brain activity.

68

Emotion, Memory and the Brain

Joseph E. LeDoux

Emotional memories—such as thestrongly felt associations behindphobias—form in a way that by-passes the brain’s higher centers.This route ensures faster responseswhen danger looms.

76

The Neurobiology of Fear

Ned H. Kalin

Clues to excessive human anxi-ety can be found by studyingfear in monkeys and other spe-cies. Their example may lead tothe development of better thera-pies for frightened people.

30

The Puzzle of Conscious Experience

David J. Chalmers

Might consciousness be an irre-ducible feature in nature, as ba-sic as mass or electrical charge?Making that radical assumption,this philosopher claims, mightbe the only way for science tomake sense of the subjective ex-perience of self.

Scrutinizing the self Patterns of excitement

An experience not soon forgotten Alarm in the rhesus monkeyIntegrated organs: the brain and the immune system

Mysteries of the Mind

Copyright 1997 Scientific American, Inc.

Page 4: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Scientific American Mysteries of the Mind (ISSN 1048-0943), Special Issue Volume 7, Number 1, 1997, pub-lished by Scientific American, Inc., 415 Madison Avenue,New York, N.Y. 10017-1111. Copyright © 1997 by Scien-tific American, Inc. All rights reserved. No part of this is-sue may be reproduced by any mechanical, photo-graphic or electronic process, or in the form of aphonographic recording, nor may it be stored in a re-trieval system, transmitted or otherwise copied for pub-lic or private use without written permission of the pub-lisher. To purchase additional quantities: 1 to 9 copies:U.S. $4.95 each plus $1.00 per copy for postage andhandling (outside U.S. 3.00 P & H); 10 to 49 copies: $4.45each, postpaid; 50 copies or more: $3.95 each, postpaid.Send payment to Scientific American, Dept. MM, 415Madison Avenue, New York, N.Y. 10017-1111. CanadianBN No. 127387652RT; QST No. Q1015332537.

Cover image by Matt Mahurin.

5

84

Phantom LimbsRonald Melzack

People who have lost anarm or leg sometimes still“feel” the missing part.Neurobiologists are be-ginning to understandmore fully what createsthis disturbing illusion.

92

AutismUta Frith

Autistic individuals seem lostin their own inner world.Their isolation stems frombiological abnormalities thatmay in part interfere withthe ability to imagine otherpeople’s mental states.

102

Seeking the Criminal Element

W. Wayt Gibbs, staff writer

Identifying people withviolent tendencies mightbe a great way to preventcrime. Or it could causestill greater injustice.

40

The Pursuit of HappinessDavid G. Myers and Ed Diener

Psychology has historical-ly dwelled on the gloom-ier side of the humancondition, but now joy isstarting to get its shareof attention. Surprisingly,people are more cheerfulthan one might suppose.

44

Manic-Depressive Illness and Creativity

Kay Redfield Jamison

Bouts of depression andmanic energy are unusu-ally common among gift-ed artists, musicians andwriters. The painful rollercoaster of their emotionsmay deepen their creativeappreciation of the ambi-guities of everyday life.

53

Depression’s Double Standard

Kristin Leutwyler, staff writer

Around the world, the ratesof depression are twice ashigh among women asamong men. The reason isunclear, but biological dif-ferences between the sexesmay contribute to this psy-chological gender gap.

58

The Meaning of Dreams

Jonathan Winson

Strangely meaningful imagesand bizarre flights of fancymay all be part of the dream-ing brain’s efforts to reviewmemories, evaluate recent ex-periences and plot new strat-egies for surviving challengesin the waking world.

Ghosts that feel real Anguish of the autistic child The violent mind

How do you feel today? The moods of genius Map of female sadness Jacob’s dream

SPECIAL ISSUE/1997

46% 27% 4%

2% 1% 0%

Copyright 1997 Scientific American, Inc.

Page 5: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

The Mind-Body Interaction in Disease8 Mysteries of the Mind

The Authors

ESTHER M. STERNBERGand PHILIP W. GOLD carryout their research on stressand immune systems at theNational Institute of MentalHealth, where Sternberg ischief of the section on neu-roendocrinology and behaviorand Gold is chief of the clini-cal neuroendocrinologybranch. Sternberg received herM.D. from McGill University.Her work on the mechanismsand molecular basis of neu-roimmune communication hasled to a growing recognitionof the importance of the mind-body interaction. She also isan authority on the L-trypto-phan eosinophilia myalgiasyndrome, which reached al-most epidemic proportions in1989. Prior to joining theNIMH in 1974, Gold receivedhis medical training at DukeUniversity and Harvard Uni-versity. Gold and his groupwere among the first to intro-duce data implicating cortico-tropin-releasing hormone andits related hormones in thepathophysiology of melan-cholic and atypical depressionand in the mechanisms of ac-tion of antidepressant drugs.

The Mind-Body Interaction in DiseaseThe brain and immune system continuously signaleach other, often along the same pathways, whichmay explain how state of mind influences health

by Esther M. Sternberg and Philip W. Gold

The belief that the mind plays an important role in physical illness goes

back to the earliest days of medicine. From the time of the ancient Greeks

to the beginning of the 20th century, it was generally accepted by both

physician and patient that the mind can affect the course of illness, and it

seemed natural to apply this concept in medical treatments of disease. After the dis-

covery of antibiotics, a new assumption arose that treatment of infectious or inflam-

matory disease requires only the elimination of the foreign organism or agent that

triggers the illness. In the rush to discover new antibiotics and drugs that cure specific

infections and diseases, the fact that the body’s own responses can influence suscepti-

bility to disease and its course was largely ignored by medical researchers.

It is ironic that research into infectious and inflammatory disease first led 20th-cen-

tury medicine to reject the idea that the mind influences physical illness, and now re-

search in the same field—including the work of our laboratory and of our collabora-

tors at the National Institutes of Health—is proving the contrary. New molecular and

pharmacological tools have made it possible for us to identify the intricate network

that exists between the immune system and the brain, a network that allows the two

systems to signal each other continuously and rapidly. Chemicals produced by im-

mune cells signal the brain, and the brain in turn sends chemical signals to restrain the

immune system. These same chemical signals also affect behavior and the response to

stress. Disruption of this communication network in any way, whether inherited or

through drugs, toxic substances or surgery, exacerbates the diseases that these systems

guard against: infectious, inflammatory, autoimmune and associated mood disorders.

The clinical significance of these findings is likely to prove profound. They hold the

promise of extending the range of therapeutic treatments available for various dis-

orders, as drugs previously known to work primarily for nervous system problems

are shown to be effective against immune maladies, and vice versa. They also help

to substantiate the popularly held impression (still discounted in some medical cir-

cles) that our state of mind can influence how well we resist or recover from infec-

tious or inflammatory diseases.

The brain’s stress response system is activated in threatening situations. The im-

mune system responds automatically to pathogens and foreign molecules. These two

response systems are the body’s principal means for maintaining an internal steady

state called homeostasis. A substantial proportion of human cellular machinery is

dedicated to maintaining it.

Immune response can be altered at the cellular level

by stress hormones.

Copyright 1997 Scientific American, Inc.

Page 6: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

The Mind-Body Interaction in Disease Mysteries of the Mind 9

ROB

ERTO

OST

I

STRESS RESPONSENerves connect the brain to ev-

ery organ and tissue. Challengingor threatening situations arouse

the brain’s stress response, whichinvolves the release of a hor-

mone that stimulates physiologi-cal arousal and regulates the im-mune system. Key components

in this stress response are the hy-pothalamus and locus ceruleusin the brain, the pituitary gland,

the sympathetic nervous systemand the adrenal glands.

IMMUNE RESPONSEThe immune system operates as

a decentralized network, re-sponding automatically to any-

thing that invades or disrupts thebody. Immune cells generated inthe bone marrow, lymph nodes,

spleen and thymus communi-cate with one another using

small proteins. These chemicalmessengers can also send signals

to the brain, through either thebloodstream or nerve pathwayssuch as the vagus nerve and nu-

cleus of the tractus solitarius.

HYPOTHALAMUS

PITUITARY GLAND

LOCUS CERULEUS

BRAIN STEM

LYMPH NODE

SPLEEN

ADRENAL GLAND

THYMUS

VAGUS NERVE

BONE MARROW

LIVER

NUCLEUS OF TRACTUSSOLITARIUS

KIDNEY

Anatomy of the Stress and Immune Systems

Copyright 1997 Scientific American, Inc.

Page 7: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

When homeostasis isdisturbed or threatened,a repertoire of molecu-lar, cellular and behav-ioral responses comesinto play. These respons-es attempt to counteractthe disturbing forces inorder to reestablish asteady state. They canbe specific to the foreigninvader or a particularstress, or they can begeneralized and non-specific when the threatto homeostasis exceedsa certain threshold. Theadaptive responses maythemselves turn intostressors capable of pro-ducing disease. We arejust beginning to under-stand the many ways inwhich the brain and theimmune system are in-terdependent, how theyhelp to regulate andcounterregulate each oth-er and how they themselves can mal-function and produce disease.

The stress response promotes physio-logical and behavioral changes that en-hance survival in threatening or taxingsituations. For instance, when we arefacing a potentially life-threatening sit-uation, the brain’s stress response goesinto action to enhance our focused at-tention, our fear and our fight-or-flightresponse, while simultaneously inhibit-ing behaviors, such as feeding, sex andsleep, that might lessen the chance ofimmediate survival. The stress re-sponse, however, must be regulated tobe neither excessive nor suboptimal;otherwise, disorders of arousal, thoughtand feeling emerge.

The immune system’s job is to barforeign pathogens from the body andto recognize and destroy those thatpenetrate its shield. The immune sys-tem also must neutralize potentiallydangerous toxins, facilitate repair ofdamaged or worn tissues, and disposeof abnormal cells. Its responses are sopowerful that they require constant reg-ulation to ensure that they are neitherexcessive nor indiscriminate and yet re-main effective. When the immune sys-tem escapes regulation, autoimmuneand inflammatory diseases or immunedeficiency syndromes result.

The immune and central nervous sys-tems appear, at first glance, to be orga-

nized in very differentways. The brain is usual-ly regarded as a central-ized command center,sending and receivingelectrical signals alongfixed pathways, muchlike a telephone net-work. In contrast, the im-mune system is decen-tralized, and its organs(spleen, lymph nodes, thy-mus and bone marrow)are located throughoutthe body. The classicalview is that the immunesystem communicates byreleasing immune cells in-to the bloodstream thatfloat, like boats, to newlocations to deliver theirmessages or to performother functions.

The central nervousand immune systems,however, are in factmore similar than differ-ent in their modes of re-

ceiving, recognizing and integrating sig-nals from the external environment andin their structural design for accom-plishing these tasks. Both the centralnervous system and the immune systempossess “sensory” elements, which re-ceive information from the environ-ment and other parts of the body, and“motor” elements, which carry out anappropriate response.

Cross Communication

Both systems also rely on chemicalmediators for communication. Elec-

trical signals along nerve pathways, forinstance, are converted to chemical sig-nals at the synapses between neurons.The chemical messengers produced byimmune cells communicate not onlywith other parts of the immune systembut also with the brain and nerves, andchemicals released by nerve cells can actas signals to immune cells. Hormonesfrom the body travel to the brain in thebloodstream, and the brain itself makeshormones. Indeed, the brain is perhapsthe most prolific endocrine organ in thebody and produces many hormonesthat act both on the brain and on tis-sues throughout the body.

A key hormone shared by the centralnervous and immune systems is cortico-tropin-releasing hormone (CRH); pro-duced in the hypothalamus and several

other brain regions, it unites the stressand immune responses. The hypothala-mus releases CRH into a specializedbloodstream circuit that conveys thehormone to the pituitary gland, whichis just beneath the brain. CRH causesthe pituitary to release adrenocortico-tropin hormone (ACTH) into thebloodstream, which in turn stimulatesthe adrenal glands to produce cortisol,the best-known hormone of the stressresponse.

Cortisol is a steroid hormone that in-creases the rate and strength of heartcontractions, sensitizes blood vessels tothe actions of norepinephrine (anadrenalinelike hormone) and affectsmany metabolic functions—actions thathelp to prepare the body to meet astressful situation. In addition, cortisolis a potent immunoregulator and anti-inflammatory agent. It plays a crucialrole in preventing the immune systemfrom overreacting to injuries and dam-aging tissues. Furthermore, cortisol in-hibits the release of CRH by the hypo-thalamus—a simple feedback loop thatkeeps this component of the stress re-sponse under control. Thus, CRH andcortisol directly link the body’s brain-regulated stress response and its im-mune response.

CRH-secreting neurons of the hypo-thalamus send fibers to regions in thebrain stem that help to regulate thesympathetic nervous system, as well asto another brain stem area called the lo-cus ceruleus. The sympathetic nervoussystem, which mobilizes the body dur-ing stress, also innervates immune or-gans, such as the thymus, lymph nodesand spleen, and helps to control inflam-matory responses throughout the body.Stimulation of the locus ceruleus leadsto behavioral arousal, fear and en-hanced vigilance.

Perhaps even more important for theinduction of fear-related behaviors isthe amygdala, where inputs from thesensory regions of the brain are chargedas stressful or not. CRH-secreting neu-rons in the central nucleus of the amyg-dala send fibers to the hypothalamusand the locus ceruleus, as well as toother parts of the brain stem. TheseCRH-secreting neurons are targets ofmessengers released by immune cellsduring an immune response. By recruit-ing the CRH-secreting neurons, the im-mune signals not only activate cortisol-mediated restraint of the immune re-sponse but also induce behaviors thatassist in recovery from illness or injury.

The Mind-Body Interaction in Disease10 Mysteries of the Mind

ROB

ERTO

OST

I

When we are fac-

ing a potentially life-

threatening situa-

tion, the brain’s

stress response goes

into action to en-

hance our focused

attention, our fear

and our fight-or-

flight arousal, while

inhibiting feeding,

sex and sleep.

Copyright 1997 Scientific American, Inc.

Page 8: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

CRH-secreting neurons also have con-nections with hypothalamic regionsthat regulate food intake and reproduc-tive behavior. In addition, there are oth-er hormonal and nerve systems, such asthe thyroid, growth and female sex hor-mones, and the sympathomedullarypathways, that influence brain–immunesystem interactions.

The Immune System’s Signals

The immune response is an elegantand finely tuned cascade of cellular

events aimed at ridding the body of for-eign substances, bacteria and viruses.

One of the major discoveries of con-temporary immunology is that whiteblood cells produce small proteins thatindirectly coordinate the responses ofother parts of the immune system topathogens. For example, the protein in-terleukin-1 (IL-1) is made by a type ofwhite blood cell called a monocyte ormacrophage. IL-1 stimulates anothertype of white blood cell, the lympho-cyte, to produce interleukin-2 (IL-2),which in turn induces lymphocytes todevelop into mature immune cells.Some mature lymphocytes, called plas-ma cells, make antibodies that fight in-fection, whereas others, the cytotoxiclymphocytes, kill viruses directly. Otherinterleukins mediate the activation ofimmune cells that are involved in aller-gic reactions.

The interleukins were originallynamed to reflect what was consideredto be their primary function: communi-cation between (“inter-”) the whiteblood cells (“leukins”). But it is nowknown that interleukins also act aschemical signals between immune cellsand many other types of cells and or-gans, including parts of the brain, andso a new name—“cytokine”—has beencoined. Cytokines are biological mole-cules that cells use to communicate.Each cytokine is a distinct protein mol-ecule, encoded by a separate gene, thattargets a particular cell type. A cytokinecan either stimulate or inhibit a re-sponse depending on the presence ofother cytokines or other stimuli and thecurrent state of metabolic activity. Thisflexibility allows the immune system totake the most appropriate actions tostabilize the local cellular environmentand to maintain homeostasis.

Cytokines from the body’s immunesystem can send signals to the brain inseveral ways. Ordinarily, a “blood-brain barrier” shields the central ner-

vous system from potentially danger-ous molecules in the bloodstream. Dur-ing inflammation or illness, however,this barrier becomes more permeable,and cytokines may be carried acrossinto the brain with nutrients from theblood. Certain cytokines, on the otherhand, readily pass through at any time.But cytokines do not have to cross theblood-brain barrier to exert their ef-fects. Cytokines made in the lining ofblood vessels in the brain can stimulatethe release of secondary chemical sig-nals in the brain tissue around theblood vessels.

Cytokines can also signal the brainvia direct nerve routes, such as the va-gus nerve, which innervates the heart,stomach, small intestine and other or-gans of the abdominal cavity. Injectionof IL-1 into the abdominal cavity acti-vates the nucleus of the tractus solitar-ius, the principal region of the brainstem for receipt of visceral sensory sig-nals. Cutting the vagus nerve blocks ac-

tivation of the tractus nucleus by IL-1.Sending signals along nerve routes isthe most rapid mechanism—on the or-der of milliseconds—by which cyto-kines signal the brain.

Activation of the brain by cytokinesfrom the peripheral parts of the bodyinduces behaviors of the stress re-sponse, such as anxiety and cautiousavoidance, that keep the affected indi-vidual out of harm’s way until full heal-ing occurs. Anyone who has experi-enced lethargy and excess sleepinessduring an illness will recognize this setof characteristic responses as “sicknessbehavior.”

Neurons and nonneuronal brain cellsalso produce cytokines of their own.Cytokines in the brain regulate nervecell growth and death, and they alsocan be recruited by the immune systemto stimulate the release of CRH. TheIL-1 cytokine system in the brain is cur-rently the best understood—all its com-ponents have been identified, including

The Mind-Body Interaction in Disease Mysteries of the Mind 11

HPA AXIS is a central component of the brain’s neuroendocrine response to stress. Thehypothalamus, when stimulated, secretes corticotropin-releasing hormone (CRH) intothe hypophyseal portal system, which supplies blood to the anterior pituitary. CRHstimulates the pituitary (red arrows show stimulatory pathways) to secrete adrenocor-ticotropin hormone (ACTH) into the bloodstream. ACTH causes the adrenal glands torelease cortisol, the classic stress hormone that arouses the body to meet a challengingsituation. But cortisol then modulates the stress response (blue arrows indicate in-hibitory effects) by acting on the hypothalamus to inhibit the continued release ofCRH. Also a potent immunoregulator, cortisol acts on many parts of the immune sys-tem to prevent it from overreacting and harming healthy cells and tissue.

ROB

ERTO

OST

I

HYPOTHALAMUS

ACTH

PITUITARYGLAND

ADRENAL GLAND

CRH

TO IMMUNESYSTEM

BRAIN STEMCORTISOL

Hypothalamus-Pituitary-Adrenal (HPA) Axis

Copyright 1997 Scientific American, Inc.

Page 9: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

receptors and a naturally occurring an-tagonist that binds to IL-1 receptorswithout activating them. The anatom-ical and cellular locations of this IL-1circuitry are being mapped out in de-tail, and this new knowledge will aidresearchers in designing drugs thatblock or enhance the actions of suchcircuits and the functions they regulate.

Excessive amounts of cytokines in thebrain can be toxic to nerves. In geneti-cally engineered mice, transplantedgenes that overexpress cytokines pro-duce neurotoxic effects. Some of theneurological symptoms of AIDS in hu-mans also may be caused by overex-pression of certain cytokines in thebrain. High levels of IL-1 and other cy-tokines have been found in the braintissue of patients living with AIDS, con-centrated in areas around the giantmacrophages that invade the patients’brain tissue.

Any disruption of com-munication between thebrain and the immune sys-tem leads to greater suscep-tibility to inflammatory dis-ease and, frequently, toincreased severity of theimmune complications. Forinstance, animals whosebrain-immune communica-tions have been disrupted(through surgery or drugs)are highly liable to lethalcomplications of inflamma-tory diseases and infectiousdiseases.

Susceptibility to inflam-matory disease that is asso-ciated with genetically im-paired stress response canbe found across species—inrats, mice, chickens and,though the evidence is lessdirect, humans. For in-stance, the Lewis strain ofrat is naturally prone tomany inflammatory diseas-es because of a severe im-pairment of its HPA axis,which greatly diminishesCRH secretion in responseto stress. In contrast, thehyperresponsive HPA-axisin the Fischer strain of ratprovides it with a strongresistance to inflammatorydisease.

Evidence of a causal linkbetween an impaired stressresponse and susceptibility

to inflammatory disease comes frompharmacological and surgical studies.Pharmacological intervention such astreatment with a drug that blocks corti-sol receptors enhances autoimmune in-flammatory disease. Injecting low dosesof cortisol into disease-susceptible ratsenhances their resistance to inflamma-tion. Strong evidence comes from surgi-cal intervention. Removal of the pitu-itary gland or the adrenal glands fromrats that normally are resistant to in-flammatory disease renders them highlysusceptible. Further proof comes fromstudies in which the transplantation ofhypothalamic tissue from disease-resis-tant rats into the brain of susceptiblerats dramatically improves their resis-tance to peripheral inflammation.

These animal studies demonstratethat disruption of the brain’s stress re-sponse enhances the body’s response to

inflammatory disease, and reconstitutionof the stress response reduces suscepti-bility to inflammation. One implicationof these findings is that disruption ofthe brain-immune communication sys-tem by inflammatory, toxic or infec-tious agents could contribute to someof the variations in the course of the im-mune system’s inflammatory response.

CRH and Depression

Although the role of the stress re-sponse in inflammatory disease in

humans is more difficult to prove, thereis a growing amount of evidence that awide variety of such diseases are associ-ated with impairment of the HPA axisand lower levels of CRH secretion,which ultimately results in a hyperac-tive immune system. Furthermore, pa-tients with a mood disorder called atyp-

ical depression also have ablunted stress response andimpaired CRH function,which leads to lethargy, fa-tigue, increased sleep andincreased feeding that oftenproduces weight gain.

Patients with other illness-es characterized by lethar-gy and fatigue, such aschronic fatigue syndrome,fibromyalgia and seasonalaffective disorder (SAD),exhibit features of both de-pression and a hyperactiveimmune system. A personwith chronic fatigue syn-drome classically manifestsdebilitating lethargy or fa-tigue lasting six months orlonger with no demonstra-ble medical cause, as well asfeverishness, aches in jointsand muscles, allergic symp-toms and higher levels ofantibodies to a variety ofviral antigens (includingEpstein-Barr virus).

Patients with fibromyal-gia suffer from muscleaches, joint pains and sleepabnormalities, symptomssimilar to early, mild rheum-atoid arthritis. Both theseillnesses are associated witha profound fatigue like thatin atypical depression. SAD,which usually occurs inwinter, is typified by lethar-gy, fatigue, increased foodintake and increased sleep.

The Mind-Body Interaction in Disease12 Mysteries of the Mind

BRAIN AND IMMUNE SYSTEM can either stimulate (red ar-rows) or inhibit (blue arrows) each other. Immune cells produce cy-tokines (chemical signals) that stimulate the hypothalamus throughthe bloodstream or via nerves elsewhere in the body. The hormoneCRH, produced in the hypothalamus, activates the HPA axis. Therelease of cortisol tunes down the immune system. CRH, acting onthe brain stem, stimulates the sympathetic nervous system, whichinnervates immune organs and regulates inflammatory responsesthroughout the body. Disruption of these communications in any wayleads to greater susceptibility to disease and immune complications.

ROB

ERTO

OST

I

Interaction of the Brain and Immune System

HYPOTHALAMUS

CORT

ISO

L

CORT

ISO

L

CY

TOKI

NES

ACTH

PITUITARYGLAND

ADRENALGLANDS

IMMUNE CELLS

IMMUNE ORGANS

VAGUSNERVE

LOCUS CERULEUS

NUCLEUS OF THETRACTUS SOLITARIUS

SYMPATHETICNERVOUS SYSTEM

CRH

Copyright 1997 Scientific American, Inc.

Page 10: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Many of its symptoms are similar tothose of atypical depression.

A deficiency of CRH could contrib-ute to lethargy in patients with chronicfatigue syndrome. Injection of CRHinto patients with fatigue syndromecauses a delayed and blunted ACTH se-cretion by the HPA axis. That same re-sponse is also seen in patients whosehypothalamus has been injured or whohave a tumor. Also, fatigue and hyper-activity of the immune response are as-sociated with cortisol deficiency, whichoccurs when CRH secretion decreases.The hormone levels and responses inpatients with fatigue syndromes sug-gest—but do not prove—that theirHPA-axis functions are impaired, re-sulting in a decrease in CRH and corti-sol secretion and an increase in immunesystem activity. Together these findingssuggest that human illness character-ized by fatigue and hyperimmunitycould possibly be treated by drugs thatmimic CRH actions in the brain.

In contrast, the classic form of de-pression, melancholia, actually is not astate of inactivation and suppression ofthought and feeling; rather it presentsas an organized state of anxiety. Theanxiety of melancholia is chiefly aboutthe self. Melancholic patients feel im-poverished and defective and often ex-press hopelessness about the prospectsfor their unworthy selves in either loveor work. The anxious hyperarousal ofmelancholic patients also manifests as apervasive sense of vulnerability, andmelancholic patients often interpret rel-atively neutral cues as harbingers ofabandonment or embarrassment.

Melancholic patients also show be-havioral alterations suggestive of physi-ological hyperarousal. They character-istically suffer from insomnia (usuallyearly-morning awakening) and experi-ence inhibition of eating, sexual activityand menstruation. One of the mostwidely found biological abnormalitiesin patients with melancholia is that ofsustained hypersecretion of cortisol.

Many studies have been conductedon patients with major depression todetermine whether the excessive level ofcortisol associated with depression cor-relates with suppressed immune re-sponses. Some have found a correlationbetween hypercortisolism and immu-nosuppression; others have not. Be-cause depression can have a variety ofmental and biochemical causes onlysome depressed patients may be im-munosuppressed.

The excessive secretion of cortisol inmelancholic patients is the result pre-dominantly of hypersecretion of CRH,caused by a defect in or above the hy-pothalamus. Thus, the clinical and bio-chemical manifestations of melancholiareflect a generalized stress response thathas escaped the usual counterregula-tion, remaining, as it were, stuck in the“on” position.

The effects of tricyclic antidepressantdrugs on components of the stress re-sponse support the concept that melan-cholia is associated with a chronicstress response. In rats, regular, but notacute, administration of the tricyclic an-tidepressant imipramine significantlylowers the levels of CRH precursors inthe hypothalamus. Imipramine givenfor two months to healthy persons withnormal cortisol levels causes a gradualand sustained decrease in CRH secre-tion and other HPA-axis functions, in-dicating that down-regulation of im-portant components of the stress re-sponse is an intrinsic effect ofimipramine.

Depression is also associated with in-flammatory disease. About 20 percentof patients with rheumatoid arthritisdevelop clinical depression at somepoint during the course of their arthrit-ic disease. A questionnaire commonlyused by clinicians to diagnose depres-sion contains about a dozen questionsthat are almost always answeredaffirmatively by patients with arthritis.

Stress and Illness

In the past, the association between aninflammatory disease and stress was

considered by doctors to be secondaryto the chronic pain and debilitation ofthe disease. The recent discovery of thecommon underpinning of the immuneand stress responses may provide an ex-planation of why a patient can be sus-ceptible to both inflammatory diseaseand depression. The hormonal dysregu-lation that underlies both inflammatorydisease and depression can lead to ei-ther illness, depending on whether theperturbing stimulus is pro-inflammato-

The Mind-Body Interaction in Disease Mysteries of the Mind 13

HYPOTHALAMIC CRH produces changes important to stress and inflammationadaptation in ways other than inducing cortisol release from the adrenal glands. Path-ways from CRH-secreting neurons in the hypothalamus extend to the locus ceruleus inthe brain stem. Separate pathways from other hypothalamic neurons to the brain steminfluence sympathetic nervous system activity, which modulates inflammatory respons-es as well as regulating metabolic and cardiovascular activities. Stimulation by CRH ofthe locus ceruleus produces protective behaviors such as arousal and fear (red indicatesstimulation, blue inhibition). The locus ceruleus, in turn, provides feedback to the hy-pothalamus for continued production of CRH and also acts on the sympathetic ner-vous system. Self-inhibitory feedback keeps the activities of CRH and the locusceruleus under control.

ROB

ERTO

OST

I

HYPOTHALAMUS

LOCUS CERULEUS–NOREPINEPHRINE SYSTEM

SYMPATHETIC NERVOUS SYSTEM

TO IMMUNE ORGANS

CRH, the Locus Ceruleus and Sympathetic Nervous System

CRH

CRH

Copyright 1997 Scientific American, Inc.

Page 11: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

ry or psychologically stressful. Thatmay explain why the waxing and wan-ing of depression in arthritic patientsdoes not always coincide with inflam-matory flare-ups.

The popular belief that stress exacer-bates inflammatory illness and that re-laxation or removal of stress amelio-rates it may indeed have a basis in fact.The interactions of the stress and im-mune systems and the hormonal re-sponses they have in common could ex-plain how conscious attempts to tonedown responsivity to stress could affectimmune responses.

How much of the responsivity tostress is genetically determined and howmuch can be consciously controlled isnot known. The set point of the stressresponse is to some extent geneticallydetermined. An event that is physiolog-ically highly stressful to one individualmay be much less so to another, de-pending on each person’s genetic ten-

dency to hormonal reactivity. The de-gree to which stress could precipitate orexacerbate inflammatory disease wouldthen depend both on the intensity ofthe stressful stimulus and on the setpoint of the stress system.

Psychological stress can affect an in-dividual’s susceptibility to infectiousdiseases. The regulation of the immunesystem by the neurohormonal stresssystem provides a biological basis forunderstanding how stress might affectthese diseases. There is evidence thatstress does affect human immune re-sponses to viruses and bacteria. In stud-ies with volunteers given a standarddose of the common cold virus (rhi-novirus), individuals who are simulta-neously exposed to stress show moreviral particles and produce more mucusthan do nonstressed individuals. Medi-cal students receiving hepatitis vaccina-tion during their final exams do not de-velop full protection against hepatitis.

These findings have impor-tant implications for thescheduling of vaccinations.People who are vaccinatedduring periods of stressmight be less likely to devel-op full antibody protection.

Animal studies providefurther evidence that stressaffects the course and severi-ty of viral illness, bacterialdisease and septic shock.Stress in mice worsens theseverity of influenza infec-tion and affects both theHPA axis and the sympa-thetic nervous system. Ani-mal studies suggest thatneuroendocrine mechanismscould play a similar role ininfections with other viruses,including HIV, and provide amechanism for understand-ing clinical observations thatstress may exacerbate thecourse of AIDS. Stress in-creases the susceptibility ofmice to infection with my-cobacteria, the bacteria thatcauses tuberculosis. It hasbeen shown that an intactHPA axis protects ratsagainst the lethal septic ef-fects of salmonella bacteria.Finally, new understandingof interactions of the im-mune and stress responsescan help explain the puzzlingobservation that classic psy-

chological conditioning of animals caninfluence their immune responses. Forexample, working with rats, RobertAder and Nicholas Cohen of the Uni-versity of Rochester paired saccharin-flavored water with an immunosup-pressive drug. Eventually the saccharinalone produced a decrease in immunefunction similar to that of the drug.

Stress is not only personal but is per-ceived through the prism of interactionswith other persons. Social interactionscan either add to or lessen psychologi-cal stress and similarly affect our hor-monal responses to it, which in turncan alter immune responses. Thus, thesocial psychological stresses that we ex-perience can affect our susceptibility toinflammatory and infectious diseasesand the course of a disease. For in-stance, studies have shown that personsexposed to chronic social stresses formore than two months have increasedsusceptibility to the common cold.

The Mind-Body Interaction in Disease14 Mysteries of the Mind

LYMPHOCYTE

Cerebral Blood VesselNEURON

NITRIC OXIDESYNTHASE

MONOCYTE

IL-1

PROSTAGLANDINSCYCLOOXYGENASE

LEAKY JUNCTIONIN BLOOD- BRAIN BARRIER

IL-1RECEPTOR

NITRICOXIDE

IMMUNE SIGNALS TO THE BRAIN via the bloodstream can occur directly or indirectly. Im-mune cells such as monocytes, a type of white blood cell, produce a chemical messenger called in-terleukin-1 (IL-1), which ordinarily will not pass through the blood-brain barrier. But certain cere-bral blood vessels contain leaky junctions, which allow IL-1 molecules to pass into the brain.There they can activate the HPA axis and other neural systems. IL-1 also binds to receptors on theendothelial cells that line cerebral blood vessels. This binding can cause enzymes in the cells toproduce nitric oxide or prostaglandins, which diffuse into the brain and act directly on neurons.

LAU

RIE

GR

AC

E

Copyright 1997 Scientific American, Inc.

Page 12: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Other studies have shown that theimmune responses of long-term care-givers, such as spouses of Alzheimer’spatients, become blunted. Immune re-sponses in unhappily married and di-vorcing couples are also blunted. Oftenthe wife has a feeling of helplessnessand experiences the greatest amount ofstress. In such a scenario, studies havefound that the levels of stress hormonesare elevated and immune responsesusually are lowered in the wife but notin the husband.

On the other hand, a positive sup-portive environment of extensive socialnetworks or group psychotherapy canenhance immune response and resis-tance to disease—even cancer. Womenwith breast cancer, for instance, whoreceive strong, positive social supportduring their illness have significantlylonger life spans than women withoutsuch support.

New Approaches to Treatment

For centuries, taking the cure at amountain sanatorium or a hot-

springs spa was the only available treat-ment for many chronic diseases. Newunderstanding of the communicationbetween the brain and immune systemprovides a physiological explanation ofwhy such cures sometimes worked.Disruption of this communication net-work leads to an increase in susceptibil-ity to disease and can worsen thecourse of the illness. Restoration ofthis communication system, whetherthrough pharmacological agents or therelaxing effects of a spa, can be the firststep on the road to recovery.

A corollary of these findings is thatpsychoactive drugs may in some casesbe used to treat inflammatory diseases,and drugs that affect the immune sys-tem may be useful in treating some psy-chiatric disorders. There is growing evi-

dence that our view of ourselves andothers, our style of handling stresses, aswell as our genetic makeup, can affectactivities of the immune system. Simi-larly, there is good evidence that diseasesassociated with chronic inflammation

significantly affect on one’s mood orlevel of anxiety. Finally, these findingssuggest that classification of illnessesinto medical and psychiatric specialties,and the boundaries that have demarcat-ed mind and body, are artificial.

The Mind-Body Interaction in Disease Mysteries of the Mind 15

Further Reading

Neural-Immune Interactions. David L. Felton and Suzanne Y. Felton in Encyclopediaof Human Biology. Academic Press, 1991.

The Concepts of Stress and Stress System Disorders. G. P. Chrousos and P. W. Goldin Journal of the American Medical Association, Vol. 267, No. 9, pages 1244–1252;March 4, 1992.

Endocrine and Immune Function. J. K. Kiecolt-Glaser, W. Malarkey, J. T. Cacioppo andR. Glaser in Handbook of Human Stress and Immunity. Edited by R. Glaser and J. K.Kiecolt-Glaser. Academic Press, 1994.

Stress: Mechanisms and Clinical Implications. G. P. Chrousos, R. McCarty, K.Pacak, G. Cizza, Esther M. Sternberg, Philip W. Gold and R. Kvetnansky in Annals of theNew York Academy of Sciences, Vol. 771, 1995.

Emotions and Disease: From Balance of Humors to Balance of Molecules. EstherM. Sternberg in Nature Medicine, Vol. 3, No. 3, pages 264–267; March 1997.

The Neurologic Basis of Fever. Clifford B. Saper and Christopher D. Breder in NewEngland Journal of Medicine, Vol. 330, No. 26, pages 1880–1886; June 30, 1997.

Emotions and Disease. Catalogue for an Exhibition at the National Library of Medicineand National Institutes of Health. Edited by Esther M. Sternberg and Elizabeth Fee.Friends of the National Library of Medicine, Bethesda Md. May 1997.

National Institutes of Health World Wide Web site for information on emotions and dis-ease: http://ohrm.od.nih.gov/ose/snapshots/

Immune Cell

DNA BINDINGREGION

HORMONE RECEPTOR

CORTISOL

HEAT-SHOCKPROTEIN

DIRECTS PROTEIN

PRODUCTION

mRNA

NUCLEUS

HORMONE-RECEPTOR

BINDING SITE

DNA

AP-1 BINDING SITE

EFFECTS ON IMMUNECELL FUNCTION

c-jun

c-fos

LAU

RIE

GR

AC

E

ALTERED GENETIC ACTIVITY in im-mune cells is an effect of cortisol. Thecortisol receptors in immune cells arefolded and bound to large “heat-shock”proteins. When cortisol enters a cell andbinds to its receptor, the protein is dis-placed and the receptor unfolds. The re-ceptor then binds to DNA in the nucleus,changing the cell’s transcription of mes-senger RNA (mRNA) and production ofproteins. (Other molecules called c-fosand c-jun bind with the receptor and con-fer more specificity on its action.) Theproteins leave the cell and directly affectcytokine and lymphocyte production.

SA

Copyright 1997 Scientific American, Inc.

Page 13: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

18 Mysteries of the Mind

The Problem of ConsciousnessIt can now be approached by scientific investigationof the visual system. The solution will require a closecollaboration among psychologists, neuroscientists and theorists

by Francis Crick and Christof Koch

Copyright 1997 Scientific American, Inc.

Page 14: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Reprinted from the September 1992 issue Mysteries of the Mind 19

The Authors

FRANCIS CRICK andCHRISTOF KOCH share aninterest in the experimentalstudy of consciousness. Crickis the co-discoverer, withJames Watson, of the doublehelical structure of DNA.While at the Medical Re-search Council Laboratory ofMolecular Biology in Cam-bridge, he worked on the ge-netic code and on develop-mental biology. Since 1976,he has been at the Salk Insti-tute for Biological Studies inSan Diego. His main interestlies in understanding the visu-al system of mammals. Kochwas awarded his Ph.D. inbiophysics by the Universityof Tübingen. After spendingfour years at the Massachu-setts Institute of Technology,he joined the California Insti-tute of Technology, where heis now professor of computa-tion and neural systems. He isstudying how single braincells process information andthe neural basis of motionperception, visual attentionand awareness. He also de-signs analog VLSI visionchips for intelligent systems.

VISUAL AWARENESS primarily involves seeing what is directly in front of you,but it can be influenced by a three-dimensional representation of the object inview retained by the brain. If you see the back of a person’s head, the brain infersthat there is a face on the front of it. We know this is true because we would bevery startled if a mirror revealed that the front was exactly like the back, as in thispainting, Reproduction Prohibited (1937), by René Magritte.

The overwhelming question in neurobiology today is the relation be-

tween the mind and the brain. Everyone agrees that what we know as

mind is closely related to certain aspects of the behavior of the brain,

not to the heart, as Aristotle thought. Its most mysterious aspect is con-

sciousness or awareness, which can take many forms, from the experience of pain to

self-consciousness. In the past the mind (or soul) was often regarded, as it was by

Descartes, as something immaterial, separate from the brain but interacting with it

in some way. A few neuroscientists, such as Sir John Eccles, still assert that the soul

is distinct from the body. But most neuroscientists now believe that all aspects of

mind, including its most puzzling attribute—consciousness or awareness—are likely

to be explainable in a more materialistic way as the behavior of large sets of inter-

acting neurons. As William James, the father of American psychology, said a centu-

ry ago, consciousness is not a thing but a process.

Exactly what the process is, however, has yet to be discovered. For many years af-

ter James penned The Principles of Psychology, consciousness was a taboo concept

in American psychology because of the dominance of the behaviorist movement.

With the advent of cognitive science in the mid-1950s, it became possible once more

for psychologists to consider mental processes as opposed to merely observing be-

havior. In spite of these changes, until recently most cognitive scientists ignored con-

sciousness, as did almost all neuroscientists. The problem was felt to be either pure-

ly “philosophical” or too elusive to study experimentally. It would not have been

easy for a neuroscientist to get a grant just to study consciousness.

In our opinion, such timidity is ridiculous, so a few years ago we began to think

about how best to attack the problem scientifically. How to explain mental events as

being caused by the firing of large sets of neurons? Although there are those who be-

lieve such an approach is hopeless, we feel it is not productive to worry too much

over aspects of the problem that cannot be solved scientifically or, more precisely,

cannot be solved solely by using existing scientific ideas. Radically new concepts

may indeed be needed—recall the modifications of scientific thinking forced on us

by quantum mechanics. The only sensible approach is to press the experimental at-

tack until we are confronted with dilemmas that call for new ways of thinking.

There are many possible approaches to the problem of consciousness. Some psy-

chologists feel that any satisfactory theory should try to explain as many aspects of

consciousness as possible, including emotion, imagination, dreams, mystical experi-

ences and so on. Although such an all-embracing theory will be necessary in the long

run, we thought it wiser to begin with the particular aspect of consciousness that is

likely to yield most easily. What this aspect may be is a matter of personal judgment.

We selected the mammalian visual system because humans are very visual animals

and because so much experimental and theoretical work has already been done on it.

It is not easy to grasp exactly what we need to explain, and it will take many care-

ful experiments before visual consciousness can be described scientifically. We did ©19

97 C

. HER

SCO

VIC

I, BR

USS

ELS;

AR

TIST

S RI

GH

TS S

OC

IETY

(ARS

), N

EW Y

ORK

Lewis Carroll’s vanishing cat can be used to study awareness.

Copyright 1997 Scientific American, Inc.

Page 15: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

not attempt to define consciousness it-self because of the dangers of prema-ture definition. (If this seems like acopout, try defining the word “gene”—you will not find it easy.) Yet the experi-mental evidence that already exists pro-vides enough of a glimpse of the natureof visual consciousness to guide re-search. In this article, we will attemptto show how this evidence opens theway to attack this profound and in-triguing problem.

Visual theorists agree that the prob-lem of visual consciousness is ill posed.The mathematical term “ill posed”means that additional constraints areneeded to solve the problem. Althoughthe main function of the visual systemis to perceive objects and events in theworld around us, the information avail-able to our eyes is not sufficient by itselfto provide the brain with its unique in-terpretation of the visual world. Thebrain must use past experience (eitherits own or that of our distant ancestors,which is embedded in our genes) to helpinterpret the information com-ing into our eyes. An examplewould be the derivation of thethree-dimensional representa-tion of the world from thetwo-dimensional signals fall-ing onto the retinas of our twoeyes or even onto one of them.

Visual theorists also would

agree that seeing is a constructive pro-cess, one in which the brain has to car-ry out complex activities (sometimescalled computations) in order to decidewhich interpretation to adopt of theambiguous visual input. “Computa-tion” implies that the brain acts to forma symbolic representation of the visualworld, with a mapping (in the mathe-matical sense) of certain aspects of thatworld onto elements in the brain.

Ray Jackendoff of Brandeis Univer-sity postulates, as do most cognitive sci-entists, that the computations carriedout by the brain are largely unconsciousand that what we become aware of isthe result of these computations. Butwhile the customary view is that thisawareness occurs at the highest levels ofthe computational system, Jackendoffhas proposed an intermediate-level the-ory of consciousness.

What we see, Jackendoff suggests, re-lates to a representation of surfaces thatare directly visible to us, together withtheir outline, orientation, color, texture

and movement. (This idea has similari-ties to what the late David C. Marr ofthe Massachusetts Institute of Technol-ogy called a “2 1/2-dimensional sketch.”It is more than a two-dimensional sketchbecause it conveys the orientation of thevisible surfaces. It is less than three-di-mensional because depth information isnot explicitly represented.) In the nextstage this sketch is processed by thebrain to produce a three-dimensionalrepresentation. Jackendoff argues thatwe are not visually aware of this three-dimensional representation.

An example may make this processclearer. If you look at a person whoseback is turned to you, you can see theback of the head but not the face. Nev-ertheless, your brain infers that the per-son has a face. We can deduce as muchbecause if that person turned aroundand had no face, you would be verysurprised.

The viewer-centered representationthat corresponds to the visible back ofthe head is what you are vividly aware

of. What your brain infersabout the front would comefrom some kind of three-di-mensional representation. Thisdoes not mean that informa-tion flows only from the sur-face representation to the three-dimensional one; it almost cer-tainly flows in both directions.

AMBIGUOUS IMAGES were frequently used by SalvadorDali in his paintings. In Slave Market with the Disappear-ing Bust of Voltaire (1940), the bust of the French philoso-pher Voltaire is apparent from a distance but transformsinto figures of three people when viewed at close range.Studies with ambiguous figures in the behaving monkey

have found that many neurons in higher cortical areas re-spond only to the currently “perceived” figure; the neu-

ronal response to the “unseen” image is suppressed.

©19

97 D

EMA

RT

PRO

AR

TE (R

), G

ENEV

A/A

RTIS

TS R

IGH

TS S

OC

IETY

(ARS

), N

EW Y

ORK

; © S

ALV

AD

OR

DA

LI M

USE

UM

, IN

C.,

ST. P

ETER

SBU

RG, F

LORI

DA

Copyright 1997 Scientific American, Inc.

Page 16: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

When you imagine the front of the face,what you are aware of is a surface rep-resentation generated by informationfrom the three-dimensional model.

It is important to distinguish betweenan explicit and an implicit representa-tion. An explicit representation is some-thing that is symbolized without furtherprocessing. An implicit representationcontains the same information but re-quires further processing to make it ex-plicit. The pattern of colored dots on atelevision screen, for example, containsan implicit representation of objects(say, a person’s face), but only the dotsand their locations are explicit. Whenyou see a face on the screen, there mustbe neurons in your brain whose firing,in some sense, symbolizes that face.

We call this pattern of firing neuronsan active representation. A latent repre-sentation of a face must also be storedin the brain, probably as a special pat-tern of synaptic connections betweenneurons. For example, you probablyhave a representation of the Statue ofLiberty in your brain, a representationthat usually is inactive. If you do thinkabout the Statue, the representation be-comes active, with the relevant neuronsfiring away.

An object, incidentally, may be repre-sented in more than one way—as a vi-sual image, as a set of words and theirrelated sounds, or even as a touch or asmell. These different representationsare likely to interact with one another.The representation is likely to be dis-tributed over many neurons, both local-ly and more globally. Such a representa-tion may not be as simple and straight-forward as uncritical introspection mightindicate. There is suggestive evidence,partly from studying how neurons fire invarious parts of a monkey’s brain andpartly from examining the effects of cer-tain types of brain damage in humans,that different aspects of a face—and ofthe implications of a face—may be rep-resented in different parts of the brain.

First, there is the representation of aface as a face: two eyes, a nose, a mouthand so on. The neurons involved areusually not too fussy about the exactsize or position of this face in the visualfield, nor are they very sensitive to smallchanges in its orientation. In monkeys,there are neurons that respond bestwhen the face is turning in a particulardirection, while others seem to be moreconcerned with the direction in whichthe eyes are gazing.

Then there are representations of the

parts of a face, as separate from thosefor the face as a whole. Further, the im-plications of seeing a face, such as thatperson’s sex, the facial expression, thefamiliarity or unfamiliarity of the face,and in particular whose face it is, mayeach be correlated with neurons firingin other places.

What we are aware of at any moment,in one sense or another, is not a simplematter. We have suggested that theremay be a very transient form of fleetingawareness that represents only rathersimple features and does not require anattentional mechanism. From this briefawareness the brain constructs a view-er-centered representation—what we see

vividly and clearly—that does requireattention. This in turn probably leadsto three-dimensional object representa-tions and thence to more cognitive ones.

Representations corresponding to viv-id consciousness are likely to have spe-cial properties. William James thoughtthat consciousness involved both atten-tion and short-term memory. Most psy-chologists today would agree with thisview. Jackendoff writes that conscious-ness is “enriched” by attention, implyingthat whereas attention may not be es-sential for certain limited types of con-sciousness, it is necessary for full con-sciousness. Yet it is not clear exactlywhich forms of memory are involved.Is long-term memory needed? Someforms of acquired knowledge are so

embedded in the machinery of neuralprocessing that they are almost certainlyused in becoming aware of something.On the other hand, there is evidencefrom studies of brain-damaged patientsthat the ability to lay down new long-term episodic memories is not essentialfor consciousness to be experienced.

It is difficult to imagine that anyonecould be conscious if he or she had nomemory whatsoever of what had justhappened, even an extremely short one.Visual psychologists talk of iconic mem-ory, which lasts for a fraction of a sec-ond, and working memory (such as thatused to remember a new telephone num-ber) that lasts for only a few seconds un-less it is rehearsed. It is not clear wheth-er both of these are essential for con-sciousness. In any case, the division ofshort-term memory into these two cate-gories may be too crude.

If these complex processes of visualawareness are localized in parts of thebrain, which processes are likely to bewhere? Many regions of the brain maybe involved, but it is almost certain thatthe cerebral neocortex plays a dominantrole. Visual information from the retinareaches the neocortex mainly by way ofa part of the thalamus (the lateral genic-ulate nucleus); another significant visualpathway from the retina is to the superi-or colliculus, at the top of the brain stem.

The cortex in humans consists of twointricately folded sheets of nerve tissue,one on each side of the head. Thesesheets are connected by a large tract ofabout half a billion axons called the cor-pus callosum. It is well known that if thecorpus callosum is cut, as is done forcertain cases of intractable epilepsy, oneside of the brain is not aware of whatthe other side is seeing. In particular, theleft side of the brain (in a right-handedperson) appears not to be aware of vi-sual information received exclusivelyby the right side. This shows that noneof the information required for visualawareness can reach the other side ofthe brain by traveling down to the brainstem and, from there, back up. In a nor-mal person, such information can get tothe other side only by using the axonsin the corpus callosum.

A different part of the brain—the hip-pocampal system—is involved in one-shot, or episodic, memories that, overweeks and months, it passes on to theneocortex. This system is so placed thatit receives inputs from, and projects to,many parts of the brain. Thus, one mightsuspect that the hippocampal system is

The Problem of Consciousness Mysteries of the Mind 21

WILLIAM JAMES, the father of Ameri-can psychology, observed that conscious-

ness is not a thing but a process.

BETT

MA

NN

AR

CH

IVE

Copyright 1997 Scientific American, Inc.

Page 17: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

the essential seat of consciousness. Thisis not the case: evidence from studies ofpatients with damaged brains showsthat this system is not essential for visu-al awareness, although naturally a pa-tient lacking one is severely handicappedin everyday life because he cannot re-member anything that took place morethan a minute or so in the past.

In broad terms, the neocortex of alertanimals probably acts in two ways. Bybuilding on crude and somewhat re-dundant wiring, produced by our genesand by embryonic processes, the neo-cortex draws on visual and other expe-rience to slowly “rewire” itself to createcategories (or “features”) it can respondto. A new category is not fully createdin the neocortex after exposure to onlyone example of it, although some smallmodifications of the neural connectionsmay be made.

The second function of the neocortex(at least of the visual part of it) is to re-spond extremely rapidly to incomingsignals. To do so, it uses the categoriesit has learned and tries to find the com-binations of active neurons that, on thebasis of its past experience, are mostlikely to represent the relevant objectsand events in the visual world at thatmoment. The formation of such coali-tions of active neurons may also be in-fluenced by biases coming from otherparts of the brain: for example, signalstelling it what best to attend to or high-level expectations about the nature ofthe stimulus.

Consciousness, as James noted, is al-ways changing. These rapidly formedcoalitions occur at different levels andinteract to form even broader coalitions.They are transient, lasting usually foronly a fraction of a second. Because co-alitions in the visual system are the basisof what we see, evolution has seen to itthat they form as fast as possible; other-wise, no animal could survive. The brainis handicapped in forming neuronal co-alitions rapidly because, by computerstandards, neurons act very slowly. Thebrain compensates for this relative slow-ness partly by using very many neu-rons, simultaneously and in parallel,and partly by arranging the system in aroughly hierarchical manner.

If visual awareness at any momentcorresponds to sets of neurons firing,then the obvious question is: Where are these neurons located in the brain,and in what way are they firing? Visualawareness is highly unlikely to occupyall the neurons in the neocortex that are

firing above their background rate at aparticular moment. We would expectthat, theoretically, at least some of theseneurons would be involved in doingcomputations—trying to arrive at thebest coalitions—whereas others wouldexpress the results of these computa-tions, in other words, what we see.

Fortunately, some experimental evi-dence can be found to back up this the-oretical conclusion. A phenomenoncalled binocular rivalry may help iden-tify the neurons whose firing symbolizesawareness. This phenomenon can beseen in dramatic form in an exhibit pre-pared by Sally Duensing and Bob Millerat the Exploratorium in San Francisco.

Conflicting Inputs

Binocular rivalry occurs when eacheye has a different visual input relat-

ing to the same part of the visual field.The early visual system on the left sideof the brain receives an input from botheyes but sees only the part of the visualfield to the right of the fixation point.The converse is true for the right side. Ifthese two conflicting inputs are rival-rous, one sees not the two inputs super-imposed but first one input, then theother, and so on in alternation.

In the exhibit, called “The CheshireCat,” viewers put their heads in a fixedplace and are told to keep the gaze fixed.By means of a suitably placed mirror,one of the eyes can look at another per-son’s face, directly in front, while theother eye sees a blank white screen to theside. If the viewer waves a hand in frontof this plain screen at the same locationin his or her visual field occupied by theface, the face is wiped out. The move-ment of the hand, being visually verysalient, has captured the brain’s atten-tion. Without attention the face cannotbe seen. If the viewer moves the eyes,the face reappears.

In some cases, only part of the facedisappears. Sometimes, for example,one eye, or both eyes, will remain. If theviewer looks at the smile on the person’sface, the face may disappear, leavingonly the smile. For this reason, the ef-fect has been called the Cheshire Cat ef-fect, after the cat in Lewis Carroll’s Al-ice’s Adventures in Wonderland.

Although it is very difficult to recordactivity in individual neurons in a hu-man brain, such studies can be done inmonkeys. A simple example of binoc-ular rivalry has been studied in a mon-key by Nikos K. Logothetis and JeffreyD. Schall, both then at M.I.T. They

The Problem of Consciousness22 Mysteries of the Mind

This simple experiment with a mirror illustrates one aspect of visual awareness.It relies on a phenomenon called binocular rivalry, which occurs when each eye

has a different input from the same part of the visual field. Motion in the field of oneeye can cause either the entire image or parts of the image to be erased. The move-ment captures the brain’s attention.

The Cheshire Cat Experiment

a

Copyright 1997 Scientific American, Inc.

Page 18: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

trained a macaque to keep its eyes stilland to signal whether it is seeing upwardor downward movement of a horizon-tal grating. To produce rivalry, upwardmovement is projected into one of themonkey’s eyes and downward move-ment into the other, so that the two im-ages overlap in the visual field. The mon-key signals that it sees up and downmovements alternatively, just as humanswould. Even though the motion stimu-lus coming into the monkey’s eyes is al-ways the same, the monkey’s perceptchanges every second or so.

Cortical area MT (which some re-searchers prefer to label V5) is an areamainly concerned with movement. Whatdo the neurons in MT do when the mon-key’s percept is sometimes up and some-times down? (The researchers studiedonly the monkey’s first response.) Thesimplified answer—the actual data arerather more messy—is that whereas thefiring of some of the neurons correlateswith the changes in the percept, for oth-ers the average firing rate is relativelyunchanged and independent of whichdirection of movement the monkey isseeing at that moment. Thus, it is un-likely that the firing of all the neuronsin the visual neocortex at one particularmoment corresponds to the monkey’s

visual awareness. Exactly which neu-rons do correspond to awareness re-mains to be discovered.

We have postulated that when weclearly see something, there must be neu-rons actively firing that stand for whatwe see. This might be called the activityprinciple. Here, too, there is some ex-perimental evidence. One example is thefiring of neurons in a specific corticalvisual area in response to illusory con-tours. Another and perhaps more strik-ing case is the filling in of the blind spot.The blind spot in each eye is caused bythe lack of photoreceptors in the area ofthe retina where the optic nerve leavesthe retina and projects to the brain. Itslocation is about 15 degrees from thefovea (the visual center of the eye). Yetif you close one eye, you do not see ahole in your visual field.

Philosopher Daniel C. Dennett ofTufts University is unusual among phi-losophers in that he is interested both inpsychology and in the brain. This inter-est is much to be welcomed. In a recentbook, Consciousness Explained, he hasargued that it is wrong to talk about fill-ing in. He concludes, correctly, that “anabsence of information is not the sameas information about an absence.” Fromthis general principle he argues that the

brain does not fill in the blind spot butrather ignores it.

Dennett’s argument by itself, howev-er, does not establish that filling in doesnot occur; it only suggests that it mightnot. Dennett also states that “your brainhas no machinery for [filling in] at thislocation.” This statement is incorrect.The primary visual cortex lacks a directinput from one eye, but normal “ma-chinery” is there to deal with the inputfrom the other eye. Ricardo Gattass andhis colleagues at the Federal Universityof Rio de Janeiro have shown that inthe macaque some of the neurons in theblind-spot area of the primary visualcortex do respond to input from botheyes, probably assisted by inputs fromother parts of the cortex. Moreover, inthe case of simple filling in, some of theneurons in that region respond as ifthey were actively filling in.

Thus, Dennett’s claim about blindspots is incorrect. In addition, psycho-logical experiments by Vilayanur S. Ra-machandran [see “Blind Spots,” Scien-tific American, May 1992] haveshown that what is filled in can be quitecomplex depending on the overall con-text of the visual scene. How, he argues,can your brain be ignoring somethingthat is in fact commanding attention?

Filling in, therefore, is not to be dis-missed as nonexistent or unusual. Itprobably represents a basic interpola-tion process that can occur at many lev-els in the neocortex. It is, incidentally, agood example of what is meant by aconstructive process.

How can we discover the neuronswhose firing symbolizes a particularpercept? William T. Newsome and hiscolleagues at Stanford University havedone a series of brilliant experimentson neurons in cortical area MT of themacaque’s brain. By studying a neuronin area MT, we may discover that it re-sponds best to very specific visual fea-tures having to do with motion. A neu-ron, for instance, might fire strongly inresponse to the movement of a bar in aparticular place in the visual field, butonly when the bar is oriented at a cer-tain angle, moving in one of the two di-rections perpendicular to its length with-in a certain range of speed.

It is technically difficult to excite just asingle neuron, but it is known that neu-rons that respond to roughly the sameposition, orientation and direction ofmovement of a bar tend to be locatednear one another in the cortical sheet.The experimenters taught the monkey a

The Problem of Consciousness Mysteries of the Mind 23

To observe the effect, a viewer divides the field of vision with a mirror placed be-tween the eyes (a). One eye sees the cat; the other eye a reflection in the mirror of awhite wall or background. The viewer then waves the hand that corresponds to theeye looking at the mirror so that the hand passes through the area in which the im-age of the cat appears in the other eye (b). The result is that the cat may disappear. Orif the viewer was attentive to a specific feature before the hand was waved, thoseparts—the eyes or even a mocking smile—may remain (c). —F.C. and C.K.

cb

PHO

TOG

RA

PHS

BY

JASO

N G

OLT

Z

Copyright 1997 Scientific American, Inc.

Page 19: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

simple task in movement discriminationusing a mixture of dots, some movingrandomly, the rest all in one direction.They showed that electrical stimulationof a small region in the right place incortical area MT would bias the mon-key’s motion discrimination, almost al-ways in the expected direction.

Thus, the stimulation of these neuronscan influence the monkey’s behavior andprobably its visual percept. Such exper-iments do not, however, show decisive-ly that the firing of such neurons is theexact neural correlate of the percept. Thecorrelate could be only a subset of theneurons being activated. Or perhapsthe real correlate is the firing of neuronsin another part of the visual hierarchythat are strongly influenced by the neu-rons activated in area MT.

These same reservations apply also tocases of binocular rivalry. Clearly, theproblem of finding the neurons whosefiring symbolizes a particular percept isnot going to be easy. It will take manycareful experiments to track them downeven for one kind of percept.

It seems obvious that the purpose ofvivid visual awareness is to feed intothe cortical areas concerned with theimplications of what we see; from therethe information shuttles on the onehand to the hippocampal system, to beencoded (temporarily) into long-termepisodic memory, and on the other tothe planning levels of the motor system.But is it possible to go from a visual in-put to a behavioral output without anyrelevant visual awareness?

That such a process can happen isdemonstrated by the remarkable classof patients with “blindsight.” These pa-tients, all of whom have suffered dam-age to their visual cortex, can point withfair accuracy at visual targets or trackthem with their eyes while vigorouslydenying seeing anything. In fact, thesepatients are as surprised as their doc-tors by their abilities. The amount of in-formation that “gets through,” howev-er, is limited: blindsight patients havesome ability to respond to wavelength,orientation and motion, yet they cannotdistinguish a triangle from a square.

It is naturally of great interest to knowwhich neural pathways are being usedin these patients. Investigators originallysuspected that the pathway ran throughthe superior colliculus. Recent experi-ments suggest that a direct albeit weakconnection may be involved betweenthe lateral geniculate nucleus and othervisual areas in the cortex. It is unclear

The Problem of Consciousness24 Mysteries of the Mind

OPTICAL ILLUSION devised by Vilayanur S. Ramachandran illustrates the brain’sability to fill in, or construct, visual information that is missing because it falls on theblind spot of the eye. When you look at the patterns of broken green bars, the visualsystem produces two illusory contours defining a vertical strip. Now shut your righteye and focus on the white square in the green series of bars. Move the page towardyour eye until the blue dot disappears (roughly six inches in front of your nose). Mostobservers report seeing the vertical strip completed across the blind spot, not the bro-ken line. Try the same experiment with the series of just three red bars. The illusoryvertical contours are less well defined, and the visual system tends to fill in the horizon-tal bar across the blind spot. Thus, the brain fills in differently depending on the over-all context of the image.

JOH

NN

Y JO

HN

SON

Copyright 1997 Scientific American, Inc.

Page 20: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

whether an intact primary visual cortexregion is essential for immediate visualawareness. Conceivably the visual sig-nal in blindsight is so weak that theneural activity cannot produce aware-ness, although it remains strong enoughto get through to the motor system.

Normal-seeing people regularly re-spond to visual signals without beingfully aware of them. In automatic ac-tions, such as swimming or driving acar, complex but stereotypical actionsoccur with little, if any, associated visu-al awareness. In other cases, the infor-mation conveyed is either very limitedor very attenuated. Thus, while we canfunction without visual awareness, ourbehavior without it is rather restricted.

Clearly, it takes a certain amount oftime to experience a conscious percept.It is difficult to determine just howmuch time is needed for an episode ofvisual awareness, but one aspect of theproblem that can be demonstrated ex-perimentally is that signals receivedclose together in time are treated by thebrain as simultaneous.

A disk of red light is flashed for, say,20 milliseconds, followed immediatelyby a 20-millisecond flash of green lightin the same place. The subject reportsthat he did not see a red light followedby a green light. Instead he saw a yel-low light, just as he would have if thered and the green light had been flashedsimultaneously. Yet the subject couldnot have experienced yellow until afterthe information from the green flashhad been processed and integrated withthe preceding red one.

Experiments of this type led psychol-ogist Robert Efron, now at the Univer-sity of California at Davis, to concludethat the processing period for percep-tion is about 60 to 70 milliseconds.Similar periods are found in experimentswith tones in the auditory system. It isalways possible, however, that the pro-cessing times may be different in higherparts of the visual hierarchy and in oth-er parts of the brain. Processing is alsomore rapid in trained, compared withnaive, observers.

Because it appears to be involved insome forms of visual awareness, it wouldhelp if we could discover the neural ba-sis of attention. Eye movement is a formof attention, since the area of the visualfield in which we see with high resolu-tion is remarkably small, roughly thearea of the thumbnail at arm’s length.Thus, we move our eyes to gaze directlyat an object in order to see it more clear-

ly. Our eyes usually movethree or four times asecond. Psychologistshave shown, however,that there appears to bea faster form of atten-tion that moves around,in some sense, when oureyes are stationary.

The exact psychologi-cal nature of this fasterattentional mechanism isat present controversial.Several neuroscientists,however, including Rob-ert Desimone and hiscolleagues at the Nation-al Institute of Mental Health, haveshown that the rate of firing of certainneurons in the macaque’s visual systemdepends on what the monkey is attend-ing to in the visual field. Thus, attentionis not solely a psychological concept; italso has neural correlates that can beobserved. A number of researchers havefound that the pulvinar, a region of thethalamus, appears to be involved in vi-sual attention. We would like to believethat the thalamus deserves to be called“the organ of attention,” but this statushas yet to be established.

Attention and Awareness

The major problem is to find whatactivity in the brain corresponds

directly to visual awareness. It has beenspeculated that each cortical area pro-duces awareness of only those visualfeatures that are “columnar,” or ar-ranged in the stack or column of neu-rons perpendicular to the cortical sur-face. Thus, the primary visual cortexcould code for orientation and area MTfor motion. So far experimentalists havenot found one particular region in thebrain where all the information neededfor visual awareness appears to cometogether. Dennett has dubbed such ahypothetical place “The Cartesian Thea-ter.” He argues on theoretical groundsthat it does not exist.

Awareness seems to be distributed notjust on a local scale, but more widelyover the neocortex. Vivid visual aware-ness is unlikely to be distributed overevery cortical area because some areasshow no response to visual signals.Awareness might, for example, be asso-ciated with only those areas that con-nect back directly to the primary visualcortex or alternatively with those areasthat project into one another’s layer 4.

(The latter areas are al-ways at the same level inthe visual hierarchy.)

The key issue, then, ishow the brain forms itsglobal representationsfrom visual signals. If at-tention is indeed crucialfor visual awareness, thebrain could form repre-sentations by attendingto just one object at atime, rapidly movingfrom one object to thenext. For example, theneurons representing allthe different aspects of

the attended object could all fire togeth-er very rapidly for a short period, possi-bly in rapid bursts.

This fast, simultaneous firing mightnot only excite those neurons that sym-bolized the implications of that objectbut also temporarily strengthen the rel-evant synapses so that this particularpattern of firing could be quickly re-called—a form of short-term memory. If only one representation needs to beheld in short-term memory, as in re-membering a single task, the neurons in-volved may continue to fire for a period.

A problem arises if it is necessary tobe aware of more than one object at ex-actly the same time. If all the attributesof two or more objects were represent-ed by neurons firing rapidly, their attri-butes might be confused. The color ofone might become attached to the shapeof another. This happens sometimes invery brief presentations.

Some time ago Christoph von derMalsburg, now at the Ruhr-UniversitätBochum, suggested that this difficultywould be circumvented if the neuronsassociated with any one object all firedin synchrony (that is, if their times offiring were correlated) but out of syn-chrony with those representing otherobjects. Recently two groups in Ger-many reported that there does appearto be correlated firing between neuronsin the visual cortex of the cat, often in arhythmic manner, with a frequency inthe 35- to 75-hertz range, sometimescalled 40-hertz, or g, oscillation.

Von der Malsburg’s proposal prompt-ed us to suggest that this rhythmic andsynchronized firing might be the neuralcorrelate of awareness and that it mightserve to bind together activity concern-ing the same object in different corticalareas. The matter is still undecided, butat present the fragmentary experimen-

The Problem of Consciousness Mysteries of the Mind 25

Rhythmic and

synchronized

firing may be the

neural correlate

of awareness and

might bind

together activity

[concerning the

same object]

in different

cortical areas.

Copyright 1997 Scientific American, Inc.

Page 21: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

tal evidence does rather little to supportsuch an idea. Another possibility is thatthe 40-hertz oscillations may help dis-tinguish figure from ground or assistthe mechanism of attention.

Correlates of Consciousness

Are there some particular types of neurons, distributed over the visu-

al neocortex, whose firing directly sym-bolizes the content of visual awareness?One very simplistic hypothesis is thatthe activities in the upper layers of thecortex are largely unconscious ones,whereas the activities in the lower lay-ers (layers 5 and 6) mostly correlatewith consciousness. We have wonderedwhether the pyramidal neurons in layer5 of the neocortex, especially the largerones, might play this latter role.

These are the only cortical neuronsthat project right out of the cortical sys-tem (that is, not to the neocortex, thethalamus or the claustrum). If visualawareness represents the results of neu-ral computations in the cortex, onemight expect that what the cortex sendselsewhere would symbolize those re-

sults. Moreover, the neurons in layer 5show a rather unusual propensity to firein bursts. The idea that layer 5 neuronsmay directly symbolize visual awarenessis attractive, but it still is too early totell whether there is anything in it.

Visual awareness is clearly a difficultproblem. More work is needed on thepsychological and neural basis of bothattention and very short term memory.Studying the neurons when a perceptchanges, even though the visual input is

constant, should be a powerful experi-mental paradigm. We need to constructneurobiological theories of visual aware-ness and test them using a combinationof molecular, neurobiological and clini-cal imaging studies.

We believe that once we have mas-tered the secret of this simple form ofawareness, we may be close to under-standing a central mystery of humanlife: how the physical events occurringin our brains while we think and act inthe world relate to our subjective sensa-tions—that is, how the brain relates tothe mind.

Postscript: There have been severalrelevant developments since this articlewas first published. It now seems likelythat there are rapid “on-line” systemsfor stereotyped motor responses such ashand or eye movement. These systemsare unconscious and lack memory. Con-scious seeing, on the other hand, seemsto be slower and more subject to visualillusions. The brain needs to form a con-scious representation of the visual scenethat it then can use for many differentactions or thoughts. Exactly how allthese pathways work and how they in-teract is far from clear.

There have been more experiments onthe behavior of neurons that respond tobistable visual percepts, such as binocu-lar rivalry, but it is probably too early todraw firm conclusions from them aboutthe exact neural correlates of visual con-sciousness. We have suggested on theo-retical grounds based on the the neuro-anatomy of the macaque monkey thatprimates are not directly aware of whatis happening in the primary visual cor-tex, even though most of the visual in-formation flows through it. This hypoth-esis is supported by some experimentalevidence, but it is still controversial.

The Problem of Consciousness26 Mysteries of the Mind

Further Reading

Perception. Irvin Rock. Scientific American Library, 1984.Consciousness and the Computational Mind. Ray Jackendoff. MIT Press/BradfordBooks, 1987.

Cold Spring Harbor Symposia on Quantitative Biology, Vol. LV: The Brain. ColdSpring Harbor Laboratory Press, 1990.

Towards a Neurobiological Theory of Consciousness. Francis Crick and ChristofKoch in Seminars in the Neurosciences, Vol. 2, pages 263–275; 1990.

The Computational Brain. Patricia S. Churchland and Terrence J. Sejnowski. MITPress/Bradford Books, 1992.

The Visual Brain in Action. A. David Milner and Melvyn A. Goodale. Oxford Universi-ty Press, 1995.

Are We Aware of Neural Activity in Primary Visual Cortex? Francis Crick andChristof Koch in Nature, Vol. 375, pages 121-123, May 11, 1995.

a c

b

BRIEF FLASHES of colored light enable researchers to infer the minimum time re-quired for visual awareness. A disk of red light is projected for 20 milliseconds (a), fol-lowed immediately by a 20-millisecond flash of green light (b). But the observer reportsseeing a single flash of yellow (c), the color that would be apparent if red and greenwere projected simultaneously. The subject does not become aware of red followed bygreen until the length of the flashes is extended to 60 to 70 milliseconds.

JASO

N G

OLT

Z

SA

Copyright 1997 Scientific American, Inc.

Page 22: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

30 Mysteries of the Mind Reprinted from the December 1995 issue

The Author

DAVID J. CHALMERS stud-ied mathematics at AdelaideUniversity and as a RhodesScholar at the University ofOxford, but a fascinationwith consciousness led himinto philosophy and cognitivescience. He has a Ph.D. inthese fields from Indiana Uni-versity and is currently in thedepartment of philosophy atthe University of California,Santa Cruz. Chalmers haspublished numerous articleson artificial intelligence andthe philosophy of mind.

Conscious experience is at once the mostfamiliar thing in the world and the mostmysterious. There is nothing we knowabout more directly than consciousness,

but it is extraordinarily hard to reconcile it with every-thing else we know. Why does it exist? What does itdo? How could it possibly arise from neural processesin the brain? These questions are among the most in-triguing in all of science.

From an objective viewpoint, the brain is relativelycomprehensible. When you look at this page, there isa whir of processing: photons strike your retina, elec-trical signals are passed upyour optic nerve and be-tween different areas of yourbrain, and eventually youmight respond with a smile,a perplexed frown or a re-mark. But there is also asubjective aspect. When youlook at the page, you areconscious of it, directly ex-periencing the images andwords as part of your pri-vate, mental life. You havevivid impressions of the col-ors and shapes of the im-ages. At the same time, youmay be feeling some emotions and forming somethoughts. Together such experiences make up con-sciousness: the subjective, inner life of the mind.

For many years, consciousness was shunned by re-searchers studying the brain and the mind. The pre-vailing view was that science, which depends on ob-jectivity, could not accommodate something as subjec-tive as consciousness. The behaviorist movement inpsychology, dominant earlier in this century, concen-trated on external behavior and disallowed any talk ofinternal mental processes. Later, the rise of cognitivescience focused attention on processes inside the head.Still, consciousness remained off-limits, fit only forlate-night discussion over drinks.

The Puzzle of Conscious ExperienceNeuroscientists and others are at last plumbing one of the most profound mysteries of existence. But knowledgeof the brain alone may not get them to the bottom of it

by David J. Chalmers

CONSCIOUSNESS,the subjective experi-ence of an inner self,could be a phenome-non forever beyond

the reach of neurosci-ence. Even a detailed

knowledge of thebrain’s workings and

the neural correlates ofconsciousness may failto explain how or why

human beings haveself-aware minds.

DU

SAN

PET

RIC

IC

REN

É M

AG

RITT

E Th

e D

oubl

e Se

cret

, 192

7; ©

199

7 C

. HER

SCO

VIC

I, B

RUSS

ELS;

ART

ISTS

RIG

HTS

SO

CIE

TY (A

RS),

NEW

YO

RK; G

IRA

UD

ON

/AR

T RE

SOU

RCE,

NEW

YO

RK

Peering into our inner selves canbe frustrating.

Copyright 1995 Scientific American, Inc.

Page 23: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Over the past several years, however,an increasing number of neuroscien-tists, psychologists and philosophershave been rejecting the idea that con-sciousness cannot be studied and are at-tempting to delve into its secrets. Asmight be expected of a field so new,there is a tangle of diverse and conflict-ing theories, often using basic conceptsin incompatible ways. To help unsnarlthe tangle, philosophical reasoning isvital.

The myriad views within the fieldrange from reductionist theories, ac-cording to which consciousness can beexplained by the standard methods of

neuroscience and psychology, to the po-sition of the so-called mysterians, whosay we will never understand conscious-ness at all. I believe that on close analy-sis both of these views can be seen to bemistaken and that the truth lies some-where in the middle.

Against reductionism I will argue thatthe tools of neuroscience cannot providea full account of conscious experience, al-though they have much to offer. Againstmysterianism I will hold that conscious-ness might be explained by a new kindof theory. The full details of such a the-ory are still out of reach, but careful rea-soning and some educated inferences

can reveal something of its general na-ture. For example, it will probably in-volve new fundamental laws, and theconcept of information may play a cen-tral role. These faint glimmerings sug-gest that a theory of consciousness mayhave startling consequences for our viewof the universe and of ourselves.

The Hard Problem

Researchers use the word “conscious-ness” in many different ways. To

clarify the issues, we first have to sepa-rate the problems that are often clus-tered together under the name. For this

Copyright 1995 Scientific American, Inc.

Page 24: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

purpose, I find it useful to distinguishbetween the “easy problems” and the“hard problem” of consciousness. Theeasy problems are by no means trivial—they are actually as challenging as mostin psychology and biology—but it iswith the hard problem that the centralmystery lies.

The easy problems of consciousnessinclude the following: How can a hu-man subject discriminate sensory stim-uli and react to them ap-propriately? How doesthe brain integrate infor-mation from many dif-ferent sources and usethis information to con-trol behavior? How is itthat subjects can verbal-ize their internal states?Although all these ques-tions are associated withconsciousness, they all

concern the objective mechanisms ofthe cognitive system. Consequently, wehave every reason to expect that contin-ued work in cognitive psychology andneuroscience will answer them.

The hard problem, in contrast, is thequestion of how physical processes inthe brain give rise to subjective experi-ence. This puzzle involves the inner as-pect of thought and perception: the waythings feel for the subject. When we see,

for example, we experi-ence visual sensations,such as that of vivid blue.Or think of the ineffablesound of a distant oboe,the agony of an intensepain, the sparkle of hap-piness or the meditativequality of a moment lostin thought. All are part ofwhat I call consciousness.It is these phenomena

that pose the real mystery of the mind.To illustrate the distinction, consider

a thought experiment devised by theAustralian philosopher Frank Jackson.Suppose that Mary, a neuroscientist inthe 23rd century, is the world’s leadingexpert on the brain processes responsi-ble for color vision. But Mary has livedher whole life in a black-and-white roomand has never seen any other colors. Sheknows everything there is to know aboutphysical processes in the brain—its biol-ogy, structure and function. This under-standing enables her to grasp all there isto know about the easy problems: howthe brain discriminates stimuli, integratesinformation and produces verbal reports.From her knowledge of color vision, sheknows how color names correspondwith wavelengths on the light spectrum.But there is still something crucial aboutcolor vision that Mary does not know:

The Puzzle of Conscious Experience32 Mysteries of the Mind

BLA

CK-

AN

D-W

HIT

E PH

OTO

GR

APH

BY

DA

N W

AG

NER

; DIG

ITA

L C

OM

POSI

TIO

N B

Y TO

M D

RA

PER

DES

IGN

ISOLATEDNEUROSCIENTIST in a black-and-white

room knows everythingabout how the brainprocesses colors but

does not know what itis like to see them. By

itself, empirical knowl-edge of the brain does

not yield completeknowledge of conscious

experience.

Continued on page 34

Copyright 1995 Scientific American, Inc.

Page 25: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

We believe that at the moment the best approach to the prob-lem of explaining consciousness is to concentrate on finding

what is known as the neural correlates of consciousness—the pro-cesses in the brain that are most directly responsible for conscious-ness. By locating the neurons in the cerebral cortex that correlatebest with consciousness, and figuring out howthey link to neurons elsewhere in the brain, wemay come across key insights into what David J.Chalmers calls the hard problem: a full accountingof the manner in which subjective experience aris-es from these cerebral processes.

We commend Chalmers for boldly recognizingand focusing on the hard problem at this earlystage, although we are not as enthusiastic aboutsome of his thought experiments. As we see it, thehard problem can be broken down into severalquestions: Why do we experience anything at all?What leads to a particular conscious experience(such as the blueness of blue)? Why are some as-pects of subjective experience impossible to con-vey to other people (in other words, why are theyprivate)? We believe we have an answer to the last problem and asuggestion about the first two, revolving around a phenomenonknown as explicit neuronal representation.

What does “explicit” mean in this context? Perhaps the best way todefine it is with an example. In response to the image of a face, say,ganglion cells fire all over the retina, much like the pixels on a televi-sion screen, to generate an implicit representation of the face. At thesame time, they can also respond to a great many other features inthe image, such as shadows, lines, uneven lighting and so on. In con-trast, some neurons high in the hierarchy of the visual cortex respondmainly to the face or even to the face viewed at a particular angle.Such neurons help the brain represent the face in an explicit manner.Their loss, resulting from a stroke or some other brain injury, leads toprosopagnosia, an individual’s inability to recognize familiar facesconsciously—even his or her own, although the person can still iden-tify a face as a face. Similarly, damage to other parts of the visual cor-tex can cause someone to lose the ability to experience color, whilestill seeing in shades of black and white, even though there is no de-fect in the color receptors in the eye.

At each stage, visual information is reencoded, typically in a semi-hierarchical manner. Retinal ganglion cells respond to a spot of light.Neurons in the primary visual cortex are most adept at responding tolines or edges; neurons higher up might prefer a moving contour. Stillhigher are those that respond to faces and other familiar objects. Ontop are those that project to pre-motor and motor structures in thebrain, where they fire the neurons that initiate such actions as speak-ing or avoiding an oncoming automobile.

Chalmers believes, as we do, that the subjective aspects of an ex-perience must relate closely to the firing of the neurons correspond-ing to those aspects (the neural correlates). He describes a well-known thought experiment, constructed around a hypothetical neu-roscientist, Mary, who specializes in color perception but has neverseen a color. We believe the reason Mary does not know what it islike to see a color, however, is that she has never had an explicit neu-ral representation of a color in her brain, only of the words and ideasassociated with colors.

In order to describe a subjective visual experience, the informationhas to be transmitted to the motor output stage of the brain, where itbecomes available for verbalization or other actions. This transmissionalways involves reencoding the information, so that the explicit infor-

mation expressed by the motor neurons is related, but not identical,to the explicit information expressed by the firing of the neurons as-sociated with color experience, at some level in the visual hierarchy.

It is not possible, then, to convey with words and ideas the exactnature of a subjective experience. It is possible, however, to convey a

difference between subjective experiences—todistinguish between red and orange, for example.This is possible because a difference in a high-levelvisual cortical area will still be associated with a dif-ference in the motor stages. The implication is thatwe can never explain to other people the subjec-tive nature of any conscious experience, only its re-lation to other ones.

The other two questions, concerning why wehave conscious experiences and what leads tospecific ones, appear more difficult. Chalmers pro-poses that they require the introduction of “experi-ence” as a fundamental new feature of the world,relating to the ability of an organism to process in-formation. But which types of neuronal informa-tion produce consciousness? And what makes a

certain type of information correspond to the blueness of blue,rather than the greenness of green? Such problems seem as difficultas any in the study of consciousness.

We prefer an alternative approach, involving the concept of“meaning.” In what sense can neurons that explicitly code for a facebe said to convey the meaning of a face to the rest of the brain? Sucha property must relate to the cells’ projective field—the pattern ofsynaptic connections to neurons that code explicitly for related con-cepts. Ultimately, these connections extend to the motor output. Forexample, neurons responding to a certain face might be connectedto ones expressing the name of the person whose face it is and toothers for her voice, memories involving her and so on. Such associa-tions among neurons must be behaviorally useful—in other words,consistent with feedback from the body and the external world.

Meaning derives from the linkages among these representationswith others spread throughout the cortical system in a vast associa-tional network, similar to a dictionary or a relational database. Themore diverse these connections, the richer the meaning. If, as in ourprevious example of prosopagnosia, the synaptic output of such faceneurons were blocked, the cells would still respond to the person’sface, but there would be no associated meaning and, therefore,much less experience. Therefore, a face would be seen but not rec-ognized as such.

Of course, groups of neurons can take on new functions, allowingbrains to learn new categories (including faces) and associate newcategories with existing ones. Certain primitive associations, such aspain, are to some extent inborn but subsequently refined in life.

Information may indeed be the key concept, as Chalmers suspects.Greater certainty will require consideration of highly parallel streamsof information, linked—as are neurons—in complex networks. Itwould be useful to try to determine what features a neural network(or some other such computational embodiment) must have to gen-erate meaning. It is possible that such exercises will suggest the neu-ral basis of meaning. The hard problem of consciousness may thenappear in an entirely new light. It might even disappear.

FRANCIS CRICK is Kieckhefer Distinguished Research Professor at theSalk Institute for Biological Studies in San Diego.

CHRISTOF KOCH is professor of computation and neural systems atthe California Institute of Technology.

Why Neuroscience May Be Able to Explain Consciousness

by Francis Crick and Christof Koch

KANIZSA TRIANGLE stimu-lates neurons that code explic-itly for such illusory contours.

The Puzzle of Conscious Experience Mysteries of the Mind 33

Copyright 1995 Scientific American, Inc.

Page 26: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

what it is like to experience a color suchas red. It follows that there are factsabout conscious experience that cannotbe deduced from physical facts aboutthe functioning of the brain.

Indeed, nobody knows why thesephysical processes are accompanied byconscious experience at all. Why is itthat when our brains process light of acertain wavelength, we have an experi-ence of deep purple? Why do we haveany experience at all? Could not an un-conscious automaton have performedthe same tasks just as well? These arequestions that we would like a theoryof consciousness to answer.

Is Neuroscience Enough?

Iam not denying that consciousnessarises from the brain. We know, for

example, that the subjective experienceof vision is closely linked to processes inthe visual cortex. It is the link itself thatperplexes, however. Remarkably, sub-jective experience seems to emergefrom a physical process. But we haveno idea how or why this is.

Given the flurry of recent work onconsciousness in neuroscience and psy-chology, one might think this mystery isstarting to be cleared up. On closer ex-amination, however, it turnsout that almost all the cur-rent work addresses only theeasy problems of conscious-ness. The confidence of thereductionist view comes fromthe progress on the easy prob-lems, but none of this makesany difference where thehard problem is concerned.

Consider the hypothesisput forward by neurobiolo-gists Francis Crick of theSalk Institute for BiologicalStudies in San Diego andChristof Koch of the Cali-fornia Institute of Technolo-gy. They suggest that con-sciousness may arise fromcertain oscillations in thecerebral cortex, which be-come synchronized as neu-rons fire 40 times per sec-ond. Crick and Koch believethe phenomenon might ex-plain how different attrib-utes of a single perceived ob-ject (its color and shape, forexample), which are pro-cessed in different parts of

the brain, are merged into a coherentwhole. In this theory, two pieces of in-formation become bound together pre-cisely when they are represented bysynchronized neural firings.

The hypothesis could conceivably elu-cidate one of the easy problems abouthow information is integrated in thebrain. But why should synchronized os-cillations give rise to a visual experience,no matter how much integration is tak-ing place? This question involves thehard problem, about which the theoryhas nothing to offer. Indeed, Crick andKoch are agnostic about whether thehard problem can be solved by scienceat all [see box on preceding page].

The same kind of critique could beapplied to almost all the recent work onconsciousness. In his 1991 book Con-sciousness Explained, philosopher Dan-iel C. Dennett laid out a sophisticatedtheory of how numerous independentprocesses in the brain combine to pro-duce a coherent response to a perceivedevent. The theory might do much to ex-plain how we produce verbal reports onour internal states, but it tells us verylittle about why there should be a sub-jective experience behind these reports.Like other reductionist theories, Den-nett’s is a theory of the easy problems.

The critical common trait among

these easy problems is that they all con-cern how a cognitive or behavioral func-tion is performed. All are ultimatelyquestions about how the brain carriesout some task—how it discriminatesstimuli, integrates information, produc-es reports and so on. Once neurobiolo-gy specifies appropriate neural mecha-nisms, showing how the functions areperformed, the easy problems are solved.

The hard problem of consciousness,in contrast, goes beyond problems abouthow functions are performed. Even ifevery behavioral and cognitive functionrelated to consciousness were explained,there would still remain a further mys-tery: Why is the performance of thesefunctions accompanied by conscious ex-perience? It is this additional conundrumthat makes the hard problem hard.

The Explanatory Gap

Some have suggested that to solve thehard problem, we need to bring in

new tools of physical explanation: non-linear dynamics, say, or new discoveriesin neuroscience, or quantum mechan-ics. But these ideas suffer from exactlythe same difficulty. Consider a proposalfrom Stuart R. Hameroff of the Univer-sity of Arizona and Roger Penrose of theUniversity of Oxford. They hold that

consciousness arises fromquantum-physical processestaking place in microtubules,which are protein structuresinside neurons. It is possible(if not likely) that such a hy-pothesis will lead to an ex-planation of how the brainmakes decisions or even howit proves mathematical theo-rems, as Hameroff and Pen-rose suggest. But even if itdoes, the theory is silentabout how these processesmight give rise to consciousexperience. Indeed, the sameproblem arises with any the-ory of consciousness basedonly on physical processing.

The trouble is that physi-cal theories are best suited to

The Puzzle of Conscious Experience34 Mysteries of the Mind

BLOOD FLOW variations in the vi-sual cortex demonstrate how a sub-ject’s brain responds to a pattern be-ing viewed. The colors in this imageshow the cortical activity corre-

UN

IVER

SITY

OF

PITT

SBU

RGH

Continued from page 32

Copyright 1995 Scientific American, Inc.

Page 27: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

explaining why systems have a certainphysical structure and how they per-form various functions. Most problemsin science have this form; to explainlife, for example, we need to describehow a physical system can reproduce,adapt and metabolize. But conscious-ness is a different sort of problem en-tirely, as it goes beyond the scientific ex-planation of structure and function.

Of course, neuroscience is not irrele-vant to the study of consciousness. Forone, it may be able to reveal the natureof the neural correlate of conscious-ness—the brain processes most directlyassociated with conscious experience. Itmay even give a detailed correspondencebetween specific processes in the brainand related components of experience.But until we know why these processesgive rise to conscious experience at all,we will not have crossed what philos-opher Joseph Levine has called the ex-planatory gap between physical process-es and consciousness. Making that leapwill demand a new kind of theory.

In searching for an alternative, a keyobservation is that not all entities in sci-ence are explained in terms of more ba-sic entities. In physics, for example,space-time, mass and charge (among oth-er things) are regarded as fundamentalfeatures of the world, as they are not re-ducible to anything simpler.Despite this irreducibility,detailed and useful theoriesrelate these entities to oneanother in terms of funda-mental laws. Together thesefeatures and laws explain agreat variety of complex andsubtle phenomena.

It is widely believed thatphysics provides a completecatalogue of the universe’sfundamental features andlaws. As physicist StevenWeinberg puts it in his 1992book Dreams of a FinalTheory, the goal of physics isa “theory of everything”from which all there is toknow about the universe canbe derived. But Weinberg

concedes that there is a problem withconsciousness. Despite the power ofphysical theory, the existence of con-sciousness does not seem to be derivablefrom physical laws. He defends physicsby arguing that it might eventually ex-plain what he calls the objective corre-lates of consciousness (that is, the neu-ral correlates), but of course to do thisis not to explain consciousness itself. Ifthe existence of consciousness cannotbe derived from physical laws, a theoryof physics is not a true theory of every-thing. So a final theory must contain anadditional fundamental component.

A True Theory of Everything

Toward this end, I propose that con-scious experience be considered a

fundamental feature, irreducible to any-thing more basic. The idea may seemstrange at first, but consistency seems todemand it. In the 19th century it turnedout that electromagnetic phenomenacould not be explained in terms of pre-viously known principles. As a conse-quence, scientists introduced electro-magnetic charge as a new fundamentalentity and studied the associated funda-mental laws. Similar reasoning shouldbe applied to consciousness. If existingfundamental theories cannot encom-

pass it, then something new is required.Where there is a fundamental proper-

ty, there are fundamental laws. In thiscase, the laws must relate experience toelements of physical theory. These lawswill almost certainly not interfere withthose of the physical world; it seems thatthe latter form a closed system in theirown right. Rather the laws will serve asa bridge, specifying how experience de-pends on underlying physical processes.It is this bridge that will cross the ex-planatory gap.

Thus, a complete theory will have twocomponents: physical laws, telling usabout the behavior of physical systemsfrom the infinitesimal to the cosmologi-cal, and what we might call psychophys-ical laws, telling us how some of thosesystems are associated with consciousexperience. These two components willconstitute a true theory of everything.

Supposing for the moment that theyexist, how might we uncover such psy-chophysical laws? The greatest hin-drance in this pursuit will be a lack ofdata. As I have described it, conscious-ness is subjective, so there is no directway to monitor it in others. But this dif-ficulty is an obstacle, not a dead end.For a start, each one of us has access toour own experiences, a rich trove thatcan be used to formulate theories. We

can also plausibly rely on in-direct information, such assubjects’ descriptions of theirexperiences. Philosophicalarguments and thought ex-periments also have a role toplay. Such methods have lim-itations, but they give us morethan enough to get started.

These theories will not beconclusively testable, so theywill inevitably be more spec-ulative than those of moreconventional scientific disci-plines. Nevertheless, there isno reason they should not bestrongly constrained to ac-count accurately for our ownfirst-person experiences, aswell as the evidence fromsubjects’ reports. If we find atheory that fits the data bet-ter than any other theory ofequal simplicity, we will havegood reason to accept it.Right now we do not haveeven a single theory that fitsthe data, so worries abouttestability are premature.

We might start by looking

The Puzzle of Conscious Experience Mysteries of the Mind 35

sponding to the subject’s view of ei-ther the vertical or horizontal half ofthe pattern. The experiment may il-luminate a neural correlate of visualconsciousness.

Copyright 1995 Scientific American, Inc.

Page 28: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

for high-level bridging laws,connecting physical process-es to experience at an ev-eryday level. The basic con-tour of such a law might begleaned from the observa-tion that when we are con-scious of something, we aregenerally able to act on itand speak about it—which are objective,physical functions. Conversely, whensome information is directly availablefor action and speech, it is generally con-scious. Thus, consciousness correlateswell with what we might call “aware-ness”: the process by which informa-tion in the brain is made globally avail-able to motor processes such as speechand bodily action.

Objective Awareness

The notion may seem trivial. But asdefined here, awareness is objective

and physical, whereas consciousness isnot. Some refinements to the definitionof awareness are needed, in order to ex-tend the concept to animals and infants,which cannot speak. But at least in fa-miliar cases, it is possible to see therough outlines of a psychophysical law:where there is awareness, there is con-sciousness, and vice versa.

To take this line of reasoning a stepfurther, consider the structure presentin the conscious experience. The experi-ence of a field of vision, for example, isa constantly changing mosaic of colors,shapes and patterns and as such has adetailed geometric structure. The factthat we can describe this structure, reachout in the direction of many of its com-

ponents and perform other actionsthat depend on it suggests that the

structure corresponds directly tothat of the information madeavailable in the brain throughthe neural processes of objec-tive awareness.

Similarly, our experiencesof color have an intrinsicthree-dimensional structurethat is mirrored in the struc-ture of information processes

in the brain’s visual cortex. Thisstructure is illustrated in the col-

or wheels and charts used by art-ists. Colors are arranged in a sys-

tematic pattern—red to green on oneaxis, blue to yellow on an-other, and black to white ona third. Colors that are closeto one another on a colorwheel are experienced assimilar [see illustration on thispage]. It is extremely likelythat they also correspond tosimilar perceptual represen-tations in the brain, as part ofa system of complex three-dimensional coding among

neurons that is not yet fully understood.We can recast the underlying concept asa principle of structural coherence: thestructure of conscious experience ismirrored by the structure of informa-tion in awareness, and vice versa.

Another candidate for a psychophys-ical law is a principle of organizationalinvariance. It holds that physical systemswith the same abstract organization willgive rise to the same kind of consciousexperience, no matter what they aremade of. For example, if the precise in-teractions between our neurons could beduplicated with silicon chips, the sameconscious experience would arise. Theidea is somewhat controversial, but I be-lieve it is strongly supported by thoughtexperiments describing the gradual re-placement of neurons by silicon chips[see box on next page]. The remarkableimplication is that consciousness mightsomeday be achieved in machines.

The ultimate goal of a theory of con-sciousness is a simple and elegant set offundamental laws, analogous to thefundamental laws of physics. The prin-ciples described above are unlikely tobe fundamental, however. Rather theyseem to be high-level psychophysicallaws, analogous to macroscopic princi-ples in physics such as those of thermo-dynamics or kinematics. What might the

underlying fundamental laws be? Noone knows, but I don’t mind speculating.

I suggest that the primary psychophys-ical laws may centrally involve the con-cept of information. The abstract no-tion of information, as put forward inthe 1940s by Claude E. Shannon of theMassachusetts Institute of Technology,is that of a set of separate states with abasic structure of similarities and differ-ences between them. We can think of a10-bit binary code as an informationstate, for example. Such informationstates can be embodied in the physicalworld. This happens whenever theycorrespond to physical states (voltages,say), and when differences betweenthem can be transmitted along somepathway, such as a telephone line.

Information: Physical and Experiential

We can also find information em-bodied in conscious experience.

The pattern of color patches in a visualfield, for example, can be seen as analo-gous to that of the pixels covering a dis-play screen. Intriguingly, it turns outthat we find the same information statesembedded in conscious experience andin underlying physical processes in thebrain. The three-dimensional encodingof color spaces, for example, suggeststhat the information state in a color ex-perience corresponds directly to an in-formation state in the brain. Thus, wemight even regard the two states as dis-tinct aspects of a single informationstate, which is simultaneously embod-ied in both physical processing and con-scious experience.

A natural hypothesis ensues. Perhapsinformation, or at least some informa-tion, has two basic aspects: a physicalone and an experiential one. This hy-pothesis has the status of a fundamen-tal principle that might underlie the re-lation between physical processes andexperience. Wherever we find consciousexperience, it exists as one aspect of aninformation state, the other aspect ofwhich is embedded in a physical pro-cess in the brain. This proposal needs tobe fleshed out to make a satisfying the-ory. But it fits nicely with the principlesmentioned earlier—systems with thesame organization will embody the sameinformation, for example—and it couldexplain numerous features of our con-scious experience.

The idea is at least compatible withseveral others, such as physicist John A.

The Puzzle of Conscious Experience36 Mysteries of the Mind

COLOR WHEELarranges hues sothat ones experi-

enced as similar areclosest. Nearby col-ors also correspondto similar perceptu-al representations

in the brain.

Copyright 1995 Scientific American, Inc.

Page 29: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Mysteries of the Mind 37The Puzzle of Conscious Experience

Wheeler’s suggestion that informationis fundamental to the physics of the uni-verse. The laws of physics might ulti-mately be cast in informational terms,in which case we would have a satisfyingcongruence between the constructs inboth physical and psychophysical laws.It may even be that a theory of physicsand a theory of consciousness couldeventually be consolidated into a singlegrander theory of information.

Is Experience Ubiquitous?

Apotential problem is posed by the ubiquity of information. Even a

thermostat embodies some information,for example, but is it conscious? Thereare at least two possible responses. First,we could constrain the fundamentallaws so that only some information hasan experiential aspect, perhaps depend-ing on how it is physically processed.Second, we might bite the bullet and al-

low that all information has an experi-ential aspect—where there is complexinformation processing, there is com-plex experience, and where there is sim-ple information processing, there is sim-ple experience. If this is so, then even athermostat might have experiences, al-though they would be much simplerthan even a basic color experience, andthere would certainly be no accompa-nying emotions or thoughts. This seemsodd at first, but if experience is trulyfundamental, we might expect it to bewidespread. In any case, the choice be-tween these alternatives should dependon which can be integrated into themost powerful theory.

Of course, such ideas may be allwrong. On the other hand, they mightevolve into a more powerful proposalthat predicts the precise structure of ourconscious experience from physical pro-cesses in our brains. If this project suc-ceeds, we will have good reason to ac-

cept the theory. If it fails, other avenueswill be pursued, and alternative funda-mental theories may be developed. Inthis way, we may one day resolve thegreatest mystery of the mind.

Further Reading

Absent Qualia, Fading Qualia, Danc-ing Qualia. David J. Chalmers in Con-scious Experience. Edited by ThomasMetzinger. Ferdinand Schöningh, 1995.

Explaining Consciousness: The “HardProblem.” Special issue of Journal ofConsciousness Studies, Vol. 2, No. 3; Au-tumn 1995.

The Conscious Mind: In Search of aFundamental Theory. David J. Chal-mers. Oxford University Press, 1996.

The Nature of Consciousness: Philo-sophical Debates. Edited by Ned Block,Owen Flanagan and Güven Güzeldere.MIT Press (in press).

Whether consciousness could arise in a complex, synthet-ic system is a question many people find intrinsically fas-

cinating. Although it may be decades or even centuries beforesuch a system is built, a simple thought experiment offers strongevidence that an artificial brain, if organized appropriately,would indeed have precisely the same kind of conscious experi-ences as a human being.

Consider a silicon-based system in which the chips are orga-nized and function in the same way asthe neurons in your brain. That is, eachchip in the silicon system does exactlywhat its natural analogue does and isinterconnected to surrounding ele-ments in precisely the same way. Thus,the behavior exhibited by the artificialsystem will be exactly the same asyours. The crucial question is: Will it beconscious in the same way that you are?

Let us assume, for the purpose of ar-gument, that it would not be. (Here weuse a reasoning technique known as re-ductio ad absurdum, in which the op-posite hypothesis is assumed and thenshown to lead to an untenable conclu-sion.) That is, it has either different expe-riences—an experience of blue, say,when you are seeing red—or no experience at all. We will con-sider the first case; the reasoning proceeds similarly in both cases.

Because chips and neurons have the same function, they areinterchangeable, with the proper interfacing. Chips thereforecan replace neurons, producing a continuum of cases in which asuccessively larger proportion of neurons are replaced by chips.Along this continuum, the conscious experience of the system

will also change. For example, we might replace all the neuronsin your visual cortex with an identically organized version madeof silicon. The resulting brain, with an artificial visual cortex, willhave a different conscious experience from the original: whereyou had previously seen red, you may now experience purple (orperhaps a faded pink, in the case where the wholly silicon sys-tem has no experience at all).

Both visual cortices are then attached to your brain, through atwo-position switch. With the switch inone mode, you use the natural visualcortex; in the other, the artificial cortexis activated. When the switch is flipped,your experience changes from red topurple, or vice versa. When the switch isflipped repeatedly, your experiences“dance” between the two different con-scious states (red and purple), known asqualia.

Because your brain’s organization hasnot changed, however, there can be nobehavioral change when the switch isthrown. Therefore, when asked aboutwhat you are seeing, you will say thatnothing has changed. You will hold thatyou are seeing red and have seen noth-ing but red—even though the two col-

ors are dancing before your eyes. This conclusion is so unreason-able that it is best taken as a reductio ad absurdum of the origi-nal assumption—that an artificial system with identicalorganization and functioning has a different conscious experi-ence from that of a neural brain. Retraction of the assumptionestablishes the opposite: that systems with the same organiza-tion have the same conscious experience. —D.J.C.

IN THOUGHT EXPERIMENT, an apple might flash from red to blue.

Dancing Qualia in a Synthetic Brain

TOM

DR

APE

R D

ESIG

N

SA

Copyright 1995 Scientific American, Inc.

Page 30: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Reprinted from the May 1996 issue

Compared with misery, hap-piness is relatively unexploredterrain for social scientists.Between 1967 and 1994,

46,380 articles indexed in PsychologicalAbstracts mentioned depression, 36,851anxiety, and 5,099 anger. Only 2,389 arti-cles spoke of happiness, 2,340 life satisfac-tion, and 405 joy.

Recently we and other researchers havebegun a systematic study of happiness. Dur-ing the past two decades, dozens of investi-gators throughout the world have askedseveral hundred thousand representativelysampled people to reflect on their happinessand satisfaction with life—or what psychol-ogists call “subjective well-being.” In theU.S. the National Opinion Research Centerat the University of Chicago has surveyed arepresentative sample of roughly 1,500 peo-ple a year since 1957; the Institute for So-cial Research at the University of Michiganhas carried out similar studies on a less reg-ular basis, as has the Gallup Organization.Government-funded efforts have also probedthe moods of European citizens.

We have uncovered some surprising find-ings. People are happier than one might ex-pect, and happiness does not appear to de-pend significantly on external circumstances.Although viewing life as a tragedy has a longand honorable history, the responses of ran-dom samples of people around the worldabout their happiness paints a much rosierpicture.

In the University of Chicago surveys, threein 10 Americans say they are very happy, forexample. Only one in 10 chooses the mostnegative description, “not too happy.” Themajority describe themselves as “pretty hap-py.” (The few exceptions to global reportsof reasonable happiness include hospitalizedalcoholics, new inmates, new psychotherapyclients, South African blacks during apart-heid, and students living under conditionsof economic and political oppression.)

How can social scientists measure some-thing as hard to pin down as happiness?Most researchers simply ask people to re-port their feelings of happiness or unhappi-ness and to assess how satisfying their livesare. Such self-reported well-being is moder-ately consistent over years of retesting. Fur-thermore, those who say they are happy andsatisfied seem happy to their close friendsand family members and to a psychologist-interviewer. Their daily mood ratings revealmore positive emotions, and they smile morethan those who call themselves unhappy.Self-reported happiness also predicts otherindicators of well-being. Comparedwith the depressed,happy people areless self-focused, lesshostile and abusive,and less susceptibleto disease.

We have foundthat the even distri-bution of happinesscuts across almost alldemographic classi-fications of age, eco-nomic class, race andeducational level. Inaddition, almost allresearch strategies for assessing subjectivewell-being—including those that samplepeople’s experience by polling them at ran-dom times with beepers—turn up similarfindings.

Interviews with representative samples ofpeople of all ages, for example, reveal thatno time of life is notably happier or unhap-pier. Similarly, men and women are equallylikely to declare themselves “very happy”and “satisfied” with life, according to a sta-tistical digest of 146 studies by Marilyn J.Haring, William Stock and Morris A. Okun,all then at Arizona State University. AlexMichalos of the University of Northern

The Pursuit of HappinessNew research uncovers some anti-intuitive insightsinto how many people are happy—and why

by David G. Myers and Ed Diener

DIM

ITRY

SC

HID

LOV

SKY

The Authors

DAVID G. MYERS and EDDIENER have been studyinghappiness for more than 10years. Myers is professor ofpsychology at Hope Collegein Michigan and author ofThe Pursuit of Happiness:Who Is Happy and Why(William Morrow, 1992). Hewon the Gordon Allport Prizefor his studies of group influ-ence. Diener is professor ofpsychology at the Universityof Illinois and investigates thedefinition and measurementof subjective well-being. Hiscurrent work focuses on cul-tural differences in subjectivewell-being and on adaptationto life events.

People who saythey are happyand satisfiedseem happy totheir family andfriends. Happypeople also areless susceptibleto disease.

40 Mysteries of the Mind

Wealth, it turns out, is not a goodpredictor of happiness.

Copyright 1997 Scientific American, Inc.

Page 31: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

British Columbia and Ronald Inglehartof the University of Michigan, summa-rizing newer surveys of 18,000 univer-sity students in 39 countries and170,000 adults in 16 countries, corrob-orate these findings.

Knowing someone’s ethnicity alsogives little clue to subjective well-being.African-Americans are only slightly lesslikely than European-Americans to feelvery happy. The National Institute ofMental Health found that rates of de-pression and alcoholism among blacksand whites are roughly equal. Socialpsychologists Jennifer K. Crocker of theUniversity of Michigan and BrendaMajor of the University of California atSanta Barbara assert that people in dis-advantaged groups maintain self-esteemby valuing things at which they excel,by making comparisons within theirown groups and by blaming problemson external sources such as prejudice.

What Money Can’t Buy

Wealth is also a poor predictor ofhappiness. People have not be-

come happier over time as their cultureshave become more affluent. Even thoughAmericans earn twice as much in today’sdollars as they did in 1957, the propor-tion of those telling surveyors from theNational Opinion Research Center thatthey are “very happy” has declined from35 to 29 percent.

Even very rich people—those surveyedamong Forbes magazine’s 100 wealthi-est Americans—are only slightly happi-er than the average American. Thosewhose income has increased over a 10-year period are not happier than thosewhose income is stagnant. Indeed, inmost nations the correlation betweenincome and happiness is negligible—

only in the poorest countries, such asBangladesh and India, is income a goodmeasure of emotional well-being.

Are people in rich countries happier,by and large, than people in not so richcountries? It appears in general that theyare, but the margin may be slim. In Por-tugal, for example, only one in 10 peo-ple reports being very happy, whereas inthe much more prosperous Netherlandsthe proportion of very happy people isfour in 10. Yet there are curious rever-sals in this correlation between nationalwealth and well-being—the Irish duringthe 1980s consistently reported greaterlife satisfaction than the wealthier WestGermans. Furthermore, other factors,such as civil rights, literacy and duration

The Pursuit of Happiness42 Mysteries of the Mind

20% 46% 27% 4% 2% 1% 0%

NA

TIO

NA

LLY

REPR

ESEN

TATI

VE

SURV

EYS GLOBAL HAPPINESS AND SATISFACTION

SCALE OF SUBJECTIVE WELL-BEING

160

140

120

100

80

60

40

20

00 1 2 3 4 5 6 7 8 9 10

Researchers use various methods to survey people’s subjective sense of well-being. Some employ images (top), others use words (middle), but all ques-

tions essentially come down to asking people how they feel about their life. Differ-ent techniques yield remarkably similar results; we have collated data from almost1,000 surveys of 1.1 million people to arrive at a global estimate of reported subjec-tive well-being (bottom). —D.G.M. and E.D.

DIM

ITRY

SC

HID

LOV

SKY

“In most ways my life is close to my ideal.”

“The conditions of my life are excellent.”

“I am satisfied with my life.”

“So far I have gotten the important things

I want in life.”

“If I could live my life over,

I would change almost nothing.”Do you strongly disagree, disagree, slightly disagree,

neither agree nor disagree, slightly agree,

agree or strongly agree?

WHICH OF THESE FACES REPRESENTS THE WAY YOU FEEL ABOUT YOUR LIFE AS A WHOLE?

Probing for Happiness

Copyright 1997 Scientific American, Inc.

Page 32: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

of democratic government, all of whichalso promote reported life satisfaction,tend to go hand in hand with nationalwealth. As a result, it is impossible totell whether the happiness of people inwealthier nations is based on money oris a by-product of other felicities.

Habits of Happy People

Although happiness is not easy to predict from material circumstanc-

es, it seems consistent for those who haveit. In one National Institute on Agingstudy of 5,000 adults, the happiest peo-ple in 1973 were still relatively happy adecade later, despite changes in work,residence and family status.

In study after study, four traits char-acterize happy people. First, especiallyin individualistic Western cultures, theylike themselves. They have high self-es-teem and usually believe themselves tobe more ethical, more intelligent, lessprejudiced, better able to get along withothers, and healthier than the average

person. (Such findings bring to mindSigmund Freud’s joke about the manwho told his wife, “If one of us shoulddie, I think I would go live in Paris.”)

Second, happy people typically feelpersonal control. Those with little or nocontrol over their lives—such as prison-ers, nursing home patients, severely im-poverished groups or individuals, andcitizens of totalitarian regimes—sufferlower morale and worse health. Third,happy people are usually optimistic.Fourth, most happy people are extro-verted. Although one might expect thatintroverts would live more happily inthe serenity of their less stressed, con-templative lives, extroverts are happi-er—whether alone or with others.

The causal arrows for these correla-tions are uncertain. Does happinessmake people more outgoing, or are out-going people more likely to be happy,perhaps explaining why they marrysooner, get better jobs and make morefriends? If extrovert traits do indeedpredispose their carriers to happiness,

people might become happierby acting in certain ways. Inexperiments, people who feignhigh self-esteem report feelingmore positively about them-selves, for example.

Whatever the reason, theclose personal relationshipsthat characterize happy livesare also correlated with health.Compared with loners, thosewho can name several intimatefriends are healthier and lesslikely to die prematurely. Formore than nine out of 10 peo-ple, the most significant alter-native to aloneness is mar-riage. Although broken mari-tal relationships can causemuch misery, a good marriageapparently is a strong sourceof support. During the 1970sand 1980s, 39 percent of mar-ried adults told the NationalOpinion Research Center theywere “very happy,” as com-pared with 24 percent of thosewho had never married. Inother surveys, only 12 percentof those who had divorcedperceived themselves to be“very happy.” The happinessgap between the married andthe never married was similarfor women and men.

Religiously active peoplealso report greater happiness.

One Gallup survey found that highlyreligious people were twice as likely asthose lowest in spiritual commitment todeclare themselves very happy. Othersurveys, including a collaborative studyof 166,000 people in 16 nations, havefound that reported happiness and lifesatisfaction rise with strength of reli-gious affiliation and frequency of atten-dance at worship services. Some re-searchers believe that religious affilia-tion entails greater social support andhopefulness.

Students of happiness are now begin-ning to examine happy people’s exercisepatterns, worldviews and goals. It is pos-sible that some of the patterns discov-ered in the research may offer clues fortransforming circumstances and behav-iors that work against well-being intoones that promote it. Ultimately, then,the scientific study of happiness couldhelp us understand how to build a worldthat enhances human well-being and toaid people in getting the most satisfac-tion from their circumstances.

The Pursuit of Happiness Mysteries of the Mind 43

PERC

ENT

OF

RESP

ON

DEN

TS

FEMALESMALES

SATISFIED

VERY HAPPY

100

80

60

40

20

0

PERSONALINCOME

SATISFACTION

AV

ERA

GE

INC

OM

E A

FTER

TA

XES

IN 1

990

DO

LLA

RS

PERC

ENTA

GE

VER

Y H

APP

Y

YEAR1930 1950 1970 1990

16,000

14,000

12,000

10,000

8,000

6,000

4,000

2,000

0

1009080706050403020100

PERC

ENTA

GE

VER

Y H

APP

Y

YEAR

MARRIED

NEVER MARRIED

50

40

30

20

10

01973 1977 1981 1985 1989

HAPPINESS APPEARS CONSISTENT across many different sectors of the population. Bothsexes report roughly the same satisfaction with life (top left), as do various age groups (topright). Among the few consistent differentials is that between married and never-married people(bottom left); other data indicate that divorced people are less happy than either of these twogroups. Personal satisfaction has remained relatively constant over at least several decades in theU.S., even as national income has increased (bottom right).

DIM

ITRY

SC

HID

LOV

SKY

AGE

PERC

ENTA

GE

SATI

SFIE

D W

ITH

LIF

E 100

80

60

40

20

015–24 25–35 35–44 45–54 55–64 65+

SOURCE: Top graphs: Based on data reported by RonaldInglehart in Culture Shift in Advanced Industrial Society,Princeton University Press, 1989

SOURCE: Bottom graphs: National Opinion Research Cen-ter, University of Chicago

SA

Copyright 1997 Scientific American, Inc.

Page 33: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Reprinted from the February 1995 issue

Men have called me mad,” wrote Edgar Allan Poe, “but the question isnot yet settled, whether madness is or is not the loftiest intelligence—

whether much that is glorious—whether all that is profound—doesnot spring from disease of thought—from moods of mind exalted at

the expense of the general intellect.”Many people have long shared Poe’s suspicion that genius and insanity are entwined. In-

deed, history holds countless examples of “that fine madness.” Scores of influential 18th-and 19th-century poets, notably William Blake, Lord Byron and Alfred, Lord Tennyson,wrote about the extreme mood swings they endured. Modern American poets John Berry-man, Randall Jarrell, Robert Lowell, Sylvia Plath, Theodore Roethke, Delmore Schwartzand Anne Sexton were all hospitalized for either mania or depression during their lives.And many painters and composers, among them Vincent van Gogh, Georgia O’Keeffe,Charles Mingus and Robert Schumann, have been similarly afflicted.

Judging by current diagnostic criteria, it seems that most of these artists—and many oth-ers besides—suffered from one of the major mood disorders, namely, manic-depressive ill-ness or major depression. Both are fairly common, very treatable and yet frequently lethaldiseases. Major depression induces intense melancholic spells, whereas manic-depression,

Manic-Depressive Illness and CreativityDoes some fine madness plague great artists? Several studies now show that creativity and mood disorders are linked

by Kay Redfield Jamison

44 Mysteries of the Mind

The Author

KAY REDFIELD JAMISONis professor of psychiatry atthe Johns Hopkins UniversitySchool of Medicine. Shewrote Touched with Fire:Manic-Depressive Illness andthe Artistic Temperament andco-authored the medical textManic-Depressive Illness.Jamison is a member of theNational Advisory Councilfor Human Genome Re-search and clinical director ofthe Dana Consortium on theGenetic Basis of Manic-De-pressive Illness. She has alsowritten and produced a seriesof public television specialsabout manic-depressive ill-ness and the arts.

Van Gogh painted flowers whilein the asylum at Saint-Rémy.

Vincentvan Gogh

Tennessee Williams

AP/

WID

E W

ORL

D P

HO

TOS

Copyright 1997 Scientific American, Inc.

Page 34: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

a strongly genetic disease, pitches pa-tients repeatedly from depressed to hy-peractive and euphoric, or intensely ir-ritable, states. In its milder form,termed cyclothymia, manic-depressioncauses pronounced but not totally de-bilitating changes in mood, behavior,sleep, thought patterns and energy lev-els. Advanced cases are marked by dra-matic, cyclic shifts.

Could such disruptive diseases con-vey certain creative advantages? Manypeople find that proposition counterin-

tuitive. Most manic-depressives do notpossess extraordinary imagination, andmost accomplished artists do not sufferfrom recurring mood swings. To assume,then, that such diseases usually promoteartistic talent wrongly reinforces simplis-tic notions of the “mad genius.” Worseyet, such a generalization trivializes avery serious medical condition and, tosome degree, discredits individuality inthe arts as well. It would be wrong tolabel anyone who is unusually accom-plished, energetic, intense, moody or

eccentric as manic-depressive. All thesame, recent studies indicate that a highnumber of established artists—far morethan could be expected by chance—

meet the diagnostic criteria for manic-depression or major depression given inthe fourth edition of the Diagnostic andStatistical Manual of Mental Disorders(DSM-IV). In fact, it seems that thesediseases can sometimes enhance or oth-erwise contribute to creativity in somepeople.

By virtue of their prevalence alone, itis clear that mood disorders do not nec-essarily breed genius. Indeed, 1 percentof the general population suffer frommanic-depression, also called bipolardisorder, and 5 percent from a majordepression, or unipolar disorder, duringtheir lifetime. Depression affects twiceas many women as men and most of-ten, but not always, strikes later in life.Bipolar disorder afflicts equal numbersof women and men, and more than athird of all cases surface before age 20.Some 60 to 80 percent of all adoles-

ARTISTS, writers and composers shownon these pages all most likely sufferedfrom manic-depressive illness or majordepressive illness, according to their let-ters and journals, medical records and ac-counts by their families and friends. Re-cent studies indicate that the tempera-ments and cognitive styles associated withmood disorders can in fact enhance cre-ativity in some individuals.

Charles Mingus

Ezra Pound

Anne Sexton

BETT

MA

NN

ARC

HIV

E

UPI

/BET

TMA

NN

HEN

RI C

AR

TIER

-BRE

SSO

N M

agnu

m

AP/

WID

E W

ORL

D P

HO

TOS

Mysteries of the Mind 45

Copyright 1997 Scientific American, Inc.

Page 35: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

cents and adults who commit suicidehave a history of bipolar or unipolar ill-ness. Before the late 1970s, when thedrug lithium first became widely avail-able, one person in five with manic-de-pression committed suicide.

Major depression in both unipolarand bipolar disorders manifests itselfthrough apathy, lethargy, hopelessness,sleep disturbances, slowed physicalmovements and thinking, impairedmemory and concentration, and a lossof pleasure in typically enjoyable events.The diagnostic criteria also include sui-cidal thinking, self-blame and inappro-priate guilt. To distinguish clinical de-

pression from normal periods of unhap-piness, the common guidelines furtherrequire that these symptoms persist fora minimum of two to four weeks andalso that they significantly interfere witha person’s everyday functioning.

Mood Elevation

During episodes of mania or hypo-mania (mild mania), bipolar pa-

tients experience symptoms that are inmany ways the opposite of those asso-ciated with depression. Their mood andself-esteem are elevated. They sleep lessand have abundant energy; their pro-

ductivity increases. Manics frequentlybecome paranoid and irritable. More-over, their speech is often rapid, excit-able and intrusive, and their thoughtsmove quickly and fluidly from one topicto another. They usually hold tremen-dous conviction about the correctnessand importance of their own ideas aswell. This grandiosity can contribute topoor judgment and impulsive behavior.

Hypomanics and manics generallyhave chaotic personal and professionalrelationships. They may spend largesums of money, drive recklessly or pur-sue questionable business ventures orsexual liaisons. In some cases, manics

Manic-Depressive Illness and Creativity46 Mysteries of the Mind

Alfred, Lord Tennyson (right), who experienced recurrent, debilitating depressions and

probable hypomanic spells, often expressed fearthat he might inherit the madness, or “taint ofblood,” in his family. His father, grandfather,two of his great-grandfathers as well as fiveof his seven brothers suffered from insanity,melancholia, uncontrollable rage or what istoday known as manic-depressive illness.His brother Edward was confined to an asy-lum for nearly 60 years before he died frommanic exhaustion. Lionel Tennyson, one of Al-fred’s two sons, displayed a mercurial tempera-ment, as did one of his three grandsons.

Modern medicine has confirmed that manic-de-pression and creativity tend to run in certain fami-

lies. Studies of twins provide strong evidencefor the heritability of manic-depressive illness.

If an identical twin has manic-depressive ill-ness, the other twin typically has a 70 to 100percent chance of also having the disease; ifthe other twin is fraternal, the chances areconsiderably lower (approximately 20 per-

cent). A review of pairs of identical twinsreared apart from birth—in which at least one

had been diagnosed as manic-depressive—found that in two thirds or more of the sets, the ill-

ness was present in both twins. —K. R. J.

ELIZABETHb. 1776

Recurrent bouts of depression

GEORGEDied in infancy,

1806

FREDERICK1807–1898

Irritability; eccentric; violent temper and volatile; obsessed with spiritualism

Recurrent depressive illness

ELIZABETH FYTCHE1781–1865

“Easy-going” and“sweet tempered”

MARY1810 –1884

“...of a wild sortof countenance”;

obsessed with spiritualism

CECILIA1817–1909

“Mental disturbanceand depression”;

eccentric

EMILY1811–1889

SOURCE: Adapted from Touched with Fire: Manic-Depressive Illness and the Artistic Temperament; based on biographies, autobiographical writings and letters.

GEORGE CLAYTON TENNYSON1778–1831

Vacillating moods “between frenzy and lethargy”; spendthrift; alcoholic; “fits”; insanity

MARYb. 1777

“Ferocious pessimism”; constant quarreling

and gloominess

CHARLES (D'EYNCOURT)1784–1861

“Inherited his father’s instabilityand fretfulness”; spendthrift

tendencies; expansive, grandiose activities and interests

Manic-depressive illness

ALFRED1809–1892

Recurrent depression that re-quired treatment; trances, possibly

epileptic but not thought so by physician; possibly transient

hypomanic episodes; “dwelling in an element of gloom”

ARTHUR1814–1899

“Suffered much from depression”;

one year in Crichton

Institutionfor the Insane

CHARLES1808 –1879

Addicted to laudanum;“complete nervous break-

down”; had to be segregated from outside world; extreme

mood swings and “recurrent fits of

psychopathic depression”

EDWARD1813–1890

Confined in insane asylum for almost 60

years; severemelancholia; death

from manic exhaustion

SEPTIMUS1815 –1866

“Suffered from ner-vous depression”; fre-quent treatments for

melancholia; “the most morbid

of all the Tennysons”

HORATIO1819–1899

“Strange personality was legendary”; “rather unused

to this planet”; per-ceived himself as vulner-

able to the “weakness of the Tennysonian temperament”

Rage, unstable moods and/or insanity

MATILDA1816–1913

“Some mental derangement,”occasionally attributed to

childhood accident; religiousobsessions; “did not entirelyescape the black-blooded-

ness of the Tennysons”

LISA

BU

RNET

T

The Tainted Blood of the Tennysons

CO

RBIS

-BET

TMA

NN

Copyright 1997 Scientific American, Inc.

Page 36: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

suffer from violent agitation and delu-sional thoughts as well as visual andauditory hallucinations.

Rates of Mood Disorders

For years, scientists have documentedsome kind of connection between

mania, depression and creative output.In the late 19th and early 20th centuries,researchers turned to accounts of mooddisorders written by prominent artists,their physicians and friends. Althoughlargely anecdotal, this work stronglysuggested that renowned writers, artistsand composers—and their first-degreerelatives—were far more likely to expe-rience mood disorders and to commitsuicide than was the general population.During the past 20 years, more system-atic studies of artistic populations haveconfirmed these findings [see illustra-tion below]. Diagnostic and psycholog-ical analyses of living writers and artistscan give quite meaningful estimates ofthe rates and types of psychopathologythey experience.

In the 1970s Nancy C. Andreasen ofthe University of Iowa completed the firstof these rigorous studies, which madeuse of structured interviews, matchedcontrol groups and strict diagnostic cri-teria. She examined 30 creative writersand found an extraordinarily high oc-currence of mood disorders and alco-holism among them. Eighty percent hadexperienced at least one episode of ma-jor depression, hypomania or mania; 43

percent reported a history of hypoma-nia or mania. Also, the relatives of thesewriters, compared with the relatives ofthe control subjects, generally performedmore creative work and more often hada mood disorder.

A few years later, while on sabbaticalin England from the University of Cali-fornia at Los Angeles, I began a studyof 47 distinguished British writers andvisual artists. To select the group as bestI could for creativity, I purposefullychose painters and sculptors who wereRoyal Academicians or Associates ofthe Royal Academy. All the playwrightshad won the New York Drama CriticsAward or the Evening Standard Drama(London Critics) Award, or both. Halfof the poets were already represented inthe Oxford Book of Twentieth CenturyEnglish Verse. I found that 38 percentof these artists and writers had in factbeen previously treated for a mood dis-order; three fourths of those treated hadrequired medication or hospitalization,or both. And half of the poets—thelargest fraction from any one group—

had needed such extensive care.Hagop S. Akiskal of the University of

California at San Diego, also affiliatedwith the University of Tennessee atMemphis, and his wife, Kareen Akis-kal, subsequently interviewed 20 award-winning European writers, poets, paint-ers and sculptors. Some two thirds oftheir subjects exhibited recurrent cyclo-thymic or hypomanic tendencies, andhalf had at one time suffered from a ma-

jor depression. In collaboration with Da-vid H. Evans of the University of Mem-phis, the Akiskals noted the same trendsamong living blues musicians. More re-cently Stuart A. Montgomery and hiswife, Deirdre B. Montgomery, of St.Mary’s Hospital in London examined50 modern British poets. One fourth metcurrent diagnostic criteria for depres-sion or manic-depression; suicide wassix times more frequent in this commu-nity than in the general population.

Ruth L. Richards and her colleaguesat Harvard University set up a systemfor assessing the degree of original think-ing required to perform certain creativetasks. Then, rather than screening formood disorders among those alreadydeemed highly inventive, they attempt-ed to rate creativity in a sample of man-ic-depressive patients. Based on theirscale, they found that compared withindividuals having no personal or fami-ly history of psychiatric disorders, man-ic-depressive and cyclothymic patients(as well as their unaffected relatives)showed greater creativity.

Biographical studies of earlier genera-tions of artists and writers also showconsistently high rates of suicide, depres-sion and manic-depression—up to 18times the rate of suicide seen in the gen-eral population, eight to 10 times thatof depression and 10 to 20 times that ofmanic-depressive illness and its mildervariants. Joseph J. Schildkraut and hisco-workers at Harvard concluded thatapproximately half of the 15 20th-cen-

Manic-Depressive Illness and Creativity Mysteries of the Mind 47

EXPECTED RATEIN GENERAL POPULATION

ANDREASEN’S STUDYOF WRITERS (1987)

JAMISON’S STUDY OFARTISTS AND WRITERS (1989)

SCHILDKRAUT AND HIRSHFELD’SSTUDY OF ARTISTS (1990)

AKISKAL AND AKISKAL’S STUDY OFARTISTS AND WRITERS (UNPUBLISHED)

JAMISON’S STUDY OF BRITISH POETS BORN BETWEEN 1705 AND 1805 (1989)

MONTGOMERY AND MONTGOMERY’SSTUDY OF POETS (1993)

LUDWIG’S STUDY OF POETS (1992)

MAJOR DEPRESSIVE ILLNESS

0 10 20 30 40

PERCENT

50 60 70 80

MANIC-DEPRESSIVE ILLNESS

CYCLOTHYMIA

SUICIDE

LISA

BU

RNET

T

INCREASED RATES OF SUICIDE, depression and manic-de-pression among artists have been established by many separatestudies. These investigations show that artists experience up to

18 times the rate of suicide seen in the general population, eightto 10 times the rate of depression and 10 to 20 times the rate ofmanic-depression and its milder form, cyclothymia.

Copyright 1997 Scientific American, Inc.

Page 37: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

tury abstract-expressionist artists theystudied suffered from depressive or man-ic-depressive illness; the suicide rate inthis group was at least 13 times the cur-rent U.S. national rate.

In 1992 Arnold M. Ludwig of theUniversity of Kentucky published an ex-tensive biographical survey of 1,005 fa-mous 20th-century artists, writers andother professionals, some of whom hadbeen in treatment for a mood disorder.He discovered that the artists and writ-ers experienced two to three times therate of psychosis, suicide attempts, mooddisorders and substance abuse that com-parably successful people in business,science and public life did. The poets inthis sample had most often been manicor psychotic and hospitalized; they alsoproved to be some 18 times more likelyto commit suicide than is the generalpublic. In a comprehensive biographi-cal study of 36 major British poets bornbetween 1705 and 1805, I found simi-larly elevated rates of psychosis and se-vere psychopathology. These poets were30 times more likely to have had man-ic-depressive illness than were their con-temporaries, at least 20 times more like-ly to have been committed to an asylumand some five times more likely to havetaken their own life.

These corroborative studies have con-firmed that highly creative individualsexperience major mood disorders moreoften than do other groups in the gener-al population. But what does this meanfor their work? How does a psychiatricillness contribute to creative achieve-ment? First, the common features ofhypomania seem highly conducive tooriginal thinking; the diagnostic criteriafor this phase of the disorder include“sharpened and unusually creative think-ing and increased productivity.” Andaccumulating evidence suggests that the

cognitive styles associated with hypo-mania (expansive thought and grandi-ose moods) can lead to increased fluen-cy and frequency of thoughts.

Mania and Creativity

Studying the speech of hypomanicpatients has revealed that they tend

to rhyme and use other sound associa-tions, such as alliteration, far more of-ten than do unaffected individuals. Theyalso use idiosyncratic words nearly threetimes as often as do control subjects.Moreover, in specific drills, they can listsynonyms or form other word associa-tions much more rapidly than is consid-ered normal. It seems, then, that both thequantity and quality of thoughts buildduring hypomania. This speed increasemay range from a very mild quickeningto complete psychotic incoherence. It isnot yet clear what causes this qualita-tive change in mental processing. Nev-ertheless, this altered cognitive statemay well facilitate the formation ofunique ideas and associations.

People with manic-depressive illnessand those who are creatively accom-plished share certain noncognitive fea-

tures: the ability to function well on afew hours of sleep, the focus needed towork intensively, bold and restless atti-tudes, and an ability to experience aprofound depth and variety of emo-tions. The less dramatic daily aspects ofmanic-depression might also providecreative advantage to some individuals.The manic-depressive temperament is,in a biological sense, an alert, sensitivesystem that reacts strongly and swiftly.It responds to the world with a widerange of emotional, perceptual, intellec-tual, behavioral and energy changes. Ina sense, depression is a view of theworld through a dark glass, and maniais that seen through a kaleidoscope—of-ten brilliant but fractured.

Where depression questions, rumi-nates and hesitates, mania answers withvigor and certainty. The constant tran-sitions in and out of constricted and thenexpansive thoughts, subdued and thenviolent responses, grim and then ebul-lient moods, withdrawn and then out-going stances, cold and then fiery states—and the rapidity and fluidity of movesthrough such contrasting experiences—

can be painful and confusing. Ideally,though, such chaos in those able to

Manic-Depressive Illness and Creativity48 Mysteries of the Mind

ROBERT SCHUMANN’S MUSICAL WORKS, charted byyear and opus number (above), show a striking relation betweenhis mood states and his productivity. He composed the mostwhen hypomanic and the least when depressed. Both of Schu-

mann’s parents were clinically depressed, and two other first-de-gree relatives committed suicide. Schumann himself attemptedsuicide twice and died in an insane asylum. One of his sonsspent more than 30 years in a mental institution.

007 001 099008

124004003002

032021018016015

072060058056055

084080065063062

115081071068066

148147140139135122

143134133132131126123118114

120064054052038037

144130129125097096090089088087083

136128121119117113112111110109107105104103100

028023020019

010005

012006

050046

061059

022011009

017014013

047044041

146145141138137108106102101098095094093092091086085082079078076075074073070069067

142127077057053051049048045043042040039036035034033031030029027026025024

’301829 ’31 ’32

SUICIDEATTEMPT

’33 ’34 ’35 ’36 ’37 ’38 ’39 ’40 ’41 ’42 ’43 ’44 ’45 ’46 ’47 ’48 ’49 ’50 ’51 ’52 ’53 ’54 ’55 1856

HYPOMANICTHROUGHOUT 1849

SUICIDEATTEMPT

DIED IN ASYLUM(SELF-

STARVATION)HYPOMANIC

THROUGHOUT 1840

SEVERE DEPRESSIONTHROUGHOUT 1844

AD

APT

ED F

ROM

E. S

LATE

R A

ND

A. M

EYER

,195

9

AR

CH

IVE

PHO

TOS

Robert Schumann

Copyright 1997 Scientific American, Inc.

Page 38: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

transcend it or shape it to their will canprovide a familiarity with transitionsthat is probably useful in artistic endeav-ors. This vantage readily accepts ambi-guities and the counteracting forces innature.

Extreme changes in mood exaggeratethe normal tendency to have conflictingselves; the undulating, rhythmic andtransitional moods and cognitive chang-es so characteristic of manic-depressiveillness can blend or harness seeminglycontradictory moods, observations andperceptions. Ultimately, these fluxes andyokings may reflect truth in humanityand nature more accurately than coulda more fixed viewpoint. The “consis-tent attitude toward life” may not, asByron scholar Jerome J. McGann of theUniversity of Virginia points out, be asinsightful as an ability to live with, andportray, constant change.

The ethical and societal implicationsof the association between mood disor-ders and creativity are important butpoorly understood. Some treatment stra-tegies pay insufficient heed to the bene-fits manic-depressive illness can bestowon some individuals. Certainly mostmanic-depressives seek relief from thedisease, and lithium and anticonvulsantdrugs are very effective therapies formanias and depressions. Nevertheless,these drugs can dampen a person’s gen-eral intellect and limit his or her emo-tional and perceptual range. For thisreason, many manic-depressive patientsstop taking these medications.

Left untreated, however, manic-de-

pressive illness often worsens over time—

and no one is creative when severely de-pressed, psychotic or dead. The attacksof both mania and depression tend togrow more frequent and more severe.Without regular treatment the diseaseeventually becomes less responsive tomedication. In addition, bipolar and uni-polar patients frequently abuse mood-altering substances, such as alcohol andillicit drugs, which can cause secondarymedical and emotional burdens formanic-depressive and depressed patients.

The Goal of Treatment

The real task of imaginative, com-passionate and effective treatment,

therefore, is to give patients more mean-ingful choices than they are now afford-ed. Useful intervention must control theextremes of depression and psychosiswithout sacrificing crucial human emo-tions and experiences. Given time andincreasingly sophisticated research, psy-chiatrists will likely gain a better under-standing of the complex biological ba-sis for mood disorders. Eventually, thedevelopment of new drugs should makeit possible to treat manic-depressive in-dividuals so that those aspects of tem-perament and cognition that are essen-tial to the creative process remain intact.

The development of more specific andless problematic therapies should beswift once scientists find the gene, orgenes, responsible for the disease. Pre-natal tests and other diagnostic measuresmay then become available; these possi-

bilities raise a host of complicated ethi-cal issues. It would be irresponsible toromanticize such a painful, destructiveand all too often deadly disease. Hence,3 to 5 percent of the Human GenomeProject’s total budget (which is conser-vatively estimated at $3 billion) hasbeen set aside for studies of the social,ethical and legal implications of geneticresearch. It is hoped that these investi-gations will examine the troubling is-sues surrounding manic-depression andmajor depression at length. To help thosewho have manic-depressive illness, orwho are at risk for it, must be a majorpublic health priority.

Manic-Depressive Illness and Creativity Mysteries of the Mind 49

The Case of Vincent van Gogh

Many clinicians have reviewed the medi-cal and psychiatric problems of the

painter Vincent van Gogh posthumously, diag-nosing him with a range of disorders, includingepilepsy, schizophrenia, digitalis and absinthepoisoning, manic-depressive psychosis, acute in-termittent porphyria and Ménière’s disease.

Richard Jed Wyatt of the National Institute ofMental Health and I have argued in detail thatvan Gogh’s symptoms, the natural course of hisillness and his family psychiatric history stronglyindicate manic-depressive illness. The extent ofthe artist’s purported use of absinthe and con-vulsive behavior remains unclear; in any event,his psychiatric symptoms long predate any pos-sible history of seizures. It is possible that he suf-fered from both an epileptic disorder and manic-depressive illness. —K. R. J.

MET

RO

POLI

TAN

MU

SEU

M O

F A

RT,

GIF

T O

F A

DEL

E R.

LEV

Y, 1

958

Further Reading

Tennyson: The Unquiet Heart. R. B.Martin. Oxford University Press, 1980.

Creativity and Mental Illness: Preva-lence Rates in Writers and TheirFirst-Degree Relatives. Nancy C. An-dreasen in American Journal of Psychia-try, Vol. 144, No. 10, pages 1288-1292;October 1987.

Manic Depressive Illness. Frederick K.Goodwin and Kay R. Jamison. OxfordUniversity Press, 1990.

Creative Achievement and Psycho-pathology: Comparison among Pro-fessions. Arnold M. Ludwig in AmericanJournal of Psychiatry, Vol. 46, No. 3,pages 330–356; July 1992.

Touched with Fire: Manic-DepressiveIllness and the Artistic Temperament.Kay R. Jamison. Free Press/Macmillan,1993.

SA

Irises, 1889

Copyright 1997 Scientific American, Inc.

Page 39: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Mysteries of the Mind 53Reprinted from the June 1995 issue

Mental health workers have long noticed a preponder-

ance of women among the clinically depressed. Until

recently, though, it was unclear whether more wom-

en than men were ill or, instead, whether more wom-

en sought help. In fact, a mounting collection of studies has confirmed

that major depression is twice as common among women as it is

among men. “This is one of the most consistent findings we have ever

had,” says Myrna M. Weissman of Columbia University. Women also

seem more susceptible to milder melancholia and to seasonal affective

disorder (SAD).

Scientists searching for explanations are challenged by the fact that a

variety of cues prompt depression in different people. Sorting out which

factors might have a greater influence on women has not proved easy.

Both sexes stand an equal chance of inheriting major depression, so

genes are most likely not to blame. Yet hormones and sleep cycles—

which differ dramatically between the sexes—can alter mood. Also,

many workers have proposed that social discrimination might put wom-

en under high levels of stress, thereby doubly disposing them to major

depressive disorder.

In 1990 Weissman and Gerald L. Klerman of Cornell University

convened an international group to examine mood disorders. In the 10

nations reviewed so far, the team has found that among generations

reaching maturity after 1945, depression seems to be on the rise and

occurs at a younger age. Although overall incidence varies regionally,

“everywhere the rates of depression among women are about twice as

Depression’s Double Standard by Kristin Leutwyler, staff writer

PET SCANS reveal that during sadness the limbic system in women’s brains (left) becomes more metabolically active than

that area does in men’s brains (right).

Studies from 10 nationsreveal that the rates of depression among

women are twice as highas they are among men.

Do women have a biological bent for

depression, or are socialdouble standards the

major cause?

MA

RK S

. GEO

RGE

Nat

iona

l Ins

titut

e of

Men

tal H

ealth

Copyright 1997 Scientific American, Inc.

Page 40: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Depression’s Double Standard 54 Mysteries of the Mind

high as they are among men,” Weissman says. In contrast, lifetime rates

for manic-depressive illness do not differ according to sex or culture.

Meanwhile neurologists and endocrinologists suggest women may

well have a biological bent for depression. Mark S. George and his col-

leagues at the National Institute of Mental Health (NIMH) recently

studied which regions of the brain have greater blood flow during peri-

ods of sadness. They asked 10 men and 10 women to feel sad while they

took a positron emission tomographic (PET) scan. The participants then

judged how much sentiment they had mustered. George found that

men and women deemed themselves equally sad, but “the brain activi-

ty of the two groups looked very different.” Both sexes had equally ac-

tivated the left prefrontal cortex, but the women showed blood flows

in the anterior limbic system that were eight times greater. He has since

compared feelings of anger, anxiety and happiness, finding no discrep-

ancies anywhere near as large. Most significant, the regions of the brain

activated during sadness are two that malfunction during clinical de-

pression. George speculates that hyperactivity of the anterior limbic sys-

tem in women experiencing sadness could, over time, exhaust that re-

gion and lead to the hypoactivity seen there during clinical depression.

If he is right, the theory would explain the gender gap, at least in part.

Hormones and Depression

Others at the NIMH have more to add. “There are hints of gen-

der differences in responses both to seasonal patterns and to

day and night, or sleep patterns,” says Ellen Leibenluft. “Ei-

ther might put women at a greater risk for depression.” Thomas Wehr,

also at the NIMH, has found that during the winter, women increase

their nightly production of melatonin, a hormone whose levels are gov-

erned by the circadian pacemaker; women produce less melatonin dur-

ing summer nights. In contrast, nocturnal secretions of melatonin in

men are unchanging.

Another intriguing find is that without time cues such as daylight,

women seem more prone to sleep excessively. (Patients who sleep a

great deal during depression are, in fact, those who most often respond

to light therapy, Leibenluft says.) Further, sleep and activity cycles are

governed by the estrus cycle. Some conjecture that testosterone, which

promotes activity, protects men against depression, whereas estrogen

may lengthen the sleep phase in women. Gonadal steroids clearly regu-

late circadian rhythms in animals, and Leibenluft plans to see if they

hold similar sway in humans.

George, too, plans to consider the effects of estrogen on brain activa-

tion levels during bouts of sadness. Epidemiological data indicate that

hormones could play an important role in the onset of depression.

Equal numbers of boys and girls experience depression before puberty,

but shortly thereafter the rate among girls doubles.

The fact that many depressed patients are women of childbearing age

must be considered in research efforts, Leibenluft emphasizes. She

notes that although most psychotropic drugs are given to women (75

percent by some estimates), there is little information on how the men-

strual cycle might influence their efficacy. Moreover, no one knows

how menopause might alter the course of a mood disorder or its treat-

ment. Because one in five American women has a history of depression,

many of those who are going through menopause could be affected—

especially as they often pursue estrogen replacement therapy, some-

times on top of an antidepressant regime. Says Leibenluft: “It is re-

markable how little work has been done on this subject.”

INCIDENCE PER 100,000

5 TO 19 YEARS OF AGE

20 TO 64 YEARS OF AGE

65 YEARSAND OLDER

DEPRESSION BY SEX AND AGE

MEN

WOMEN

0 3,000

INCIDENCE PER 100,000

DEPRESSION BY REGION

0 2,500

AFRICA

AMERICAS

E. MEDITERRANEAN

EUROPE

SOUTHEAST ASIA

WESTERN PACIFIC

INCIDENCE PER 100,000

DEPRESSION BY LEVEL OF ECONOMIC DEVELOPMENT

0 2,500

DEVELOPED

DEVELOPING

LEASTDEVELOPED

ECONOMIESIN TRANSITION

INCIDENCE OF DEPRESSION among womenworldwide is nearly twice as great as amongmen (top) and less common in the young,

according to data gathered by the World HealthOrganization (WHO). Although incidence of

depression in the population varies by region(middle), women in every region have substan-tially higher rates of depression than men. Theincidence of depression is highest in economi-cally developed regions and lowest in the least

developed nations (bottom).

BRYA

N C

HRI

STIE

SA

SOURCE: World Health Organization

Copyright 1997 Scientific American, Inc.

Page 41: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Throughout history human beings have sought to understandthe meaning of dreams. The ancient Egyptians believed dreamspossessed oracular power—in the Bible, for example, Joseph’selucidation of Pharaoh’s dream averted seven years of famine.

Other cultures have interpreted dreams as inspirational, curative or alterna-tive reality. During the past century, scientists have offered conflicting psy-chological and neuroscientific explanations for dreams. In 1900, with thepublication of The Interpretation of Dreams, Sigmund Freud proposed thatdreams were the “royal road” to the unconscious; that they revealed in dis-guised form the deepest elements of an individual’s inner life.

More recently, in contrast, dreams have been characterized as meaningless,the result of random nerve cell activity. Dreaming has also been viewed as themeans by which the brain rids itself of unnecessary information—a processof “reverse learning,” or unlearning.

Based on recent findings in my own and other neuroscientific laboratories,I propose that dreams are indeed meaningful. Studies of the hippocampus (abrain structure crucial to memory), of rapid eye movement (REM) sleep andof a brain wave called theta rhythm suggest that dreaming reflects a pivotalaspect of the processing of memory. In particular, studies of theta rhythm insubprimate animals have provided an evolutionary clue to the meaning ofdreams. They appear to be the nightly record of a basic mammalian memo-ry process: the means by which animals form strategies for survival andevaluate current experience in light of those strategies. The existence of thisprocess may explain the meaning of dreams in human beings.

Stages of Sleep and Dreaming

The physiology of dreaming was first understood in 1953, when re-searchers characterized the human sleep cycle. They found that sleep in

humans is initiated by the hypnogogic state, a period of several minuteswhen thoughts consist of fragmented images or minidramas. The hyp-nogogic state is followed by slow-wave sleep, so called because at that timethe brain waves of the neocortex (the convoluted outer mantle of the brain)are low in frequency and large in amplitude. These signals are measured aselectroencephalographic (EEG) recordings.

Researchers also discovered that a night’s sleep is punctuated by periods inwhich the EEG readings are irregular in frequency and low in amplitude—

similar to those observed in awake individuals. These periods of mental ac-tivity are called REM sleep. Dreaming takes place solely during these peri-ods. While in REM sleep, motor neurons are inhibited, preventing the bodyfrom moving freely but allowing extremities to remain slightly active. Eyes

The Meaning of DreamsDreams may reflect a fundamental aspect of mammalianmemory processing. Crucial information acquired duringthe waking state may be reprocessed during sleep

by Jonathan Winson

58 Mysteries of the Mind Reprinted from the November 1990 issue

JACOB’S LADDER, painted in 1973 by Marc Chagall,depicts a biblical story. Jacob dreams of angels ascend-

ing to and descending from heaven on a ladder.

The Author

JONATHAN WINSONstarted his career as anaeronautical engineer, grad-uating with an engineeringdegree from the CaliforniaInstitute of Technology in1946. He completed hisPh.D. in mathematics atColumbia University andthen turned to business for15 years. Because of hiskeen interest in neurosci-ence, however, Winsonstarted to do research at theRockefeller University onmemory processing duringwaking and sleeping states.In 1979 he became an asso-ciate professor there and asprofessor emeritus contin-ues his work on memoryand dreaming. His researchhas been supported by theNational Institute of Men-tal Health, the National Sci-ence Foundation and theHarry F. Guggenheim Foundation.

Dreaming occurs solely during REM sleep.

SCA

LA/A

RT

RESO

URC

E

LAB

AT/J

ERRI

CA

N P

hoto

Res

earc

hers

, Inc

.

Copyright 1990 Scientific American, Inc.

Page 42: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

move rapidly in unison under closedlids, breathing becomes irregular andheart rate increases.

The first REM stage of the night fol-lows 90 minutes of slow-wave sleep andlasts for 10 minutes. The second andthird REM periods follow shorter slow-wave sleep episodes but grow progres-sively longer themselves. The fourth andfinal REM interval lasts 20 to 30 min-utes and is followed by awakening. If adream is remembered at all, it is mostoften the one that occurred in this lastphase of REM sleep.

This sleep cycle—alternating slow-wave and REM sleep—appears to bepresent in all placental and marsupialmammals. Mammals exhibit the vari-ous REM-associated characteristics ob-served in humans, including EEG read-ings similar to those of the awake state.Animals also dream. By destroying neu-rons in the brain stem that inhibit move-ment during sleep, researchers foundthat sleeping cats rose up and attackedor were startled by invisible objects—os-tensibly images from dreams.

By studying subprimate animals, sci-

entists have discovered additional neu-rophysiological aspects of REM sleep.They determined that neural control ofthis stage of the sleep cycle is centeredin the brain stem (the brain region clos-est to the spinal cord) and that duringREM sleep neural signals—called pon-tine-geniculate-occipital (PGO) cortexspikes—proceed from the brain stem tothe center of visual processing, the visu-al cortex. Brain stem neurons also initi-ate a sinusoidal wave (one resembling asine curve) in the hippocampus. Thisbrain signal is called theta rhythm.

Copyright 1990 Scientific American, Inc.

Page 43: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

At least one animal experiences slow-wave but not REM sleep—and, conse-quently, does not exhibit theta rhythmwhen asleep. This animal is the echidna,or spiny anteater, an egg-laying mam-mal (called a monotreme) that providessome insight into the origin of dream-ing. The absence of REM sleep in theechidna suggests that this stage of thesleep cycle evolved some 140 millionyears ago, when marsupials and placen-tals diverged from the monotreme line.(Monotremes were the first mammalsto develop from reptiles.)

By all evolutionary criteria, the perpet-uation of a complex brain process suchas REM sleep indicates that it serves animportant function for the survival ofmammalian species. Understanding thatfunction might reveal the meaning ofdreams.

When Freud wrote The Interpretationof Dreams, the physiology of sleep was

unknown. In light of the discovery ofREM sleep, certain elements of his psy-choanalytic theory were modified, andthe stage was set for more neurological-ly based theories. Dreaming came to beunderstood as part of a biologically de-termined sleep cycle. Yet the centralconcept of Freud’s theory—namely, thebelief that dreams reveal a censoredrepresentation of our innermost uncon-scious feelings and concerns—continuesto be used in psychoanalysis.

Some theorists abandoned Freud al-together following the neurological dis-coveries. In 1977 J. Allan Hobson andRobert McCarley of Harvard MedicalSchool proposed the “activation-syn-thesis” hypothesis. They suggested thatdreaming consists of associations andmemories elicited from the forebrain (theneocortex and associated structures) inresponse to random signals from thebrain stem such as PGO spikes. Dreams

were merely the “best fit” the forebraincould provide to this random bombard-ment from the brain stem. Althoughdreams might at times appear to havepsychological content, their bizarrenesswas inherently meaningless.

The sense, or plot, of dreams resultedfrom order that was imposed on thechaos of neural signals, Hobson said.“That order is a function of our ownpersonal view of the world, our remotememories,” Hobson wrote. In otherwords, the individual’s emotional vo-cabulary could be relevant to dreams.In a further revision of the original hy-pothesis, Hobson also suggested thatbrain stem activation may merely serveto switch from one dream episode toanother.

Reverse Learning

Although Hobson and McCarley had presented an explanation of dream

content, the basic function of REM sleepadmittedly remained unknown. In 1983Francis Crick of the Salk Institute in LaJolla, Calif., and Graeme Mitchison ofthe University of Cambridge in Englandproposed the idea of reverse learning.Working from the Hobson-McCarleyassumption of random neocortical bom-bardment by PGO waves and their ownknowledge of the behavior of stimulatedneural networks, they postulated that acomplex associational neural networksuch as the neocortex might becomeoverloaded by vast amounts of incominginformation. The neocortex could thendevelop false, or “parasitic,” thoughtsthat would jeopardize the true and or-derly storage of memory.

According to Crick and Mitchison’shypothesis, REM sleep served to erasethese spurious associations on a regularbasis. Random PGO waves impinged onthe neocortex, resulting in erasure, or un-learning, of the false information. Thisprocess served an essential function: itallowed the orderly processing of mem-ory. In humans, dreams were a runningrecord of these parasitic thoughts—ma-terial to be purged from memory. “Wedream to forget,” Crick and Mitchisonwrote.

The two researchers proposed a revi-sion in 1986. Erasure of parasiticthoughts accounted only for bizarredream content. Nothing could be saidabout dream narrative. Furthermore,dreaming to forget, they said, was bet-ter expressed as dreaming to reducefantasy or obsession.

The Meaning of Dreams60 Mysteries of the Mind

ANATOMY OF THE BRAIN and cross section of the hippocampus show some of the regions involved in dreaming. In the hippocampus, incoming information is processedsequentially in the dentate gyrus and the CA3 and the CA1 pyramidal cells. In subpri-mate species, theta rhythm is generated in the dentate gyrus and the CA1 cells.

CA

RO

L D

ON

NER

PREFRONTAL CORTEX

SEPTUM

HIPPOCAMPUS

DENTATE GYRUS

CA1 CELLS

SPINAL CORD

BRAIN STEM

HIPPOCAMPUS

VISUALCORTEXENTORHINAL

CORTEX

CA3CELLS

Copyright 1990 Scientific American, Inc.

Page 44: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

None of these hypotheses seems to ex-plain adequately the function of dream-ing. On the one hand, Freud’s theorylacked physiological evidence. (AlthoughFreud had originally intended to de-scribe the neurology of the unconsciousand of dreams in his proposed “Projectfor a Scientific Psychology,” the under-taking was premature, and he limitedhimself to psychoanalysis.) On the oth-er hand, despite revisions to incorpo-rate elements of psychology, most ofthe later theories denied that dreamshad meaning.

Exploring the neuroscientific aspectsof REM sleep and of memory process-ing seemed to me to hold the greatestpotential for understanding the mean-ing and function of dreams. The key tothis research was theta rhythm.

Theta rhythm was discovered in 1954in awake animals by John D. Green andArnaldo A. Arduini of the University ofCalifornia at Los Angeles. The research-ers observed a regular sinusoidal signalof six cycles per second in the hippocam-pus of rabbits when the animals wereapprehensive of stimuli in their envi-ronment. They named the signal thetarhythm after a previously discoveredEEG component of the same frequency.

Theta rhythm was subsequently re-corded in the tree shrew, mole, rat andcat. Although it was consistently ob-served in awake animals, theta rhythmwas correlated with very different be-haviors in each species. For example, in marked contrast to the rabbit, envi-ronmental stimuli did not induce the-

ta rhythm in the rat. Rats demonstratedtheta rhythm only during movement,typically when they explored. In 1969,however, Case H. Vanderwolf of the Uni-versity of Western Ontario discoveredthere was one behavior during which theanimals he studied, including the rat,showed theta rhythm: REM sleep.

In 1972 I published a commentarypointing out that the different occurrenc-es of theta rhythm could be understoodin terms of animal behavior. Awake an-imals seemed to show theta rhythmwhen they were behaving in ways mostcrucial to their survival. In other words,theta rhythm appeared when they exhib-ited behavior that was not geneticallyencoded—such as feeding or sexual be-havior—but rather a response to chang-ing environmental information. Preda-tory behavior in the cat, prey behaviorin the rabbit, and exploration in the ratare, respectively, most important to theirsurvival. (For example, a hungry rat willexplore before it eats even if food isplaced in front of it.)

Role of Theta Rhythm

Furthermore, because the hippocam-pus is involved in memory process-

ing, the presence of theta rhythm dur-ing REM sleep in that region of thebrain might be related to that activity. Isuggested that the theta rhythm reflect-ed a neural process whereby informationessential to the survival of a species—

gathered during the day—was repro-cessed into memory during REM sleep.

In 1974, by recording signals from thehippocampus of freely moving rats andrabbits, I found the source from whichtheta rhythm was generated in the hip-pocampus. Together with the neocortex,the hippocampus is believed to providethe neural basis for memory storage. Thehippocampus (from the Greek word for“seahorse,” which it resembles in shape)is a sequential structure composed ofthree types of neurons. Information fromall sensory and associational areas of theneocortex converges in a region calledthe entorhinal cortex; from there it istransmitted to the three successive neu-ronal populations of the hippocampus.The signal arrives first at the granulecells of the dentate gyrus, then at theCA3 pyramidal cells (so called becauseof their triangular shape) and finally atthe pyramidal cells of CA1. After infor-mation is processed by this trio of cells,it is retransmitted to the entorhinal cor-tex and then back to the neocortex.

My studies showed that theta rhythmwas produced in two regions within thehippocampus: the dentate gyrus and theCA1 neurons. The rhythms in these twoareas were synchronous. Subsequently,Susan Mitchell and James B. Ranck, Jr.,of the State University of New YorkDownstate Medical Center identified athird synchronous generator in the en-torhinal cortex, and Robert Verdes ofWayne State University discovered thebrain stem neurons that control thetarhythm. These neurons transmit signalsto the septum (a forebrain structure)that activate theta rhythm in the hippo-

The Meaning of Dreams Mysteries of the Mind 61

THETA RHYTHM is present during different waking behav-iors in different species. Each of these behaviors is pivotal to the

animal’s survival. In placental and marsupial animals, thetarhythm is present during rapid eye movement (REM) sleep.

PATR

ICIA

J. W

YNN

E

THETA RHYTHM

REM SLEEP

REM SLEEPREM SLEEP

EXPLORATIONPREDATION APPREHENSION

1 SEC

Copyright 1990 Scientific American, Inc.

Page 45: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

campus and the entorhinal cortex. Thus,the brain stem activates the hippocam-pus and the neocortex—the core memo-ry system of the brain.

To determine the relation betweentheta rhythm and memory, I made a le-sion in the rat septum. Rats that hadpreviously learned, using spatial cues,to locate a particular position in a mazewere no longer able to do so after theirseptums were disabled. Without thetarhythm, spatial memory was destroyed.

Studies of the cellular changes thatbring about memory illustrated the roleof theta rhythm. In particular, the dis-covery in 1973 of long-term potentia-tion (LTP)—a change in neural behavior

that reflects previous activity—showedthe means by which memory might beencoded. Timothy V. P. Bliss and A. R.Gardner-Medwin of the National Insti-tute of Medical Research in Londonand Terje Lømo of the University ofOslo found changes in nerve cells thathad been intensely stimulated with elec-trical pulses.

Long-Term Memory Storage

Earlier studies had shown that if onestimulated the pathway from the

entorhinal cortex to the granule cells ofthe hippocampus, the response of thesecells could be measured with a record-

ing electrode. Using this technique, Blissand his colleagues measured the normalresponse to a single electrical pulse. Thenthey applied a long series of high-fre-quency signals—called tetanic pulses—

to this pathway. After the train of tetanicstimuli, a single electrical pulse causedmuch greater firing in the granule cellsthan had been observed prior to the ex-periment. The heightened effect persistedfor as long as three days. This phenom-enon of LTP was precisely the kind ofincrease in neuronal strength that couldbe capable of sustaining memory. LTPis now considered a model for learningand memory.

LTP is achieved by the activity of theNMDA (N-methyl-D-aspartate) recep-tor. This molecule is embedded in thedendrites of the granule cells and theCA1 cells of the hippocampus as wellas in neurons throughout the neocor-tex. Like other neuronal receptors, theNMDA receptor is activated by a neu-rotransmitter—glutamate in this case.Glutamate momentarily opens a non-NMDA channel in the granule cell den-drite, allowing sodium from the extra-cellular space to flow into the neuron.This influx causes the granule cell to be-come depolarized. If the depolarizationis sufficient, the granule cell fires, trans-mitting information to other nerve cells.

Unlike other neuronal receptors,NMDA possesses an additional proper-ty. If a further activation of glutamateoccurs while the granule cell is depolar-ized, a second channel opens up, allow-ing an influx of calcium. Calcium isthought to act as a second messenger, ini-tiating a cascade of intracellular eventsthat culminates in long-lasting synapticchanges—or LTP. (The description giv-en here has been necessarily simplified.LTP is the subject of extensive ongoinginvestigation.)

Because the tetanic impulses appliedby Bliss and his colleagues did not oc-cur naturally in the brain, the questionremained as to how LTP was achievedunder normal circumstances. In 1986John Larson and Gary S. Lynch of theUniversity of California at Irvine andGregory Rose and Thomas V. Dunwid-die of the University of Colorado atDenver suggested that the occurrence ofLTP in the hippocampus was linked totheta rhythm. They applied a small num-ber of electrical pulses to CA1 cells inthe rat hippocampus and produced LTP,but only when the pulses were separat-ed by the normal time that elapses be-tween two theta waves—approximately

The Meaning of Dreams62 Mysteries of the Mind

CA

ROL

DO

NN

ERG

AB

OR

KISS

NMDA RECEPTOR activation induces long-term potentiation (LTP), a model formemory. The release of the neurotransmitter glutamate (left panel) opens a non–NMDA receptor channel, allowing the influx of sodium, which depolarizes the neuron.If a further release of glutamate occurs while the cell is depolarized (center panel), theNMDA receptor opens a second channel, which allows calcium to flow in, leading toLTP. LTP occurs as a result of increased sodium through the non-NMDA channel(right panel) and the subsequent greater depolarization of the cell.

PRESYNAPTICMEMBRANE

NEUROTRANSMITTERION CHANNEL

SYNAPTIC CLEFT

POSTSYNAPTICMEMBRANE

GLUTAMATE

CALCIUM

SODIUM

DEPOLARIZATIONSECOND RELEASEOF GLUTAMATE LTP

NMDARECEPTOR

Copyright 1990 Scientific American, Inc.

Page 46: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

200 milliseconds. Theta rhythm is ap-parently the natural means by which theNMDA receptor is activated in neuronsin the hippocampus.

Work in my laboratory at the Rocke-feller University duplicated Larson andLynch’s CA1 findings, but this time inthe hippocampal granule cells. Constan-tine Pavlides, Yoram J. Greenstein and Ithen demonstrated that LTP was depen-dent on the presence and phase of thetarhythm. If electrical pulses were appliedto the cells at the peak of the thetawave, LTP was induced. But if the samepulse were applied at the trough of thewaves—or when theta rhythm was ab-sent—LTP was not induced.

A coherent picture of memory process-ing was emerging. As a rat explores, forexample, brain stem neurons activatetheta rhythm. Olfactory input (whichin the rat is synchronized with thetarhythm, as is the twitching of whiskers)and other sensory information convergeon the entorhinal cortex and the hippo-campus. There they are partitioned into200-millisecond “bites” by theta rhythm.The NMDA receptors, acting in con-junction with theta rhythm, allow forlong-term storage of this information.

A similar process occurs during REMsleep. Although there is no incoming in-formation or movement during REMsleep, the neocortical-hippocampal net-work is once again paced by thetarhythm. Theta rhythm might producelong-lasting changes in memory.

Storing Spatial Memory

The results of one of my further ex-periments served to show that spa-

tial memory was indeed being stored inthe rat hippocampus during sleep. JohnO’Keefe and J. Dostrovsky of the Uni-

versity College London had demonstrat-ed that individual CA1 neurons in therat hippocampus fired when the awakeanimal moved to a particular location—

namely, the neuron’s place field. Theimplication of this finding was that theCA1 neuron fired to map the environ-ment, thereby committing it to memory.

In 1989 Pavlides and I located twoCA1 neurons in the rat hippocampusthat had different place fields. We re-corded from both cells simultaneously.After determining the normal firing ratesin awake and asleep animals, we posi-tioned a rat in the place field of one ofthe neurons. The neuron fired vigorous-ly, mapping that location. The secondcell fired only sporadically because itwas not coding space. We continued re-cording from the two pairs of neuronsas the rat moved about and then enteredseveral sleep cycles. Six pairs of neuronswere studied in this manner.

We found that neurons that had cod-ed space fired at a normal rate as theanimal moved about prior to sleep. Insleep, however, they fired at a signifi-cantly higher rate than their previoussleeping baseline. There was no such in-crease in firing rate during sleep in neu-rons that had not mapped space. Thisexperiment suggested that the reprocess-ing or strengthening of information en-coded when the animal was awake oc-curred in sleep at the level of individualneurons.

Bruce L. McNaughton and his col-leagues at the University of Arizona havedeveloped a technique for simultaneous-ly recording from a large number of neu-rons in the hippocampus that map loca-tions. Their technique allows definitivepatterns of firing to be identified. In an-imal studies, they found that ensemblesof place-field neurons that code space

in the waking state reprocess informa-tion during slow-wave sleep and then inREM sleep. These results suggest thatsleep processing of memory may havetwo stages—a preliminary stage in slow-wave sleep and a later phase in REMsleep, when dreaming occurs.

Evolution of REM Sleep

Evidence that theta rhythm encodesmemories during REM sleep may

be derived not only from neuroscientif-ic studies but also from evolution. Theemergence of a neural mechanism toprocess memory in REM sleep suggestsdifferences in brain anatomy betweenmammals that have that aspect of thesleep cycle and those that do not. Andin fact, such differences clearly exist be-tween the echidna and the marsupialsand placentals.

The echidna has a large convolutedprefrontal cortex, larger in relation tothe rest of the brain than that of anyother mammal, even humans. I believeit needed this huge prefrontal cortex toperform a dual function: to react to in-coming information in an appropriatemanner based on past experience and toevaluate and store new information toaid in future survival. Without thetarhythm during REM sleep, the echidnawould not be able to process informa-tion while it slept. (The echidna does,however, show theta rhythm when for-aging for food.) For higher capabilitiesto develop, the prefrontal cortex wouldhave to become increasingly large—be-yond the capacity of the skull—unlessanother brain mechanism evolved.

REM sleep could have provided thisnew mechanism, allowing memory pro-cessing to occur “off-line.” Coincidentwith the apparent development of REM

The Meaning of Dreams Mysteries of the Mind 63

LTP in the granule cells of the hippocampus is achieved by thetarhythm. Electrical pulses, which have been separated by 200

milliseconds (the time between the peaks of two theta waves),result in LTP when applied at the peak of theta rhythm.

PULSES AT PEAK

THETA RHYTHM

200 MILLISECONDS

ELECTRICAL PULSE

PULSES AT TROUGH

MIL

LIV

OLT

S

MILLISECONDS

0

0 10 20

LTP

NO LTP

3

6

MIL

LIV

OLT

S

MILLISECONDS

0

0 10 20

3

6

GA

BO

R KI

SS

Copyright 1990 Scientific American, Inc.

Page 47: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

sleep in marsupial and placental mam-mals was a remarkable neuroanatomicalchange: the prefrontal cortex was dra-matically reduced in size. Far less pre-frontal cortex was required to processinformation. That area of the brain coulddevelop to provide advanced perceptu-al abilities in higher species.

The nature of REM sleep supportsthis evolutionary argument. During theday, animals gather information that in-volves locomotion and eye movement.The reprocessing of this informationduring REM sleep would not be easilyseparated from the locomotion relatedto the experience—such disassociationmight be expecting too great a revisionof brain circuitry. So to maintain sleep,locomotion had to be suppressed by in-hibiting motor neurons. Suppressing eyemovement was unnecessary because thisactivity does not disturb sleep.

Eye movement potentials, similar to

PGO spikes, accompany rapid eye move-ment in the waking state and also dur-ing REM sleep. The function of thesesignals has not yet been established, butthey may serve to alert the visual cortexto incoming information when the ani-mal is awake and may reflect the repro-cessing of this information during REMsleep. In any case, PGO spikes do notdisturb sleep and do not have to be sup-pressed—unlike motor neurons.

Strategy for Survival

With the evolution of REM sleep,each species could process the in-

formation most important for its sur-vival, such as the location of food or themeans of predation or escape—thoseactivities during which theta rhythm ispresent. In REM sleep this informationmay be accessed again and integratedwith past experience to provide an on-

going strategy for behavior. Althoughtheta rhythm has not yet been demon-strated in primates, including humans,the brain signal provides a clue to theorigin of dreaming in humans. Dreamsmay reflect a memory-processing mech-anism inherited from lower species, inwhich information important for sur-vival is reprocessed during REM sleep.This information may constitute the coreof the unconscious.

Because animals do not possess lan-guage, the information they process dur-ing REM sleep is necessarily sensory.Consistent with our early mammalianorigins, dreams in humans are sensory,primarily visual. Dreams do not take theform of verbal narration.

Also in keeping with the role REMsleep played in processing memories inanimals, there is no functional necessityfor this material to become conscious.Consciousness arose later in evolutionin humans. But neither is there any rea-son for the material of dreams not toreach consciousness. Therefore, dreamscan be remembered—most readily ifawakening occurs during or shortly af-ter a REM sleep period.

Consistent with evolution and evi-dence derived from neuroscience andreports of dreams, I suggest that dreamsreflect an individual’s strategy for sur-vival. The subjects of dreams are broad-ranging and complex, incorporatingself-image, fears, insecurities, strengths,grandiose ideas, sexual orientation, de-sire, jealousy and love.

Dreams clearly have a deep psycho-logical core. This observation has beenreported by psychoanalysts since Freudand is strikingly illustrated by the workof Rosalind Cartwright of Rush-Presby-terian–St. Luke’s Hospital in Chicago.Cartwright is studying a series of 90 sub-jects who are undergoing marital sepa-ration and divorce. All the subjects areclinically evaluated and psychologicallytested to ascertain their attitudes andresponses to their personal crisis. Cart-wright’s subjects are also awakenedfrom REM sleep to report their dreams,

The Meaning of Dreams64 Mysteries of the Mind

EVOLUTIONARY TREE shows the di-vergence of placentals and marsupialsfrom monotremes. The echidna, whichdoes not possess REM sleep, has a largerprefrontal cortex compared with the restof its brain than does any mammal, evenhumans. It is larger than in similarly sizedanimals, including the opossum and cat.

CAT

PREFRONTALCORTEX

250

200

150

100

50

0

PLACENTALS MARSUPIALSMONOTREMES

REM SLEEP

OPOSSUM ECHIDNA

SLOW-WAVE SLEEP

LIVE-BEARINGMAMMALS

EGG-LAYINGMAMMALS

FIRST TRUE MAMMALS

MAMMALLIKEREPTILES

YEA

RS A

GO

(MIL

LIO

NS)

GA

BO

R KI

SSC

AR

OL

DO

NN

ER

Copyright 1990 Scientific American, Inc.

Page 48: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

which are then interpreted by the sub-jects themselves without questions thatmight influence their interpretation. Inthe 70 individuals studied to date, thedream content conveys the person’s un-conscious thoughts and is strongly cor-related with the manner in which he orshe is coping with the crisis while awake.

Although the topic “chosen” for con-sideration during a night’s sleep is un-predictable, certain of life’s difficulties—

as in the case of Cartwright’s subjects—

so engage psychological survival thatthey are selected for REM sleep process-ing. In the ordinary course of events, de-pending on the individual’s personality,the themes of dreams may be freewheel-ing. Moreover, when joined with the in-tricate associations that are an intrinsicpart of REM sleep processing, thedream’s statement may be rather obscure.

Nevertheless, there is every reason tobelieve that the cognitive process takingplace in Cartwright’s subjects occurs inevery individual. Interpretation of thecoherent statement that is being madedepends on the individual’s tracing ofrelevant or similar events. These associ-ations are strongly biased toward earlychildhood experience.

My hypothesis also offers an expla-nation for the large amount of REMsleep in infants and children. Newbornsspend eight hours a day in REM sleep.The sleep cycle is disorganized at thisage. Sleep occurs in 50- to 60-minutebouts and begins with REM ratherthan with slow-wave sleep. By the ageof two, REM sleep is reduced to threehours a day, and the adult pattern hasbeen established. Thereafter, the timespent in REM sleep gradually diminish-es to a little less than two hours.

REM sleep may perform a specialfunction in infants. A leading theoryproposes that it stimulates nerve growth.Whatever the purpose in infants may be,I suggest that at about the age of two,when the hippocampus, which contin-ues to develop after birth, becomes func-tional, REM sleep takes on its interpre-tive memory function. The waking in-formation to be integrated at this pointin development constitutes the basic cog-nitive substrate for memory—the con-cept of the real world against which lat-er experiences must be compared andinterpreted. The organization in memoryof this extensive infrastructure requiresthe additional REM sleep time.

For reasons he could not possibly haveknown, Freud set forth a profound truth

in his work. There is an unconscious, anddreams are indeed the “royal road” toits understanding. The characteristics ofthe unconscious and associated pro-cesses of brain functioning, however, arevery different from what Freud thought.Rather than being a cauldron of un-tamed passions and destructive wishes,I propose that the unconscious is a co-hesive, continually active mental struc-ture that takes note of life’s experiencesand reacts according to its own schemeof interpretation. Dreams are not dis-guised as a consequence of repression.Their unusual character is a result ofthe complex associations that are culledfrom memory.

Memory Consolidation

Research on REM sleep suggeststhat there is a biologically relevant

reason for dreaming. The revised versionof the Hobson-McCarley activation-synthesis hypothesis acknowledges thedeep psychological core of dreams. Inits present truncated form, the hypothe-sis of random brain stem activation haslittle explanatory or predictive power.

The Crick-Mitchison hypothesis pro-vides a function for REM sleep—reverselearning—but it does not apply to narra-tive, only to the bizarre elements of thedream. What this implies with regard toREM processing in lower species mustbe defined before the theory can be eval-uated further. In addition, the Crick-

Mitchison hypothesis as applied to thehippocampus would suggest that neu-rons fire randomly during REM sleep,providing reverse learning. Instead, in myexperiment on the neurons that codedspace, these neurons fired selectively, im-plying an orderly processing of memory.

Recently Avi Karni and his colleaguesat the Weizmann Institute of Science inIsrael were able to show that memoryprocessing occurs in humans duringREM sleep. In their experiment, indi-viduals learned to identify particular pat-terns on a screen. The memory of thisskill improved after a night with REMsleep. When the subjects were deprivedof REM sleep, memory consolidationdid not occur. This study is an impor-tant breakthrough and opens a particu-larly promising field for exploration.

Further study will continue to eluci-date the meaning of dreams. In particu-lar, an experiment is needed to deter-mine whether eliminating theta rhythmduring REM sleep alone results in amemory deficit. Because theta rhythmhas not been demonstrated in primates,it may have disappeared as vision re-placed olfaction as the dominant sense.An equivalent neural mechanism mayexist in the hippocampus that periodi-cally activates the NMDA receptor.These studies in animals and others tocome in humans will probe fundamen-tal aspects of memory processing andthe neuroscientific basis of human psy-chological structure.

The Meaning of Dreams Mysteries of the Mind 67

Further Reading

Interspecies Differences in the Occurrence of Theta. Jonathan Winson in Behav-ioral Biology, Vol. 7, No. 4, pages 479–487; 1972.

Loss of Hippocampal Theta Rhythm Results in Spatial Memory Deficit in theRat. Jonathan Winson in Science, Vol. 201, No. 435, pages 160–163; 1978.

Brain and Psyche: The Biology of the Unconscious. Jonathan Winson. Anchor Press,Doubleday, 1985.

Long-Term Potentiation in the Dentate Gyrus Is Induced Preferentially on thePositive Phase of Q-Rhythm. Constantine Pavlides, Yoram J. Greenstein, Mark Grud-man and Jonathan Winson in Brain Research, Vol. 439, pages 383–387; 1988.

Influences of Hippocampal Place Cell Firing in the Awake State on the Activityof These Cells during Subsequent Sleep Episodes. Constantine Pavlides andJonathan Winson in Journal of Neuroscience, Vol. 9, No. 8, pages 2907–2918; August,1989.

Dependence on REM Sleep of Overnight Improvement of a Perceptual Skill. AviKarni, David Tanne, Barton S. Rubenstein, Jean J. M. Askenasy and Dov Sagi in Science,Vol. 265, pages 679–682; July 29, 1994.

Reactivation of Hippocampal Ensemble Memories during Sleep. Mathew A. Wilsonand Bruce L. McNaughton in Science, Vol. 265, pages 676–679; July 29, 1994.

REM Sleep and the Reactivation of Recent Correlation Patterns in Hippocam-pal Neuronal Ensembles. H. S. Kudrimoto, W. F. Skaggs, C. A. Barnes, B. L. Mc-Naughton, J. L. Gerrard, M. S. Suster and K. L. Weaver in Society for Neuroscience Ab-stracts, page 1871; 1996.

SA

Copyright 1990 Scientific American, Inc.

Page 49: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Reprinted from the June 1994 issue

Despite millennia of preoccupation with every facet of human emotion,we are still far from explaining in a rigorous physiological sense thispart of our mental experience. Neuroscientists have, in modern times,been especially concerned with the neural basis of such cognitive pro-

cesses as perception and memory. They have for the most part ignored the brain’s rolein emotion. Yet in recent years, interest in this mysterious mental terrain has surged.Catalyzed by breakthroughs in understanding the neural basis of cognition and by anincreasingly sophisticated knowledge of the anatomical organization and physiologyof the brain, investigators have begun to tackle the problem of emotion.

One quite rewarding area of research has been the inquiry into the relation betweenmemory and emotion. Much of this examination has involved studies of one particu-lar emotion—fear—and the manner in which specific events or stimuli come, throughindividual learning experiences, to evoke this state. Scientists, myself included, havebeen able to determine the way in which the brain shapes how we form memoriesabout this basic, but significant, emotional event. We call this process “emotionalmemory.”

By uncovering the neural pathways through which a situation causes a creature tolearn about fear, we hope to elucidate the general mechanisms of this form of memo-ry. Because many human mental disorders—including anxiety, phobia, post-traumat-ic stress syndrome and panic attack—involve malfunctions in the brain’s ability tocontrol fear, studies of the neural basis of this emotion may help us further understandand treat these disturbances.

Most of our knowledge about how the brain links memory and emotion has beengleaned through the study of so-called classical fear conditioning. In this process thesubject, usually a rat, hears a noise or sees a flashing light that is paired with a brief,mild electric shock to its feet. After a few such experiences, the rat responds automat-ically to the sound or light, even in the absence of the shock. Its reactions are typicalto any threatening situation: the animal freezes, its blood pressure and heart rate in-crease, and it startles easily. In the language of such experiments, the noise or flash is aconditioned stimulus, the foot shock is an unconditioned stimulus, and the rat’s reac-tion is a conditioned response, which consists of readily measured behavioral andphysiological changes.

Emotion, Memory and the Brain

The neural routes underlying the formation of memories about primitive emotional experiences,

such as fear, have been traced

by Joseph E. LeDoux

68 Mysteries of the Mind

The Author

JOSEPH E. LEDOUX is in-terested in the neural founda-tion of memory and emotion.He studies the anatomy, physiology and behavioralorganization of these aspectsof mental functioning. Le-Doux, the Henry and LucyMoses Professor of Science atNew York University, is therecipient of two National In-stitute of Mental Health dis-tinctions: a Merit Award anda Research Scientist Develop-ment Award. He has also re-ceived an Established Investi-gator Award from the Ameri-can Heart Association.

MEH

AM

KU

LYK

SPL

/Pho

to R

esea

rche

rs, I

nc.

Memories of disturbing experiences form deep within

our brains.

Copyright 1997 Scientific American, Inc.

Page 50: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Conditioning of this kind happensquickly in rats—indeed, it takes place asrapidly as it does in humans. A singlepairing of the shock to the sound orsight can bring on the conditioned ef-fect. Once established, the fearful reac-tion is relatively permanent. If the noiseor light is administered many times with-out an accompanying electric shock, therat’s response diminishes. This change iscalled extinction. But considerable evi-dence suggests that this behavioral al-teration is the result of the brain’s con-trolling the fear response rather thanthe elimination of the emotional memo-ry. For example, an apparently extin-guished fear response can recover spon-taneously or can be reinstated by an ir-relevant stressful experience. Similarly,stress can cause the reappearance ofphobias in people who have been suc-

Emotion, Memory and the Brain Mysteries of the Mind 69

HIPPOCAMPUS

AUDITORYTHALAMUS

AMYGDALA

ANATOMY OF EMOTION includesseveral regions of the brain. Shown herein the rat (above), parts of the amygdala,the thalamus and the cortex interact tocreate memories about fearful experiencesassociated, in this case, with sound. Re-cent work has located precise areas wherefear is learned and remembered: certainparts of the thalamus (light pink at topright) communicate with areas in theamygdala (light yellow at bottom right)that process the fear-causing sound stim-uli. Because these neural mechanisms arethought to be similar in humans, the studyof emotional memory in rodents may illu-minate aspects of fear disorders in people. PH

OTO

GRA

PHS

BY

AN

DRE

W L

EON

ARD

APL

Mic

rosc

opic

ROB

ERTO

OST

I

AUDITORYCORTEX HIPPOCAMPUS

Copyright 1997 Scientific American, Inc.

Page 51: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

cessfully treated. This resurrection dem-onstrates that the emotional memory un-derlying the phobia was rendered dor-mant rather than erased by treatment.

Fear and Emotional Memory

Fear conditioning has proved an ide-al starting point for studies of emo-

tional memory for several reasons. First,it occurs in nearly every animal groupin which it has been examined: fruitflies, snails, birds, lizards, fish, rabbits,rats, monkeys and people. Although noone claims that the mechanisms are pre-cisely the same in all these creatures, itseems clear from studies to date that thepathways are very similar in mammalsand possibly in all vertebrates. We there-fore are confident in believing that manyof the findings in animals apply to hu-mans. In addition, the kinds of stimulimost commonly used in this type ofconditioning are not signals that rats—

or humans, for that matter—encounterin their daily lives. The novelty and ir-relevance of these lights and sounds helpto ensure that the animals have not al-ready developed strong emotional reac-tions to them. So researchers are clearlyobserving learning and memory atwork. At the same time, such cues donot require complicated cognitive pro-cessing from the brain. Consequently,the stimuli permit us to study emotionalmechanisms relatively directly. Finally,our extensive knowledge of the neuralpathways involved in processing acous-tic and visual information serves as anexcellent starting point for examiningthe neurological foundations of fearelicited by such stimuli.

My work has focused on the cerebralroots of learning fear, specifically fearthat has been induced in the rat by as-sociating sounds with foot shock. As domost other investigators in the field, Iassume that fear conditioning occurs

because the shock modifies the way inwhich neurons in certain important re-gions of the brain interpret the soundstimulus. These critical neurons arethought to be located in the neural path-way through which the sound elicits theconditioned response.

During the past 10 years, researchersin my laboratory, as well as in others,have identified major components of thissystem. Our study began at Cornell Uni-versity Medical College, where I workedseveral years ago, when my colleaguesand I asked a simple question: Is the au-ditory cortex required for auditory fearconditioning?

In the auditory pathway, as in othersensory systems, the cortex is the high-est level of processing; it is the culmina-tion of a sequence of neural steps thatstarts with the peripheral sensory recep-tors, located, in this case, in the ear. Iflesions in (or surgical removal of) partsof the auditory cortex interfered with

Emotion, Memory and the Brain70 Mysteries of the Mind

BLO

OD

PRE

SSU

RE (

MIL

LIM

ETER

S O

F M

ERC

URY

)

20

15

10

5

0

10

00 100 10 0 10 0 10 0 10 0 10

CES

SATI

ON

OF

MO

VEM

ENT

(CU

MU

LATI

VE

SEC

ON

DS)

BLO

OD

PRE

SSU

RE

20

15

10

5

0

BLO

OD

PRE

SSU

RE

20

15

10

5

0

10

0

CES

SATI

ON

OF

MO

VEM

ENT

10

0

CES

SATI

ON

OF

MO

VEM

ENT

2

4

6

8

2

4

6

8

2

4

6

8

TIME (SECONDS)

SOUND

ELECTRICITY

CLASSICAL FEAR CONDITIONING can be brought aboutby pairing a sound and a mild electric shock to the foot of a rat.In one set of experiments, the rat hears a sound (left), which haslittle effect on the animal’s blood pressure or patterns of move-ment. Next, the rat hears the same sound, coupled with a foot

shock (center). After several such pairings, the rat’s blood pres-sure rises at the same time that the animal holds still for an ex-tended period when it hears the sound. The rat has been fear-conditioned (right): sound alone achieves the same physiologicalchanges as did sound and shock together.

IAN

WO

RPO

LE

Copyright 1997 Scientific American, Inc.

Page 52: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

fear conditioning, we could concludethat the region is indeed necessary forthis activity. We could also deduce thatthe next step in the conditioning path-way would be an output from the audi-tory cortex. But our lesion experimentsin rats confirmed what a series of otherstudies had already suggested: the audi-tory cortex is not needed in order tolearn many things about simple acous-tic stimuli.

We then went on to make lesions inthe auditory thalamus and the auditorymidbrain, sites lying immediately belowthe auditory cortex. Both these areasprocess auditory signals: the midbrainprovides the major input to the thala-mus; the thalamus supplies the majorinput to the cortex. Lesions in both re-gions completely eliminated the rat’ssusceptibility to conditioning. This dis-covery suggested that a sound stimulusis transmitted through the auditory sys-tem to the level of the auditory thalamusbut that it does not have to reach thecortex for fear conditioning to occur.

This possibility was somewhat puz-zling. We knew that the primary nervefibers that carry signals from the audi-tory thalamus extend to the auditorycortex. So David A. Ruggiero, Donald

J. Reis and I looked again and foundthat, in fact, cells in some regions of theauditory thalamus also give rise to fibersthat reach several subcortical locations.Could these neural projections be theconnections through which the stimu-lus elicits the response we identify withfear? We tested this hypothesis by mak-ing lesions in each one of the subcorti-cal regions with which these fibers con-nect. The damage had an effect in onlyone area: the amygdala.

Filling in the Picture

That observation suddenly created aplace for our findings in an already

accepted picture of emotional process-ing. For a long time, the amygdala hasbeen considered an important brain re-gion in various forms of emotional be-havior. In 1979 Bruce S. Kapp and hiscolleagues at the University of Vermontreported that lesions in the amygdala’scentral nucleus interfered with a rab-bit’s conditioned heart rate responseonce the animal had been given a shockpaired with a sound. The central nucle-us connects with areas in the brain steminvolved in the control of heart rate,respiration and vasodilation. Kapp’s

work suggested that the central nucleuswas a crucial part of the system throughwhich autonomic conditioned respons-es are expressed.

In a similar vein, we found that lesionsof this nucleus prevented a rat’s bloodpressure from rising and limited its abil-ity to freeze in the presence of a fear-causing stimulus. We also demonstrat-ed, in turn, that lesions of areas to whichthe central nucleus connects eliminatedone or the other of the two responses.Michael Davis and his associates at YaleUniversity determined that lesions of thecentral nucleus, as well as lesions of an-other brain stem area to which the cen-tral nucleus projects, diminished yet an-other conditioned response: the increasedstartle reaction that occurs when an an-imal is afraid.

The findings from various laborato-ries studying different species and mea-suring fear in different ways all impli-cated the central nucleus as a pivotalcomponent of fear-conditioning circuit-ry. It provides connections to the vari-ous brain stem areas involved in the con-trol of a spectrum of responses.

Despite our deeper understanding ofthis site in the amygdala, many detailsof the pathway remained hidden. Does

Emotion, Memory and the Brain Mysteries of the Mind 71

AUDITORYTHALAMUS

AUDITORYMIDBRAIN

AUDITORYCORTEX

EFFECTS OF LESIONSON FEAR CONDITIONING

IMPLICATION OF LESIONS FOR FEAR-CONDITIONING PATHWAY

AUDITORY PATHWAYIN THE RAT BRAIN

CONDITIONINGDISRUPTED

CONDITIONINGDISRUPTED

EARAUDITORYNERVE

SOUND

LESION

NO EFFECT

UNKNOWN LOCATION

BRAIN LESIONS have been crucial to pinpointing the sites in-volved in experiencing and learning about fear. When a sound isprocessed by the rat brain, it follows a pathway from ear to mid-brain to thalamus to cortex (left). Lesions can be made in vari-ous sites in the auditory pathway to determine which areas are

necessary for fear conditioning (center). Only damage to thecortex does not disrupt the fear response, which suggests thatsome other areas of the brain receive the output of the thalamusand are involved in establishing memories about experiencesthat stimulate fear (right).

ROB

ERTO

OST

I

IAN

WO

RPO

LE

Copyright 1997 Scientific American, Inc.

Page 53: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

sound, for example, reach the centralnucleus directly from the auditory thal-amus? We found that it does not. Thecentral nucleus receives projections fromthalamic areas next to, but not in, theauditory part of the thalamus. Indeed,an entirely different area of the amyg-dala, the lateral nucleus, receives inputsfrom the auditory thalamus. Lesions ofthe lateral nucleus prevented fear con-ditioning. Because this site gets informa-tion directly from the sensory system,we have come to think of it as the senso-ry interface of the amygdala in fear con-ditioning. In contrast, the central nucle-us appears to be the interface with thesystems that control responses.

Mapping the Mechanism

These findings seemed to place us onthe threshold of being able to map

the entire stimulus response pathway.But we still did not know how informa-tion received by the lateral nucleus ar-rived at the central nucleus. Earlier stud-ies had suggested that the lateral nucleusprojects directly to the central nucleus,but the connections were fairly sparse.Working with monkeys, David Amaraland Asla Pitkanen of the Salk Institutefor Biological Studies in San Diego dem-onstrated that the lateral nucleus ex-tends directly to an adjacent site, called

the basal or basolateral nucleus, which,in turn, projects to the central nucleus.

Collaborating with Lisa Stefanacciand other members of the Salk team,Claudia R. Farb and C. Genevieve Goin my laboratory at New York Univer-sity found the same connections in therat. We then showed that these connec-tions form synaptic contacts—in otherwords, they communicate directly, neu-ron to neuron. Such contacts indicatethat information reaching the lateralnucleus can influence the central nucleusvia the basolateral nucleus. The lateralnucleus can also influence the centralnucleus by way of the accessory basalor basomedial nucleus. Clearly, ampleopportunities exist for the lateral nucle-us to communicate with the central nu-cleus once a stimulus has been received.

The emotional significance of such astimulus is determined not only by thesound itself but by the environment inwhich it occurs. Rats must thereforelearn not only that a sound or visual cueis dangerous, but under what conditionsit is so. Russell G. Phillips and I exam-ined the response of rats to the cham-ber, or context, in which they had beenconditioned. We found that lesions ofthe amygdala interfered with the ani-mals’ response to both the tone and thechamber. But lesions of the hippocam-pus—a region of the brain involved in

declarative memory—interfered onlywith response to the chamber, not thetone. (Declarative memory involves ex-plicit, consciously accessible informa-tion, as well as spatial memory.) At aboutthe same time, Michael S. Fanselow andJeansok J. Kim of the University of Cal-ifornia at Los Angeles discovered thathippocampal lesions made after fearconditioning had taken place also pre-vented the expression of responses tothe surroundings.

These findings were consistent withthe generally accepted view that the hip-pocampus plays an important role inprocessing complex information, suchas details about the spatial environmentwhere activity is taking place. Phillipsand I also demonstrated that the subicu-lum, a region of the hippocampus thatprojects to other areas of the brain,communicated with the lateral nucleusof the amygdala. This connection sug-gests that contextual information mayacquire emotional significance in thesame way that other events do—viatransmission to the lateral nucleus.

Although our experiments had iden-tified a subcortical sensory pathway thatgave rise to fear conditioning, we did notdismiss the importance of the cortex.The interaction of subcortical and cor-tical mechanisms in emotion remains ahotly debated topic. Some researchersbelieve cognition is a vital precursor toemotional experience; others think thatcognition—which is presumably a corti-cal function—is necessary to initiateemotion or that emotional processing isa type of cognitive processing. Still oth-ers question whether cognition is neces-sary for emotional processing.

It became apparent to us that the au-ditory cortex is involved in, though notcrucial to, establishing the fear response,at least when simple auditory stimuli areapplied. Norman M. Weinberger andhis colleagues at the University of Cali-fornia at Irvine have performed elegantstudies showing that neurons in the au-ditory cortex undergo specific physiolog-ical changes in their reaction to soundsas a result of conditioning. This findingindicates that the cortex is establishingits own record of the event.

Experiments by Lizabeth M. Roman-ski in my laboratory have determinedthat in the absence of the auditory cor-tex, rats can learn to respond fearfullyto a single tone. If, however, projectionsfrom the thalamus to the amygdala areremoved, projections from the thalamusto the cortex and then to the amygdala

Emotion, Memory and the Brain72 Mysteries of the Mind

AMYGDALASTIMULUS

HIPPOCAMPUSTHALAMUS

BASALNUCLEUS

ACCESSORYBASALNUCLEUS

LATERALNUCLEUS

CENTRALNUCLEUS

CORTEX

Structure of the Amygdala

The amygdala plays an important role in emotional behavior. Experiments inrodents have elucidated the structures of various regions of the amygdala

and their role in learning about and remembering fear. The lateral nucleus re-ceives inputs from sensory regions of the brain and transmits these signals to thebasal, the accessory basal and the central nuclei. The central nucleus connects tothe brain stem, bringing about physiological changes. —J.E.LeD.

Copyright 1997 Scientific American, Inc.

Page 54: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

are sufficient. Romanski went on to es-tablish that the lateral nucleus can re-ceive input from both the thalamus andthe cortex. Her work in the rat comple-ments earlier research in primates.

Once we had a clear understandingof the mechanism through which fearconditioning is learned, we attempted tofind out how emotional memories areestablished and stored on a molecularlevel. Farb and I showed that the excita-tory amino acid transmitter glutamateis present in the thalamic cells that reachthe lateral nucleus. Together with ChiyeJ. Aoki, we showed that it is also pres-ent at synapses in the lateral nucleus.Because glutamate transmission is impli-cated in memory formation, we seemedto be on the right track.

Long-Term Potentiation

Glutamate has been observed in aprocess called long-term potentia-

tion, or LTP, that has emerged as a mod-el for the creation of memories. This pro-cess, which is most frequently studiedin the hippocampus, involves a changein the efficiency of synaptic transmis-sion along a neural pathway—in otherwords, signals travel more readily alongthis pathway once LTP has taken place.The mechanism seems to involve gluta-mate transmission and a class of post-synaptic excitatory amino acid recep-tors known as NMDA receptors.

Various studies have found LTP inthe fear-conditioning pathway. Marie-Christine Clugnet and I noted that LTPcould be induced in the thalamo-amyg-dala pathway. Thomas H. Brown andPaul Chapman and their colleagues atYale discovered LTP in a cortical pro-jection to the amygdala. Other research-ers, including Davis and Fanselow, havebeen able to block fear conditioning byblocking NMDA receptors in the amyg-dala. And Michael T. Rogan in my lab-oratory found that the processing ofsounds by the thalamo-amygdala path-way is amplified after LTP has been in-duced. The fact that LTP can be dem-onstrated in a conditioning pathway of-fers new hope for understanding howLTP might relate to emotional memory.

In addition, recent studies by FabioBordi, also in my laboratory, have sug-gested hypotheses about what could begoing on in the neurons of the lateral nu-cleus during learning. Bordi monitoredthe electrical state of individual neuronsin this area when a rat was listening tothe sound and receiving the shock. He

and Romanski found that essentially ev-ery cell responding to the auditory stim-uli also responded to the shock. The ba-sic ingredient of conditioning is thuspresent in the lateral nucleus.

Bordi was able to divide the acousti-cally stimulated cells into two classes:habituating and consistently responsive.Habituating cells eventually stopped re-sponding to the repeated sound, suggest-ing that they might serve to detect anysound that was unusual or different.They could permit the amygdala to ig-nore a stimulus once it became familiar.Sound and shock pairing at these cellsmight reduce habituation, thereby al-lowing the cells to respond to, ratherthan ignore, significant stimuli.

The consistently responsive cells hadhigh-intensity thresholds: only loudsounds could activate them. That findingis interesting because of the role loud-ness plays in judging distance. Nearbysources of sound are presumably moredangerous than those that are far away.Sound coupled with shock might act onthese cells to lower their threshold, in-creasing the cells’ sensitivity to the samestimulus. Consistently responsive cellswere also broadly tuned. The joining ofa sound and a shock could make thecells responsive to a narrower range offrequencies, or it could shift the tuningtoward the frequency of the stimulus.In fact, Weinberger has recently shownthat cells in the auditory system do altertheir tuning to approximate the condi-tioned stimulus. Bordi and I have de-tected this effect in lateral nucleus cellsas well.

The apparent permanence of thesememories raises an important clinicalquestion: Can emotional learning beeliminated, and, if not, how can it betoned down? As noted earlier, it is actu-ally quite difficult to get rid of emotion-al memories, and at best we can hope

Emotion, Memory and the Brain Mysteries of the Mind 73

PRESYNAPTICNEURON

POSTSYNAPTICNEURON

NMDARECEPTOR

GLUTAMATE

NEU

RAL

IMPU

LSE

(MIL

LIV

OLT

S)

TIME (MILLISECONDS)0 10 20

AFTER LTP

BEFORE LTP

MEMORY FORMATION has beenlinked to the establishment of long-termpotentiation, or LTP. In this model ofmemory the neurotransmitter glutamateand its receptors, called NMDA receptors(top), bring about strengthened neuraltransmission. Once LTP is established, thesame neural signals produce larger re-sponses (top, middle). Emotional mem-ories may also involve LTP in the amyg-dala. Glutamate (red circle in top photo-graph) and NMDA receptors (red circlein bottom photograph) have been foundin the region of the amygdala where fearconditioning takes place. PH

OTO

GR

APH

S B

Y JO

SEPH

E. L

E DO

UX

IAN

WO

RPO

LEIA

N W

ORP

OLE

Copyright 1997 Scientific American, Inc.

Page 55: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

only to keep them under wraps. Studiesby Maria A. Morgan in my laboratoryhave begun to illuminate how the brainregulates emotional expressions. Mor-gan has shown that when part of theprefrontal cortex is damaged, emotionalmemory is very hard to extinguish. Thisdiscovery indicates that the prefrontalareas—possibly by way of the amyg-dala—normally control expression ofemotional memory and prevent emotion-al responses once they are no longer use-ful. A similar conclusion was proposedby Edmund T. Rolls and his colleaguesat the University of Oxford during stud-ies of primates. The researchers studiedthe electrical activity of neurons in thefrontal cortex of the animals.

Functional variation in the pathway

between this region of the cortex andthe amygdala may make it more difficultfor some people to change their emotion-al behavior. Davis and his colleagueshave found that blocking NMDA recep-tors in the amygdala interferes with ex-tinction. Those results hint that extinc-tion is an active learning process. At thesame time, such learning could be situ-ated in connections between the pre-frontal cortex and the amygdala. Moreexperiments should disclose the answer.

Placing a basic emotional memoryprocess in the amygdalic pathway yieldsobvious benefits. The amygdala is acritical site of learning because of itscentral location between input and out-put stations. Each route that leads tothe amygdala—sensory thalamus, sen-

sory cortex and hippocampus—deliversunique information to the organ. Path-ways originating in the sensory thala-mus provide only a crude perception ofthe external world, but because they in-volve only one neural link, they are quitefast. In contrast, pathways from the cor-tex offer detailed and accurate represen-tations, allowing us to recognize an ob-ject by sight or sound. But these path-ways, which run from the thalamus tothe sensory cortex to the amygdala, in-volve several neural links. And each linkin the chain adds time.

Emotion, Memory and the Brain74 Mysteries of the Mind

VISUAL THALAMUSVISUALCORTEX

AMYGDALA

BLOOD PRESSURE MUSCLE

HEART RATE

CORTICAL AND SUBCORTICAL PATHWAYS in the brain—

generalized from our knowledge of the auditory system—maybring about a fearful response to a snake on a hiker’s path. Visu-al stimuli are first processed by the thalamus, which passesrough, almost archetypal, information directly to the amygdala(red). This quick transmission allows the brain to start to re-spond to the possible danger (green). Meanwhile the visual cor-

tex also receives information from the thalamus and, with moreperceptual sophistication and more time, determines that thereis a snake on the path (blue). This information is relayed to theamygdala, causing heart rate and blood pressure to increase andmuscles to contract. If, however, the cortex had determined thatthe object was not a snake, the message to the amygdala wouldquell the fear response.

ROB

ERTO

OST

I

Copyright 1997 Scientific American, Inc.

Page 56: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Conserving time may be the reasonthere are two routes—one cortical andone subcortical—for emotional learning.Animals, and humans, need a quick-and-dirty reaction mechanism. The thalamusactivates the amygdala at about the sametime as it activates the cortex. The ar-rangement may enable emotional re-sponses to begin in the amygdala beforewe completely recognize what it is weare reacting to or what we are feeling.

The thalamic pathway may be partic-ularly useful in situations requiring arapid response. Failing to respond todanger is more costly than respondinginappropriately to a benign stimulus.For instance, the sound of rustling leavesis enough to alert us when we are walk-ing in the woods without our havingfirst to identify what is causing thesound. Similarly, the sight of a slendercurved shape lying flat on the path aheadof us is sufficient to elicit defensive fearresponses. We do not need to go througha detailed analysis of whether or notwhat we are seeing is a snake. Nor dowe need to think about the fact thatsnakes are reptiles and that their skinscan be used to make belts and boots. Allthese details are irrelevant and, in fact,detrimental to an efficient, speedy andpotentially lifesaving reaction. The brainsimply needs to be able to store primi-tive cues and detect them. Later, coordi-nation of this basic information with thecortex permits verification (yes, this is asnake) or brings the response (scream-ing, sprinting) to a stop.

Storing Emotional Memory

Although the amygdala stores primi-tive information, we should not

consider it the only learning center. Theestablishment of memories is a functionof the entire network, not just of onecomponent. The amygdala is certainlycrucial, but we must not lose sight ofthe fact that its functions exist only byvirtue of the system to which it belongs.

Memory is generally thought to bethe process by which we bring back tomind some earlier conscious experience.The original learning and the remem-bering, in this case, are both consciousevents. Workers have determined thatdeclarative memory is mediated by thehippocampus and the cortex. But re-moval of the hippocampus has little ef-fect on fear conditioning—except con-ditioning to context.

In contrast, emotional learning thatcomes about through fear conditioning

is not declarative learn-ing. Rather it is mediat-ed by a different system,which in all likelihoodoperates independentlyof our conscious aware-ness. Emotional infor-mation may be storedwithin declarative mem-ory, but it is kept thereas a cold declarativefact. For example, if aperson is injured in anautomobile accident inwhich the horn getsstuck in the on position,he or she may later havea reaction when hearing the blare of carhorns. The person may remember thedetails of the accident, such as whereand when it occurred, who else was in-volved and how awful it was. These aredeclarative memories that are depen-dent on the hippocampus. The individ-ual may also become tense, anxious anddepressed, as the emotional memory isreactivated through the amygdalic sys-tem. The declarative system has storedthe emotional content of the experi-ence, but it has done so as a fact.

Emotional and declarative memoriesare stored and retrieved in parallel, andtheir activities are joined seamlessly inour conscious experience. That does notmean that we have direct conscious ac-cess to our emotional memory; it meansinstead that we have access to the con-sequences—such as the way we behave,the way our bodies feel. These conse-quences combine with current declara-tive memory to form a new declarativememory. Emotion is not just uncon-scious memory: it exerts a powerful in-fluence on declarative memory and oth-er thought processes. As James L. Mc-Gaugh and his colleagues at theUniversity of California at Irvine haveconvincingly shown, the amygdala playsan essential part in modulating the stor-age and strength of memories.

The distinction between declarativememory and emotional memory is animportant one. W. J. Jacobs of the Uni-versity of British Columbia and LynnNadel of the University of Arizona haveargued that we are unable to remembertraumatic events that take place early inlife because the hippocampus has notyet matured to the point of formingconsciously accessible memories. Theemotional memory system, which maydevelop earlier, clearly forms and storesits unconscious memories of these events.

And for this reason, thetrauma may affect men-tal and behavioral func-tions in later life, albeitthrough processes thatremain inaccessible toconsciousness.

Because pairing a toneand a shock can bringabout conditioned re-sponses in animalsthroughout the phyla, itis clear that fear condi-tioning cannot be de-pendent on conscious-ness. Fruit flies andsnails, for example, are

not creatures known for their consciousmental processes. My way of interpret-ing this phenomenon is to consider feara subjective state of awareness broughtabout when brain systems react to dan-ger. Only if the organism possesses asufficiently advanced neural mechanismdoes conscious fear accompany bodilyresponse. This is not to say that onlyhumans experience fear but, rather, thatconsciousness is a prerequisite to sub-jective emotional states.

Thus, emotions or feelings are con-scious products of unconscious process-es. It is crucial to remember that thesubjective experiences we call feelingsare not the primary business of the sys-tem that generates them. Emotional ex-periences are the result of triggeringsystems of behavioral adaptation thathave been preserved by evolution. Sub-jective experience of any variety is chal-lenging turf for scientists. We have,however, gone a long way toward un-derstanding the neural system that un-derlies fear responses, and this samesystem may in fact give rise to subjec-tive feelings of fear. If so, studies of theneural control of emotional responsesmay hold the key to understanding sub-jective emotion as well.

Emotion, Memory and the Brain Mysteries of the Mind 75

Further Reading

The Amygdala: Neurobiological As-pects of Emotion, Memory and Men-tal Dysfunction. Edited by John P. Ag-gleton. Wiley-Liss, 1992.

Brain Mechanisms of Emotion andEmotional Learning. J. E. LeDoux inCurrent Opinion in Neurobiology. Vol. 2,No. 2, pages 191–197; April 1992.

The Role of the Amygdala in Fear andAnxiety. M. Davis in Annual Review ofNeuroscience, Vol. 15, pages 353–375;1992.

Emotional and

declarative

memories are

joined seamlessly

in our conscious

experience. That

does not mean

we have direct

conscious access

to our emotional

memory.

SA

Copyright 1997 Scientific American, Inc.

Page 57: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Over the years, most people acquire a repertoireof skills for coping with a range of frighteningsituations. They will attempt to placate a vexedteacher or boss and will shout and run when

chased by a mugger. Some individuals, though, become over-whelmed in circumstances others would consider only minimal-ly stressful: fear of ridicule might cause them to shake uncon-trollably when called on to speak in a group, or terror of strang-ers might lead them to hide at home, unable to work or shopfor groceries. Why do certain people fall prey to excessive fear?

At the University of Wisconsin at Madison, my colleagueSteven E. Shelton and I are addressing this problem by identify-ing specific brain processes that regulate fear and its associatedbehaviors. Despite the availability of noninvasive imaging tech-niques, such information is still extremely difficult to obtain inhumans. Hence, we have turned our attention to another pri-mate, the rhesus monkey (Macaca mulatta). These animals un-dergo many of the same physiological and psychological devel-opmental stages that humans do, but in a more compressedtime span. As we gain more insight into the nature and opera-tion of neural circuits that modulate fear in monkeys, it shouldbe possible to pinpoint the brain processes that cause inordinateanxiety in people and to devise new therapies to counteract it.

Effective interventions would be particularly beneficial ifthey were applied at an early age. Growing evidence suggestsoverly fearful youngsters are at high risk for later emotionaldistress. Jerome Kagan and his colleagues at Harvard Uni-versity have shown, for example, that a child who is profound-ly shy at the age of two years is more likely than a less inhibitedchild to suffer from anxiety and depression later in life.

This is not to say these ailments are inevitable. But it is easy tosee how excessive fear could contribute to a lifetime of emotion-al struggle. Consider a child who is deeply afraid of other chil-dren and is therefore taunted by them at school. That youngstermight begin to feel unlikable and, in turn, to withdraw further.With time the growing child could become mired in a viciouscircle leading to isolation, low self-esteem, underachievementand the anxiety and depression noted by Kagan.

There are indications that unusually fearful children might

The Neurobiology of FearResearchers are teasing apart the neurochemical processes that give rise to different fears in monkeys. The results may lead to new ways to treat anxiety in humans

by Ned H. Kalin

76 Mysteries of the Mind Reprinted from the May 1993 issue

The Author

NED H. KALIN, a clinicianand researcher, is professor ofpsychiatry and psychologyand chairman of the depart-ment of psychiatry at the Uni-versity of Wisconsin–MadisonMedical School. He is also ascientist at the Wisconsin Re-gional Primate Research Cen-ter and the Harlow PrimateLaboratory at the university.He earned his B.S. degree in1972 from Pennsylvania StateUniversity and his M.D. in1976 from Jefferson MedicalCollege in Philadelphia. Be-fore joining his current de-partments, he completed aresidency program in psychi-atry at Wisconsin and a post-doctoral fellowship in clinicalneuropharmacology at theNational Institute of MentalHealth.

RHESUS MONKEY REGISTERS ALARM (right) as another monkey approaches her baby. The mother’s fear is evident in her “threat” face: the open mouth and piercing

stare serve to intimidate would-be attackers and intruders.

Wild monkey displays cooing—a typical fear behavior.

NED

H. K

ALI

N

Copyright 1997 Scientific American, Inc.

Page 58: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

also be prone to physical illness. Manyyoungsters who become severely inhib-ited in unfamiliar situations chronicallyoverproduce stress hormones, includingthe adrenal product cortisol. In times ofthreat, these hormones are critical. Theyensure that muscles have the energyneeded for “fight or flight.” But someevidence indicates long-term elevationsof stress hormones may contribute togastric ulcers and cardiovascular disease.

Further, through unknown mecha-nisms, fearful children and their fami-lies are more likely than others to sufferfrom allergic disorders. Finally, in ro-

dents and nonhuman primates, persis-tent elevation of cortisol has been shownto increase the vulnerability of neuronsin the hippocampus to damage by othersubstances; this brain region is involvedin memory, motivation and emotion.Human neurons probably are affectedin a similar manner, although direct ev-idence is awaited.

When we began our studies about 10years ago, Shelton and I knew we wouldfirst have to find cues that elicit fear andidentify behaviors that reflect differenttypes of anxiety. With such informationin hand, we could proceed to determine

the age at which monkeys begin tomatch defensive behaviors selectively tospecific cues. By also determining theparts of the brain that reach maturityduring the same time span, we couldgain clues to the regions that underliethe regulation of fear and fear-relatedbehavior.

The experiments were carried out atthe Wisconsin Regional Primate Re-search Center and the Harlow PrimateLaboratory, both at the University ofWisconsin. We discerned varied behav-iors by exposing monkeys between sixand 12 months old to three related situ-

The Neurobiology of Fear Mysteries of the Mind 77

Copyright 1997 Scientific American, Inc.

Page 59: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

ations. In the alone condition, an ani-mal was separated from its mother andleft by itself in a cage for 10 minutes. Inthe no-eye-contact condition, a personstood motionless outside the cage andavoided looking at the solitary infant.In the stare condition, a person wasagain present and motionless but, as-suming a neutral expression, peered di-rectly at the animal. These conditionsare no more frightening than those pri-mates encounter frequently in the wildor those human infants encounter everytime they are left at a day-care center.

Three Typical Fear Behaviors

In the alone condition, most monkeysbecame very active and emitted fre-

quent “coo” calls. These fairly melodi-ous sounds are made with pursed lips.They start at a low pitch, become high-er and then fall. More than 30 yearsago Harry F. Harlow, then at Wiscon-sin, deduced that when an infant mon-key is separated from its mother, its pri-mary goal is affiliative—that is, it yearnsto regain the closeness and sense of se-curity provided by nearness to the par-ent. Moving about and cooing help todraw the mother’s attention.

In contrast, in the more frighteningno-eye-contact situation, the monkeysreduced their activity greatly and some-times “froze,” remaining completely stillfor prolonged periods. When an infantspots a possible predator, its goal shiftsfrom attracting the mother to becominginconspicuous. Inhibiting motion andfreezing—common responses in manyspecies—reduce the likelihood of attack.

If the infant perceives that it has beendetected, its aim shifts again, to ward-ing off an attack. And so the stare con-dition evoked a third set of responses.The monkeys made several hostile ges-tures, among them “barking” (forcingair from the abdomen through the vo-

cal cords to emit agrowllike sound), star-ing back, producing so-called threat faces [seeillustration on preced-ing page], baring theirteeth and shaking thecage. Sometimes the an-imals mixed the threat-ening displays with sub-missive ones, such asfear grimaces, whichlook something likewary grins, or grinding of the teeth. Inthis condition, too, cooing increasedover the amount heard when the ani-mals were alone. (As will be seen, wehave recently come to think the cooingdisplayed in the stare condition mayserve a somewhat different function thanit does in the alone situation.)

Monkeys, by the way, are not uniquein becoming aroused by stares and inusing them reciprocally to intimidatepredators. Animals as diverse as crabs,lizards and birds all perceive staring asa threat. Some fishes and insects haveevolved protective spots that resembleeyes; these spots either avert attackscompletely or redirect them to nonvitalparts of the body. In India, field-work-ers wear face masks be-hind their heads to dis-courage tigers frompouncing at their backs.Studies of humans showthat we, too, are sensi-tive to direct gazes: brainactivity increases whenwe are stared at, andpeople who are anxiousor depressed tend toavoid direct eye contact.

Having identifiedthree constellations ofdefensive behaviors, weset about determiningwhen infant monkeys

first begin to apply them effectively.Several lines of work led us to surmisethat the ability to make such choicesemerges sometime around an infant’stwo-month birthday. For instance, rhe-sus mothers generally permit childrento venture off with their peers at thattime, presumably because the adults arenow confident that the infants can pro-tect themselves reasonably well. Wealso knew that by about 10 weeks ofage infant monkeys respond with dif-ferent emotions to specific expressionson the faces of other monkeys—a signthat at least some of the innate wiringor learned skills needed to discriminatethreatening cues are in place.

To establish the critical period of de-

The Neurobiology of Fear78 Mysteries of the Mind

TYPICAL BEHAVIORS induced by the alone, no-eye-contact andstare conditions in the laboratory—such as cooing (left), freezing (cen-ter) and hostile display of the teeth (right)—are also seen in frightenedinfants and adults living in the wild. In this case, the setting is CayoSantiago, an island off the mainland of Puerto Rico.

THREE EXPERIMENTAL CONDITIONS elicit distinct fear-relatedbehaviors in rhesus monkeys older than about two months. When iso-lated in a cage (left), youngsters become quite active and emit “coo”sounds to attract their mothers. If a human appears but avoids eyecontact (center), the monkeys try to evade discovery, such as by stay-ing completely still (freezing) or hiding behind their food bin. If the in-truder stares at the animals (right), they become aggressive.

CA

RO

L D

ON

NER

Copyright 1997 Scientific American, Inc.

Page 60: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

velopment, we examined four groups ofmonkeys ranging in age from a few daysto 12 weeks old. We separated the ba-bies from their mothers and let them ac-climate to an unfamiliar cage. Then weexposed them to the alone, no-eye-con-tact and stare conditions. All sessionswere videotaped for analysis.

We found that infants in the youngestgroup (newborns to two-week-olds) en-gaged in defensive behaviors. But theylacked some motor coordination andseemed to act randomly, as if they wereoblivious to the presence or gaze of thehuman intruder. Babies in our two in-termediate-age groups had good motorcontrol, but their actions seemed un-related to the test condition. This finding

meant motor control was not the primedeterminant of selective responding.

Only animals in our oldest group(nine- to 12-week-olds) conducted them-selves differently in each situation, andtheir reactions were both appropriateand identical to those of mature mon-keys. Nine to 12 weeks, then, is the crit-ical age for the appearance of a monkey’sability to adaptively modulate its defen-sive activity to meet changing demands.

Studies by other workers, who pri-marily examined rodents, suggested thatthree interconnected parts of the brainregulate fearfulness. We suspected thatthese regions become functionally ma-ture during the nine- to 12-week periodand thus give rise to the selective reac-

tivity we observed. One of these regionsis the prefrontal cortex, which takes upmuch of the outer and side areas of thecerebral cortex in the frontal lobe [seetop illustration on next page]. A cogni-tive and emotional area, the prefrontalcortex is thought to participate in theinterpretation of sensory stimuli and isprobably a site where the potential fordanger is assessed.

The second region is the amygdala, a part of a primitive area in the braincalled the limbic system (which includesthe hippocampus). The limbic system ingeneral and the amygdala in particularhave been implicated in generating fear.

The final region is the hypothalamus.Located at the base of the brain, it is a

The Neurobiology of Fear Mysteries of the Mind 79

PHO

TOG

RA

PHS

BY

NED

H. K

ALI

N

Copyright 1997 Scientific American, Inc.

Page 61: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

constituent of what is called the hypo-thalamic-pituitary-adrenal system. Inresponse to stress signals from elsewherein the brain, such as the limbic systemand other cortical regions, the hypotha-lamus secretes corticotropin-releasinghormone. This small protein spurs the

pituitary gland, located just below thebrain, to secrete adrenocorticotropichormone (ACTH), which prods theadrenal gland to release cortisol, whichprepares the body to defend itself.

In neuroanatomic data collected inother laboratories, we found support for

our suspicion that maturation of thesebrain regions underlies selective respond-ing in the nine- to 12-week period. Forinstance, during this time the formationof synapses (contact points between neu-rons) has been shown to reach its peakin the prefrontal cortex and the limbic

The Neurobiology of Fear80 Mysteries of the Mind

THREE BRAIN REGIONS that are interconnected by neuralpathways (shown schematically by red lines) are critically im-portant in regulating fear-related behaviors. The prefrontal cor-tex (purple) participates in assessing danger. The amygdala(dark blue) is a major constituent of the emotion-producing lim-

bic system (light blue). And the hypothalamus (green), in re-sponse to signals from the prefrontal cortex, amygdala and hip-pocampus, directs the release of hormones (red arrows in box)that support motor responses to perceived threats. (Gray arrowsrepresent inhibitory activity by cortisol.)

INFANT (left) has strayed a short distance from its mother(center) and is producing a rudimentary threat face in an at-tempt to keep a photographer (the author) at bay. Rhesus mon-keys become adept at matching their behavior to the severity

and type of a threat when they are between nine and 12 weeksold, probably because certain neuronal pathways in three re-gions of the brain—the prefrontal cortex, amygdala and hy-pothalamus—reach functional maturity during this same period.

NED

H. K

ALI

N

HIPPOCAMPUSAMYGDALAHYPOTHALAMUSPREFRONTALCORTEX C

ARO

L D

ON

NER

INCREASED DELIVERYOF FUEL TO HEART, BRAINAND SKELETAL MUSCLE

HYPOTHALAMUS

CORTISOL

CORTICOTROPIN-RELEASINGHORMONE

PITUITARYGLAND

ACTH

ADRENALGLAND

Copyright 1997 Scientific American, Inc.

Page 62: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

system (including the amygdala), as wellas in the motor and visual cortices andother sensory areas. Patricia S. Gold-man-Rakic of Yale University has alsoestablished that as the prefrontal cortexmatures in rhesus monkeys, the abilityto guide behavior based on experienceemerges. This skill is necessary if one isto contend successfully with danger.

Maturation of the prefrontal cortexlikewise seems important for enablinghumans to distinguish among threaten-ing cues. Harry T. Chugani and his co-workers at the University of Californiaat Los Angeles have shown that activityin the prefrontal cortex increases whenhuman offspring are seven to 12 monthsof age. During this span—which appearsto be analogous to the time when mon-keys begin to respond selectively to fear—children begin to display marked fearof strangers. They also become adept atwhat is called social referencing; theyregulate their level of fear based on in-terpreting the expressions they observeon a parent’s face.

But what of the hypothalamus, thethird brain region we assumed couldparticipate in regulating fear-related be-havior? Published research did not tellus much about its development or aboutthe development of the complete hypo-thalamic-pituitary-adrenal system inmonkeys. Our own investigations, how-ever, revealed that the full system ma-tures in parallel with that of the pre-frontal cortex and the limbic system.

In these studies, we used the pituitaryhormone ACTH as a marker of the sys-tem’s function. We again examined fourgroups of infants aged a few days to 12weeks. From each subject, we measuredACTH levels in blood drawn while theyoungster was with its mother. Thisreading provided a baseline. We alsomeasured ACTH levels in blood sam-ples obtained 20 minutes after the infantwas separated from its parent. Hormon-al levels rose in all four age groups dur-ing separation, but they jumped pro-foundly only in the oldest (nine- to 12-week-old) monkeys.

The relatively weak response in theyounger animals, particularly in thoseunder two weeks old, is consistent withfindings in rat pups, whose stress hor-mone response is also blunted duringthe first two weeks of life. The develop-ment of the rodent and primate stresshormone system may well be delayedduring early life to protect young neu-rons from the potentially damaging ef-fects of cortisol.

Assured that the hypothalamic-pitu-itary-adrenal system becomes function-ally mature by nine to 12 weeks, wepressed the inquiry forward to deter-mine whether levels of cortisol andACTH might partly account for indi-vidual differences in defensive behavior.We were also curious to know whetherthe responses of the infants resembledthose of their mothers; a correspon-dence would indicate that further anal-yses of mothers and their infants couldhelp reveal the relative contributions ofinheritance and learning to fearfulness.We mainly examined the propensity forfreezing, which we had earlier foundwas a stable trait in our subjects.

Maturing of the Fear Response

In one set of studies, we measuredbaseline levels of cortisol in monkeys

four months to a year old and then ob-served how much time the youngstersfroze in the no-eye-contact condition.Monkeys that started off with relativelylow levels of cortisol froze for shorterperiods than did their counterparts withhigher cortisol levels—a pattern we alsonoted in separate studies of adult fe-males. In other studies, we observed thatas youngsters pass through their firstyear of life, they become progressivelylike their mothers hormonally and be-haviorally. By the time infants are aboutfive months old, their stress-induced ris-es in ACTH levels parallel those of themothers. And by the time they are a yearold, the duration of freezing in the no-

eye-contact condition also correspondsto that of the mother.

Strikingly, some of these results echoedthose obtained in humans. Extremely in-hibited children often have parents whosuffer from anxiety. Moreover, Kaganand his colleagues have found that bas-al cortisol levels are predictive of suchchildren’s reaction to a frightening situ-ation. They measured cortisol concen-trations in saliva of youngsters at home(where they are presumably most re-laxed) and then observed the childrenconfronting an unfamiliar situation inthe laboratory; high basal cortisol levelswere associated with greater inhibitionin the strange setting.

These similarities between humansand monkeys again imply that monkeysare reasonable models of human emo-tional reactivity. The link between basalcortisol levels and duration of freezingor inhibition suggests as well that levelsof stress hormones influence how ap-propriately animals and people behavein the face of fear. (This effect may part-ly be mediated by the hippocampus,where the concentration of cortisol re-ceptors is high.) And the likeness ofhormonal and behavioral responses inmothers and infants implies that genet-ic inheritance might predispose someindividuals to extreme fearfulness, al-though we cannot rule out the contri-bution of experience.

No one can yet say to what extent theactivity of the hypothalamic-pituitary-adrenal system controls, and is con-trolled by, other brain regions that reg-

The Neurobiology of Fear Mysteries of the Mind 81

DECREASES

INCREASES

NO EFFECT

NO EFFECT

NO EFFECT

DECREASES

NO EFFECT

NO EFFECT

DECREASES

MORPHINE(OPIATE)

NALOXONE(OPIATE BLOCKER)

DIAZEPAM(BENZODIAZEPINE)

COOING FREEZING BARKING

EFFECTS ON COOING, FREEZING AND BARKING were evaluated some yearsago for three drugs that act on neurons responsive to opiates (top two rows) or to ben-zodiazepines (bottom row). The results implied that opiate-sensitive pathways in thebrain control affiliative behaviors (those that restore closeness to the mother, as cooingoften does), whereas benzodiazepine-sensitive pathways control responses to immedi-ate threats (such as freezing and barking). Newer evidence generally supports this con-clusion but adds some complexity to the picture.

CA

RO

L D

ON

NER

Copyright 1997 Scientific American, Inc.

Page 63: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

ulate the choice of defensive behavior.We have, however, begun to identifydistinct neurochemical circuits, or sys-tems, in the brain that affect differentbehaviors. The two systems we havestudied most intensely seemed at first tohave quite separate functions. But morerecent work implies that the controlson defensive behavior are rather morecomplicated than the original analysesimplied.

We gathered our initial data threeyears ago by treating six- to 12-month-old monkeys with two different classesof neuroactive chemicals—opiates (mor-phinelike substances) and benzodiaze-pines (chemicals that include the anti-anxiety drug diazepam, or Valium). Wechose to look at opiates and benzodia-zepines because neurons that release ortake up those chemicals are abundantin the prefrontal cortex, the amygdalaand the hypothalamus. The opiates areknown to have natural, or endogenous,counterparts, called endorphins and en-kephalins, that serve as neurotransmit-ters; after the endogenous chemicals arereleased by certain neurons, they bindto receptor molecules on other nervecells and thereby increase or decreasenerve cell activity. Receptors for benzo-diazepines have been identified, but in-vestigators are still trying to isolate en-dogenous benzodiazepinelike molecules.

Selective Drug Effects

Once again, our subjects were ex-posed to the alone, no-eye-contact

and stare conditions. We delivered thedrugs before the infants were separatedfrom their mothers and then recordedthe animals’ behavior. Morphine de-creased the amount of cooing normallydisplayed in the alone and stare condi-tions. Conversely, cooing was increasedby naloxone, a compound that binds toopiate receptors but blocks the activityof morphine and endogenous opiates.Yet morphine and naloxone had no in-fluence on the frequency of stare-in-duced barking and other hostile behav-iors, nor did they influence duration offreezing in the no-eye-contact situation.We concluded that opiate-using neuralpathways primarily regulate affiliativebehaviors (such as those induced by dis-tress over separation from the mother),but those pathways seem to have littlepower over responses to direct threats.

The benzodiazepine we studied—di-azepam—produced a contrary picture.The drug had no impact on cooing, but

it markedly reduced freezing, barkingand other hostile gestures. Thus, benzo-diazepine-using pathways seemed pri-marily to influence responses to directthreats but to have little power over af-filiative behavior.

We still think the opiate and benzodi-azepine pathways basically serve theseseparate functions. Nevertheless, thesimple model we initially envisionedgrew more interesting as we investigat-ed two additional drugs: a benzodiaze-pine called alprazolam (Xanax) and acompound called beta-carboline, whichbinds to benzodiazepine receptors butelevates anxiety and typically produceseffects opposite to those of diazepamand its relatives.

When we administered alprazolam indoses that lower anxiety enough to de-crease freezing, this substance, like dia-zepam, minimized hostility in the threat-ening, stare condition. And beta-carbo-line enhanced hostility. No surprises here.Yet, unlike diazepam, these drugs mod-ulated cooing, which we had consideredto be an affiliative (opiate-controlled),not a threat-related (benzodiazepine-controlled), behavior. Moreover, boththese compounds decreased cooing. Wecannot explain the similarity of effect,but we have some ideas about whydrugs that act on benzodiazepine recep-tors might influence cooing.

It may be that, contrary to our earlyview, benzodiazepine pathways can in

fact regulate affiliative behavior. We fa-vor a second interpretation, however.Cooing displayed in the stare conditionmay not solely reflect an affiliative need(a desire for mother’s comfort); at times,it may also be an urgent, threat-inducedplea for immediate help. One behavior,then, might serve two different functionsand be controlled by different neuro-chemical pathways. (This conclusionwas strengthened for me recently, whenI tried to photograph a rhesus infantthat had become separated from itsmother in the wild—where we are nowinitiating additional studies. Its persis-tent, intense coos attracted the mother,along with a pack of protectors. Thestrategy worked: I retreated rapidly.)

More generally, our chemical studieslead us to suspect that the opiate- andbenzodiazepine-sensitive circuits bothoperate during stress; the relative degreeof activity changes with the characteris-tics of a worrisome situation. As thecontribution of each pathway is altered,so, too, are the behaviors that appear.

Exactly how neurons in the opiateand benzodiazepine pathways functionand how they might cooperate are un-clear. But one plausible scenario goes likethis: When a young monkey is separatedfrom its mother, opiate-releasing and,consequently, opiate-sensitive neuronsbecome inhibited. Such inhibition givesrise to yearning for the mother and ageneralized sense of vulnerability. This

The Neurobiology of Fear82 Mysteries of the Mind

RELAXED MOTHER (left) barely reacts to the presence of the camera-wielding author,whereas a more sensitive mother becomes frightened (right), as evinced by her “fear gri-

PHO

TOG

RA

PHS

BY

NED

H. K

ALI

N

Copyright 1997 Scientific American, Inc.

Page 64: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

reduction of activity in opiate-sensitivepathways enables motor systems in thebrain to produce cooing. When a poten-tial predator appears, neurons that se-crete endogenous benzodiazepines be-come suppressed to some degree. Thischange, in turn, leads to elevated anxi-ety and the appearance of behaviors andhormonal responses that accompanyfear. As the sense of alarm grows, mo-tor areas prepare for fight or flight. Thebenzodiazepine system may also influ-ence the opiate system, thereby alteringcooing during threatening situations.

We are now refining our model ofbrain function by testing other com-pounds that bind to opiate and benzo-diazepine receptors. We are also exam-ining behavioral responses to substances,such as the neurotransmitter serotonin,that act on other receptors. (Serotoninreceptors occur in many brain regionsthat participate in the expression offear.) And we are studying the activitiesof substances that directly control stresshormone production, including corti-cotropin-releasing hormone, which isfound throughout the brain, not solelyin the hypothalamus.

In collaboration with Richard J. Da-vidson, here at Wisconsin, Shelton andI have recently identified at least onebrain region where the benzodiazepinesystem exerts its effects. Davidson hadshown that the prefrontal cortex of theright hemisphere is unusually active in

extremely inhibited children. We there-fore wondered whether we would seethe same asymmetry in frightened mon-keys and whether drugs that reducedfear-related behavior in the animalswould dampen right frontal activity.

This time we used mild restraint as astress. As we anticipated, neuronal fir-ing rose more in the right frontal cortexthan in the left. Moreover, when we de-livered diazepam in doses we knew low-ered hostility, the drug returned the re-straint-induced electrical activity to nor-mal. In other words, the benzodiazepinesystem influences defensive behavior atleast in part by acting in the right pre-frontal cortex.

These findings have therapeutic im-plications. If human and monkey brainsdo operate similarly, our data would

suggest that benzodiazepines might bemost helpful in those adults and chil-dren who exhibit elevated electrical ac-tivity in the right prefrontal cortex. Be-cause of the potential for side effects,many clinicians are cautious about de-livering antianxiety medications to chil-dren over a long time. But administra-tion of such drugs during critical peri-ods of brain development might provesufficient to alter the course of later de-velopment. It is also conceivable thatbehavioral training could teach extreme-ly inhibited youngsters to regulate ben-zodiazepine-sensitive systems withouthaving to be medicated. Alternatively,by screening compounds that are help-ful in monkeys, investigators might dis-cover new drugs that are quite safe forchildren. As the workings of other fear-modulating neurochemical systems inthe brain are elucidated, similar strate-gies could be applied to manage thosecircuits.

Our discovery of cues that elicit threedistinct sets of fear-related behaviors inrhesus monkeys has thus enabled us togain insight into the development andregulation of defensive strategies in theseanimals. We propose that the opiate andbenzodiazepine pathways in the pre-frontal cortex, the amygdala and the hy-pothalamus play a major part in deter-mining which strategies are chosen. Andwe are currently attempting to learnmore about the ways in which these andother neural circuits cooperate with oneanother. We have therefore laid thegroundwork for deciphering the rela-tive contributions of various brain sys-tems to inordinate fear in humans. Wecan envision a time when treatments willbe tailored to normalizing the specificsignaling pathways that are disrupted ina particular child, thereby sparing thatyoungster enormous unhappiness laterin life.

The Neurobiology of Fear Mysteries of the Mind 83

Further Reading

Love in Infant Monkeys. Harry F. Harlow in Scientific American, Vol. 200, No. 6, pages68–74; June 1959.

The Ethology of Predation. Eberhard Curio. Springer-Verlag, 1976.Stress and Coping in Early Development. Jerome Kagan in Stress, Coping, and Devel-opment in Children. Edited by N. Garmezy and M. Rutter. McGraw-Hill, 1983.

Defensive Behaviors in Infant Rhesus Monkeys: Environmental Cues and Neuro-chemical Regulation. Ned H. Kalin and Steven E. Shelton in Science, Vol. 243, pages1718–1721; March 31, 1989.

Stress in the Wild. Robert M. Sapolsky in Scientific American, Vol. 262, No. 1; January1990.

Defensive Behaviors in Infant Rhesus Monkeys: Ontogeny and Context-Depen-dent Selective Expression. N. H. Kalin, S. E. Shelton and L. K. Takahashi in Child De-velopment, Vol. 62, No. 5, pages 1175–1183; October 1991.

mace.” The author hopes explorations of the neural bases for such differences in mon-keys will facilitate development of new therapies for excessively anxious human beings.

SA

Copyright 1997 Scientific American, Inc.

Page 65: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

PhantomLimbs

84 Mysteries of the Mind

The Author

RONALD MELZACK is E. P.Taylor Professor of Psychologyat McGill University and re-search director of the Pain Clin-ic at Montreal General Hospi-tal. His work on the neurophysi-ology of pain spans almost fourdecades. After obtaining hisPh.D. in psychology at McGillin 1954 and taking up fellow-ships in the U.S. and abroad, hejoined the faculty of the Mas-sachusetts Institute of Technolo-gy in 1959. There he andPatrick D. Wall began discus-sions that led to the publicationin 1965 of their now famous“gate control” theory of pain.Melzack joined the McGill fac-ulty in 1963. This is his fourtharticle for Scientific American.

Stimulation elsewhere in thebody can sometimes be felt

in a phantom limb.

CA

ROL

DO

NN

ER

Copyright 1997 Scientific American, Inc.

Page 66: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

In 1866 S. Weir Mitchell, theforemost American neurologistof his time, published his first ac-count of phantom limbs, not in a

scientific journal but in the AtlanticMonthly, as an anonymously writtenshort story. In his tale, “The Case ofGeorge Dedlow,” the protagonist losesan arm to amputation during the CivilWar. Later, he awakens in the hospitalafter, unbeknownst to him, both his legshave also been amputated.

“[I was] suddenly aware of a sharpcramp in my left leg. I tried to get atit . . .with my single arm, but, findingmyself too weak, hailed an attendant.‘Just rub my left calf,. . . if you please.’”

“‘Calf?. . .You ain’t got none, pard-ner. It’s took off.’”

Some historians have speculated thatMitchell chose to publish in the Atlan-tic as a way of testing the reaction of hispeers to the concept of phantom limbs.He feared they would not believe ampu-tated arms and legs could be felt afterthe limbs were gone.

In fact, the phenomenon of phantomlimbs is common. So is the occurrenceof terrible pain in these invisible appen-dages. Yet neither the cause of phan-toms nor the associated suffering is wellunderstood. My colleagues and I haverecently proposed explanations that areleading to fresh research into treatmentsfor the often intractable pain. The con-cepts also raise questions about basicassumptions of contemporary psychol-ogy and neuroscience.

The most extraordinary feature of

phantoms is their reality to the amputee.Their vivid sensory qualities and preciselocation in space—especially at first—make the limbs seem so lifelike that apatient may try to step off a bed onto aphantom foot or lift a cup with a phan-tom hand. The phantom, in fact, mayseem more substantial than an actuallimb, particularly if it hurts.

In most cases, a phantom arm hangsstraight down at the side when the per-son sits or stands, but it moves in per-fect coordination with other limbs dur-ing walking; that is, it behaves like anormal limb. Similarly, a phantom legbends as it should when its owner sits;it stretches out when the individual liesdown; and it becomes upright duringstanding.

Sometimes, however, the amputee issure the limb is stuck in some unusualposition. One man felt that his phan-tom arm extended straight out from theshoulder, at a right angle to the body.He therefore turned sideways wheneverhe passed through doorways, to avoidhitting the wall. Another man, whosephantom arm was bent behind him,slept only on his abdomen or on his sidebecause the phantom got in the waywhen he tried to rest on his back.

The eerie reality of phantoms is oftenreinforced by sensations that mimic feel-ings in the limb before amputation. Forexample, a person may feel a painful ul-cer or bunion that had been on a foot oreven a tight ring that had been on a fin-ger. Such individuals are not merely rec-ollecting sensations but are feeling themwith the full intensity and detail of anongoing experience. The reality of thephantom is also enhanced by wearingan artificial arm or leg; the phantomusually fills the prosthesis as a hand fitsa glove.

The sense of reality is also strength-ened by the wide range of sensations aphantom limb can have. Pressure,warmth, cold and many different kinds

of pain are common. A phantom canfeel wet (as when an artificial foot is seenstepping into a puddle). Or it can itch,which can be extremely distressing, al-though scratching the apparent site ofdiscomfort can sometimes actually re-lieve the annoyance. The person mayalso feel as if the limb is being tickled oris sweaty or prickly.

Naturally, of all the sensations inphantom limbs, pain, which as many as70 percent of amputees suffer, is themost frightening and disturbing. It isoften described asburning, cramping orshooting and can varyfrom being occasionaland mild to continuousand severe. It usuallystarts shortly after am-putation but some-times appears weeks,months or years later.A typical complaint isthat a hand is clenched,fingers bent over thethumb and digginginto the palm, so thatthe whole hand is tiredand achy. In the leg thediscomfort may be feltas a cramp in the calf. Many patientsreport that their toes feel as if they arebeing seared by a red-hot poker.

A final striking feature of phantoms,which reinforces the reality still further,is that they are experienced as a part ofoneself. That is, patients perceive themas integral parts of the body. A phantomfoot is described not only as real but asunquestionably belonging to the person.Even when the foot is felt to be danglingin the air several inches beneath thestump and unconnected to the leg, it isstill experienced as part of one’s body,and it moves appropriately with theother limbs and with the torso.

Amputation is not essential for theoccurrence of a phantom. In some acci-

Reprinted from the April 1992 issue Mysteries of the Mind 85

TYPICAL EXAMPLES of phantom limbs reported by patients are combined in this human figure. Some parts of the phantom are felt especially vividly (highlighted areas in transparent limbs). The phantom limb is perceived as perfectly real to the patient, who describes it as being in various positionsand often reports feeling pain in it.

People who have lost an arm or a leg often perceive the limb as though it were still there.

Treating the pain of these ghostly appendagesremains difficult

by Ronald Melzack

The existenceof phantoms

in peopleborn without alimb suggests

that the neural networks for

perceiving thebody are builtinto the brain.

Copyright 1997 Scientific American, Inc.

Page 67: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

dents, particularly when a rider is thrownoff a motorcycle and hits the pavement,the shoulder is wrenched forward sothat all the nerves from the arm areripped from the spinal cord, a conditionknown as a brachial plexus avulsion.The resulting phantom occupies the nowuseless true arm and is usually coordi-nated with it. But if the victim’s eyes areclosed, the phantom will remain in itsoriginal position when the real arm ismoved by someone else. Although theflesh-and-blood arm is incapable of re-sponding to stimulation, the phantomversion is usually extremely painful. Re-grettably, even surgical removal of thetrue arm has no effect on the phantomor on the pain.

Similarly, paraplegics—persons whohave had a complete break of the spinal

cord and therefore have no feeling in,or control over, their body below thebreak—often have phantom legs and oth-er body parts, including genitals. Imme-diately after an accident, the phantommay be dissociated from the real body.For instance, a person may feel as if thelegs are raised over the chest or headeven when he or she can see that they arestretched out on the road. Later, though,phantoms move in coordination withthe body, at least when the person’s eyesare open. Some paraplegics complainthat their legs make continuous cyclingmovements, producing painful fatigue,even though a patient’s actual legs arelying immobile on the bed. Phantomsare also reported by patients whose spi-nal cords are anesthetized, such as by aspinal block during labor.

The oldest explanation for phantomlimbs and their associated pain is thatthe remaining nerves in the stump,which grow at the cut end into nodulescalled neuromas, continue to generateimpulses. The impulses flow up throughthe spinal cord and parts of the thalamus(which is a central way station in thebrain) to the somatosensory areas of thecortex. These cortical areas are the pre-sumed centers for sensation in classicalconcepts of the nervous system.

On the basis of this explanation, treat-ments for pain have attempted to halt thetransmission of impulses at every levelof the somatosensory projection system.The nerves from the stump have beencut, usually just above the neuroma orat the roots—small bundles of fibers thatarise when the sensory nerves divide intosmaller branches, just before they enterthe spinal cord. Pathways within the spi-nal cord have been cut as well, and theareas of the thalamus and cortex thatultimately receive sensory informationfrom the limb have been removed.

Although these approaches may pro-vide relief for months or even years, thepain usually returns. Moreover, none ofthese procedures abolish the phantomlimb itself. Hence, neuroma activity can-not by itself account either for the phe-nomenon of the phantom limb or for thesuffering.

A related hypothesis moves the sourceof phantom limbs from neuromas to thespinal cord, suggesting that phantomsarise from excessive, spontaneous firingof spinal cord neurons that have losttheir normal sensory input from thebody. The output of the cells is transmit-ted to the cortex, just as if the spinalneurons had received external stimula-tion. This proposal grew in part out ofresearch done in the 1960s showingthat after sensory nerves in the body arecut, neurons in the spinal cord sponta-neously generate a high level of electri-cal impulses, often in an abnormal,bursting pattern.

Other observations indicate that thisexplanation is insufficient. Paraplegicswho have suffered a complete break ofthe spinal cord high in the upper bodysometimes feel severe pain in the legsand groin. Yet the spinal neurons thatcarry messages from those areas to thebrain originate well below the level ofthe break, which means that any nerveimpulses arising in those neurons wouldnot traverse the break.

Some recent work has led to the pro-posal that phantom limbs can arise still

Phantom Limbs86 Mysteries of the Mind

SOMATOSENSORY CORTEX

THALAMUS

LIMBIC SYSTEM

MEDIAL PATHWAYS

LATERAL PATHWAYS

MIDBRAIN

INHIBITORY PATHWAY

SENSORY INPUT FROM STUMP

NEUROMA SPINAL CORD

CA

RO

L D

ON

NER

PATHWAYS OF SIGNALS from the body to the brain are shown. After the loss of alimb, nerve cells in the denervated areas of the spinal cord and brain fire spontaneous-ly at high levels and with abnormal bursting patterns.

Copyright 1997 Scientific American, Inc.

Page 68: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

higher in the central nervous system—

in the brain itself. One hypothesis holdsthat phantoms are caused by changes inthe flow of signals through the somato-sensory circuit in the brain.

For example, Frederick A. Lenz, thenat the University of Toronto, observedabnormally high levels of activity and abursting pattern in cells of the thalamusin a paraplegic patient who had a fullbreak of the spinal cord just below theneck but nonetheless suffered pain inthe lower half of his body. The overac-tive cells, it turned out, also respondedto touches the head and neck, eventhough the cells were in the area of thethalamus that normally responds onlyto stimulation of the body below thelevel of the cut. This finding suggestedthat neural inhibition was lifted on theflow of signals across existing but pre-viously unused synapses in sensory neu-rons projecting to the thalamus fromthe head and neck.

Such changes in the somatosensorythalamus or cortex could help explainwhy certain feelings arise in limbs thatno longer exist or can no longer trans-

mit signals to the brain. Nevertheless, al-terations in this system cannot by them-selves account for phantoms and theirpain. If this explanation were sufficient,removal of the affected parts of the so-matosensory cortex or thalamus wouldsolve both problems.

Clearly, the source of phantom limbsis more complex than any of these the-ories would suggest. No other hypothe-ses have been proposed, however. As anoutgrowth of my interest in the brainmechanisms that give rise to pain, I havepondered the causes of phantoms andphantom-limb pain and studied patientswith these problems for many years.

Self-Awareness Neuromatrix

My work and that of others haveled me to conclude that, to a

great extent, phantom limbs originatein the brain, as the work of Lenz wouldsuggest. But much more of the cere-brum than the somatosensory system isinvolved.

Any explanation must account forthe rich variety of sensations a person

can feel, the intense reality of the phan-tom and the conviction that even free-floating phantoms belong to the self. Ihave proposed such a model. It has beenwell received, but it must, of course, betested more fully before its value can beassessed completely. Meanwhile, though,it has already generated new ideas for re-search into stopping the pain that arisesfrom phantom limbs.

In essence, I postulate that the braincontains a neuromatrix, or network ofneurons, that, in addition to respondingto sensory stimulation, continuouslygenerates a characteristic pattern of im-pulses indicating that the body is intactand unequivocally one’s own. I havecalled this pattern a neurosignature. Ifsuch a matrix operated in the absenceof sensory inputs from the periphery ofthe body, it would create the impressionof having a limb even after that limbhas been removed.

To produce all the qualities I have de-scribed for phantoms, the matrix wouldhave to be quite extensive, including atleast three major neural circuits in thebrain. One of them, of course, is the clas-

Phantom Limbs Mysteries of the Mind 87

Phantom seeing and hearing, like phantom limbs, are alsogenerated by the brain in the absence of sensory input.

People whose vision has been impaired by cataracts or by theloss of a portion of the visual processing system in the brainsometimes report highly detailed visual experiences. This syn-drome was first described in 1769, when the philosopher CharlesBonnet wrote an article on the remarkable visual experiences ofhis grandfather, Charles Lullin, who had lost most of his visionbecause of cataracts but was otherwise in good physical andpsychological health. Since then, many mentally sound individu-als have reported similarly vivid phantom visual experiences.

Phantom seeing often coexists with a limited amount of nor-mal vision. The person experiencing the phantom has no diffi-culty in differentiating between the two kinds of vision. Phan-tom visual episodes appear suddenly and unexpectedly whenthe eyes are open. People usually describe the visual phantomsas seeming real despite the obvious impossibility of their exis-tence. Common phantom images include people and largebuildings. Rarer perceptions include miniature people and smallanimals. Phantom sights are not mere memories of earlier expe-riences; they often contain events, places or people that havenever before been encountered.

First appearances of phantom images can be quite startling. Awoman in one of our studies who had lost much of her vision be-cause of retinal degeneration reported being shocked when shelooked out a window and saw a tall building in what she knew tobe a wooded field. Even though she realized that the buildingwas a phantom, it seemed so real that she could count its stepsand describe its other details. The building soon disappeared,only to return several hours later. The phantom vision continues

to come and go unexpectedly, she explained to my student Geoffrey Schultz.

Phantom seeing occurs most among the elderly, presumablybecause vision tends to deteriorate with age. Some 15 percent ofthe people who lose all or part of their vision report phantom vi-sual experiences. The proportion may be higher because somepeople avoid discussing phantom vision for fear of being labeledas psychologically disturbed.

Phantom sounds are also extremely common, although fewpeople recognize them for what they are. People who lose theirhearing commonly report noises in their heads. These noises,called tinnitus, are said to sound like whistling, clanging, screech-ing or the roaring of a train. They can be so loud and unpleasantthat the victim needs help to cope with the distress they cause.

Some people with tinnitus report hearing “formed sounds,”such as music or voices. A woman who had been a musician

before losing her hearing says she “hears” piano concertos andsonatas. The impression is so real that at first she thought thesounds were coming from a neighbor’s radio. The woman re-ports that she cannot turn off the music and that it often getslouder at night when she wants to go to sleep. Another woman,who had lost much of her sight and hearing, experienced bothphantom sight and sound. In one instance, she described seeinga circus and hearing the music that accompanied the acts.

Phantom sights and sounds, like phantom limbs, occur whenthe brain loses its normal input from a sensory system. In the ab-sence of input, cells in the central nervous system become moreactive. The brain’s intrinsic mechanisms transform that neuronalactivity into meaningful experiences. —R.M.

Phantom Seeing and Hearing

Copyright 1997 Scientific American, Inc.

Page 69: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

sical sensory pathway passing throughthe thalamus to the somatosensory cor-tex. A second system must consist ofthe pathways leading through the retic-ular formation of the brain stem to thelimbic system, which is critical for emo-tion and motivation. I include this cir-cuit in part because I and others havenoted that paraplegics who suffer acomplete spinal break high in the upperbody continue to experience themselvesas still being in their old body, and theydescribe the feelings in the denervatedareas with the same kinds of affectiveterms as they did before they were in-jured, such as “painful,” “pleasurable”or “exhausting.”

A final system consists of cortical re-gions important to recognition of theself and to the evaluation of sensory sig-nals. A major part of this system is theparietal lobe, which in studies of brain-damaged patients has been shown to beessential to the sense of self.

Indeed, patients who have suffered alesion of the parietal lobe in one hemi-sphere have been known to push one oftheir own legs out of a hospital bed be-cause they were convinced it belongedto a stranger. Such behavior shows thatthe damaged area normally imparts asignal that says, “This is my body; it isa part of my self.”

I believe that when sensory signalsfrom the periphery or elsewherereach the brain, they pass througheach of these systems in parallel. Asthe signals are analyzed, informationabout them is shared among the threesystems and converted into an inte-grated output, which is sent to otherparts of the brain. Somewhere in thebrain the output is transformed intoa conscious perception, although noone knows exactly where the trans-formation that leads to awarenesstakes place.

As dynamic as this description mayseem, the processing is probably stillmore dynamic than that. I furtherpropose that as the matrix analyzessensory information, it imprints itscharacteristic neurosignature on theoutput. Thus, the output carries in-formation about sensory input aswell as the assurance that the sensa-tion is occurring in one’s own body.The neurosignature may be likenedto the basic theme of an orchestralpiece. The collective sound changeswhen different instruments play theirparts (the input), but the product iscontinually shaped by the underly-

ing theme (the neurosignature), whichprovides the continuity for the work,even as the details of its rendition change.

Genetically Prewired Matrix

The specific neurosignature of an in-dividual would be determined by

the pattern of connectivity among neu-rons in the matrix—that is, by such fac-tors as which neurons are connected toone another and by the number, typesand strengths of the synapses. Readersfamiliar with neuroscience will note thatmy conception of the neuromatrix hassimilarities to the notion of the cell as-sembly proposed long ago by Donald O.Hebb of McGill University. Hebb ar-gued that when sensory input activatestwo brain cells simultaneously, synapsesbetween the cells form stronger connec-tions. Eventually the process gives riseto whole assemblies of linked neurons,so that a signal going into one part of anassembly spreads through the rest, evenif the assembly extends across broad ar-eas of the brain.

I depart from Hebb, however, in thatI visualize the neuromatrix as an assem-bly whose connections are primarilydetermined not by experience but by thegenes. The matrix, though, could later

be sculpted by experience, which wouldadd or delete, strengthen or weaken,existing synapses. For instance, experi-ence would enable the matrix to storethe memory of a pain from a gangren-ous ulcer and might thus account forthe frequent reappearance of the samepain in phantom limbs.

I think the matrix is largely prewired,for the simple reason that my colleaguesand I have encountered many peoplewho were born without an arm or a legand yet experience a vivid phantom. Forexample, an intelligent and serious eight-year-old boy, who was born with para-lyzed legs and a right arm that ends atthe elbow, tells us that when he fits hiselbow into a small cup so as to manipu-late a lever that allows him to move hiswheelchair, phantom fingers, “like ev-eryone else’s fingers,” emerge from hiselbow and grasp the edges of the cup.Phantoms such as these may persist intoadulthood: a 32-year-old engineer whowas born without a leg below the kneereports that his phantom leg and footremain vivid but vanish for several hoursonce or twice a week. He reports thathe is always astonished and delightedwhen they return.

Parenthetically, I should note that thelong-held belief that phantoms are ex-

perienced only when an amputationhas occurred after the age of six orseven is not true. My postdoctoralstudent Renée Lacroix and I haveconfirmed earlier reports that chil-dren who lose a limb when they areas young as one or two years old canhave phantom limbs. We have alsoencountered children who have pain-ful phantoms of legs that were lostbefore age two.

Under normal circumstances, then,the myriad qualities of sensation peo-ple experience emerge from varia-tions in sensory input. This input isboth analyzed and shaped into com-plex experiences of sensation and selfby the largely prewired neuromatrix.Yet even in the absence of externalstimuli, much the same range of ex-periences can be generated by othersignals passing through the neuro-matrix—such as those produced bythe spontaneous firing of neurons inthe matrix itself or the spinal cord orproduced by neuromas. Regardlessof the source of the input to the ma-trix, the result would be the same:rapid spread of the signals through-out the matrix and perception of alimb that is located within a unitary

Phantom Limbs88 Mysteries of the Mind

REFERRED SENSATIONS in a painful phan-tom arm were reported by a woman receivingelectrical stimuli at two points (dots). Stimula-tion at the stump gave the sensation of electricshocks that jumped from finger to finger. Stimu-lation on the right ear made the left phantom el-bow feel warm and caused a pulsation thattraveled down the phantom wrist and thumb.The observations were made by Joel Katz, nowat the University of Toronto, and the author.

CA

ROL

DO

NN

ER

Copyright 1997 Scientific American, Inc.

Page 70: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

self, even when the actual limb is gone.The fading of phantom limbs and their

pain, which sometimes occurs over time,would be explained if cerebral neuronsthat once responded to lost or paralyzedlimbs develop increasingly strong con-nections with still sensate parts of thebody and then begin to serve those re-gions. In the process the neurosignaturepattern would change, resulting inchanges in the phantom and the pain.But phantoms do not usually disappearforever. In fact, they may return decadesafter they seem to have gone, which in-dicates that the neuromatrix, even whenmodified, retains many of its featurespermanently.

My students Anthony L. Vaccarino,

John E. McKenna and Terence J. Co-derre and I have already gathered somedirect evidence supporting my sugges-tion that the brain—and by implication,the neuromatrix—can generate sensa-tion on its own. Our studies relied onwhat is called the formalin pain test.

We injected a dilute solution of for-malin (formaldehyde dissolved in wa-ter) under the skin of a rat’s paw, whichproduces pain that rapidly rises and fallsin intensity during the first five minutesafter the injection. (The degree and du-ration of discomfort are assessed by suchbehaviors as licking the paw.) This “ear-ly” response is followed by “late” pain,which begins about 15 minutes after theinjection and persists for about an hour.

By means of this test, we found thatan anesthetic block of the paw complete-ly obliterates the late pain, but only if theanesthetic is delivered in time to preventthe early response. Once the early painoccurs, the drug only partly reduces thelater response. This observation of paincontinuing even after the nerves carry-ing pain signals are blocked implies thatlong-lasting pain (such as that in phan-toms) is determined not only by sensorystimulation during the discomfort butalso by brain processes that persist with-out continual priming.

Phantom-Limb Pain

But what exactly causes the pain inphantom limbs? The most common

complaint is a burning sensation. Thisfeeling could stem from the loss of sen-sory signaling from the limb to the neu-romatrix. Without its usual sensory stim-ulation, the neuromatrix would proba-bly produce high levels of activity in abursting pattern, such as Lenz observedin the thalamus. This kind of signal mayvery well be transformed into an aware-ness of burning.

Other pain may result from the effortof the neuromatrix to make the limbsmove as they normally would. Whenthe limbs do not respond in amputeesand paraplegics, the neuromatrix (whichwould be prewired to “assume” thelimbs can indeed move) may issue morefrequent and stronger messages urgingthe muscles to move the limb. Theseoutputs may be perceived as cramping.Similar output messages might also befelt as shooting pain.

Research to test some of these ideasand explore new ways of eliminatingpain is still in its infancy, but some in-triguing results are beginning to emerge.The need for such treatments is urgent,both because the suffering can be severeand persistent and because, sadly, fewmethods are permanently effective.

At the moment, a number of differenttherapies are used. Stimulation of thestump with electric currents, a vibratoror acupuncture helps some amputees.Relaxation and hypnosis aid others.Some individuals obtain considerable re-lief from drugs that are usually given tocounteract epilepsy or depression, andother patients find their pain is eased bya combination of an antidepressant anda narcotic (such as methadone). Butabout half of those with persistent, long-term phantom pain fail to respond toany approach.

Phantom Limbs90 Mysteries of the Mind

REAL ARM made insensate by an inflated pressure cuff resembles a phantom arm.The subject could not see the arm, because the table was covered by a black cloth. Thepositions of the hand felt before the cuff was inflated and at intervals thereafter, as thehand seemed to be closer to the body, are shown. This study was carried out with YigalGross, now at Bar-Ilan University in Israel.

AFTER40 MINUTES

AFTER30 MINUTES

BEFORE CUFF INFLATION

CA

RO

L D

ON

NER

Copyright 1997 Scientific American, Inc.

Page 71: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

On a more promising note, an exper-imental treatment called the DREZ (dor-sal root entry zone) procedure selective-ly abolishes phantom-limb pain, but notthe phantoms themselves, in about 60percent of the patients treated. In thismethod, developed by Blaine S. Nash-old of Duke University, neurosurgeonsdestroy the spinal cells that receive inputdirectly from the sensory nerves of thestump, specifically eliminating the cellsat the site where the sensory roots enterthe spinal cord. (Past efforts at dampen-ing the somatosensory projection systemgenerally cut the sensory roots or thetransmission pathways in the spinalcord.) The DREZ procedure is so newthat no one yet knows how long the re-lief persists.

Because my model of brain function-

ing posits that the neuromatrixas a whole may contribute topain, the model also suggeststhat altering the activity of path-ways outside the somatosenso-ry system might be important,either alone or in combinationwith other treatments. Oneplace to begin work is the lim-bic system. Until now, limbicstructures have been relegatedto a secondary role in efforts totreat pain, because injuriousstimuli do not activate them di-rectly. Nevertheless, if the lim-bic system contributes to out-put by the neuromatrix, as Ihave proposed, it might wellcontribute to the pain felt inphantom limbs.

Vaccarino, McKenna, Coderreand I have begun to test the val-ue of manipulating the limbicsystem as a way of easing pain.We have shown that localizedinjection of lidocaine (a relativeof cocaine that prevents neu-rons from transmitting signals)into diverse areas of the limbicsystem produces striking de-creases in several types of ex-perimentally produced pain inrats, including a model of phan-tom-limb pain. A similar ap-proach could be feasible for re-

lieving phantom-limb pain in humansbut needs more study.

The phenomenon of phantom limbsis more than a challenge to medical

management. It raises doubts aboutsome fundamental assumptions in psy-chology. One such assumption is thatsensations are produced only by stimuliand that perceptions in the absence ofstimuli are psychologically abnormal.Yet phantom limbs, as well as phantomseeing and hearing, indicate this notionis wrong. The brain does more than de-tect and analyze inputs; it generatesperceptual experience even when no ex-ternal inputs occur. We do not need abody to feel a body.

The Brain’s Body Image

Another entrenched assumption is that perception of one’s body re-

sults from sensory inputs that leave amemory in the brain; the total of thesesignals becomes the body image. But theexistence of phantoms in people bornwithout a limb or who have lost a limbat an early age suggests that the neuralnetworks for perceiving the body andits parts are built into the brain. Theabsence of inputs does not stop the net-works from generating messages aboutmissing body parts; they continue toproduce such messages throughout life.

In short, phantom limbs are a mysteryonly if we assume the body sends sen-sory messages to a passively receivingbrain. Phantom limbs become compre-hensible once we recognize that thebrain generates the experience of thebody. Sensory inputs merely modulatethat experience; they do not directlycause it.

Phantom Limbs Mysteries of the Mind 91

Further Reading

Body Image: Dissociation of Real and Perceived Limbs by Pressure-Cuff Is-chemia. Y. Gross and R. Melzack in Experimental Neurology, Vol. 61, No. 3, pages680–688; September 15, 1978.

Phantom Limbs, the Self and the Brain: The D. O. Hebb Memorial Lecture. R.Melzack in Canadian Psychology, Vol. 30, No. 1, pages 1–16; January 1989.

Central Nervous System Plasticity in the Tonic Pain Response to SubcutaneousFormalin Injection. T. J. Coderre, A. L. Vaccarino and R. Melzack in Brain Research,Vol. 535, No. 1, pages 155–158; December 3, 1990.

Pain “Memories” in Phantom Limbs: Review and Clinical Observations. J. Katzand R. Melzack in Pain, Vol. 43, No. 3, pages 319–336; December 1990.

The Role of the Cingulum Bundle in Self-Mutilation following PeripheralNeurectomy in the Rat. A. L. Vaccarino and R. Melzack in Experimental Neurology,Vol. 111, No. 1, pages 131–134; January 1991.

CA

ROL

DO

NN

ERa SOMATOSENSORY SYSTEM

PARIETAL CORTEX

POSTERIORTHALAMUS

b LIMBIC SYSTEMCINGULATEGYRUS

RAPHE NUCLEI

THALAMUS

AMYGDALA

SOURCE OF PHANTOM LIMBS is thought by the author to involve activityin three of the brain’s neural circuits. One of them (a) is the somatosensory re-ceiving areas and the adjacent parietal cortex, which process information re-lated to the body. The second area (b) is the limbic system, which is concernedwith emotion and motivation. The third (c) encompasses the widespread corti-cal networks involved in cognitive activities, among them the memory of pastexperience and the evaluation of sensory inputs in relation to the self.

c COGNITIVE SYSTEM

ANTERIORCINGULATEGYRUS

FRONTALCORTEX

TEMPORAL CORTEX

PARIETALCORTEX

POSTERIORCINGULATE

GYRUS

VISUALCORTEX

SA

Copyright 1997 Scientific American, Inc.

Page 72: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

The image often invoked to de-scribe autism is that of a beau-tiful child imprisoned in a glassshell. For decades, many par-

ents have clung to this view, hoping that oneday a means might be found to break the in-visible barrier. Cures have been proclaimed,but not one of them has been backed by ev-idence. The shell remains intact. Perhaps thetime has come for the whole image to beshattered. Then at last we might be able tocatch a glimpse of what the minds of autis-tic individuals are truly like.

Psychological and physiological researchhas shown that autistic people are not liv-ing in rich inner worlds but instead are vic-tims of a biological defect that makes theirminds very different from those of normalindividuals. Happily, however, autistic peo-ple are not beyond the reach of emotionalcontact and attachment to others.

Thus, we can make the world more hos-pitable for autistic individuals just as we can,say, for the blind. To do so, we need to un-derstand what autism is like—a most chal-lenging task. We can imagine being blind,but autism seems unfathomable. For cen-turies, we have known that blindness is of-ten a peripheral defect at the sensory-motorlevel of the nervous system, but only recent-ly has autism been appreciated as a centraldefect at the highest level of cognitive pro-cessing. Autism, like blindness, persiststhroughout life, and it responds to specialefforts in compensatory education. It cangive rise to triumphant feats of coping butcan also lead to disastrous secondary conse-quences—anxiety, panic and depression.Much can be done to prevent problems. Un-derstanding the nature of the handicap mustbe the first step in any such effort.

Autism existed long before it was de-scribed and named by Leo Kanner of theJohns Hopkins Children’s Psychiatric Clin-ic. Kanner published his landmark paper in1943 after he had observed 11 children

who seemed to him to form a recognizablegroup. All had in common four traits: apreference for aloneness, an insistence onsameness, a liking for elaborate routinesand some abilities that seemed remarkablecompared with the deficits.

Concurrently, though quite independent-ly, Hans Asperger of the University PediatricClinic in Vienna prepared his doctoral the-sis on the same type of child. He also usedthe term “autism” to refer to the core fea-tures of the disorder. Both men borrowedthe label from adult psychiatry, where it hadbeen used to refer to the progressive loss ofcontact with the outside world experiencedby schizophrenics. Autistic children seemedto suffer such a lack of contact with theworld around them from a very early age.

Kanner’s first case, Donald, has long servedas a prototype for diagnosis. It had been ev-ident early in life that the boy was differentfrom other children. At two years of age, hecould hum and sing tunes accurately frommemory. Soon he learned to count to 100and to recite both the alphabet and the 25questions and answers of the Presbyteriancatechism. Yet he had a mania for makingtoys and other objects spin. Instead of play-ing like other toddlers, he arranged beadsand other things in groups of different col-ors or threw them on the floor, delighting inthe sounds they made. Words for him had aliteral, inflexible meaning.

Donald was first seen by Kanner at agefive. Kanner observed that the boy paid noattention to people around him. Whensomeone interfered with his solitary activi-ties, he was never angry with the interferingperson but impatiently removed the handthat was in his way. His mother was theonly person with whom he had any signifi-cant contact, and that seemed attributablemainly to the great effort she made to shareactivities with him. By the time Donald wasabout eight years old, his conversation con-sisted largely of repetitive questions. His re-lation to people remained limited to his im-mediate wants and needs, and his attemptsat contact stopped as soon as he was toldor given what he had asked for.

Some of the other children Kanner de-scribed were mute, and he found that eventhose who spoke did not really communicatebut used language in a very odd way. Forexample, Paul, who was five, would parrotspeech verbatim. He would say “You want

AutismAutistic people suffer from a biological defect. Although they cannot be cured, much can be done to improve their lives

by Uta Frith

92 Mysteries of the Mind Reprinted from the June 1993 issue

The Author

UTA FRITH is a senior scien-tist in the Cognitive Develop-ment Unit of the Medical Re-search Council in London.Born in Germany, she took adegree in psychology in 1964at the University of the Saar-land in Saarbrücken, whereshe also studied the history ofart. Four years later she ob-tained her Ph.D. in psycholo-gy at the University of Lon-don. Besides autism, her in-terests include reading devel-opment and dyslexia. She hasedited a book in the field ofreading development, Cog-nitive Processes in Spelling,and is the author of Autism:Explaining the Enigma.

CHARACTERISTIC ALONENESS of autistic children is exhibited by a boy atthe Association in Manhattan for Autistic

Children, Inc. All the accompanying photographs were taken there.RO

DIC

A P

RATO

The “glass shell” metaphor for autism is misleading.

Copyright 1997 Scientific American, Inc.

Page 73: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

candy” when he meant “I want candy.”He was in the habit of repeating, al-most every day, “Don’t throw the dogoff the balcony,” an utterance his moth-er traced to an earlier incident with atoy dog.

Twenty years after he had first seenthem, Kanner reassessed the membersof his original group of children. Someof them seemed to have adapted sociallymuch better than others, although theirfailure to communicate and to form re-lationships remained, as did their ped-antry and single-mindedness. Two pre-requisites for better adjustment, though

no guarantees of it, were the presenceof speech before age five and relativelyhigh intellectual ability. The brightestautistic individuals had, in their teens,become uneasily aware of their pecu-liarities and had made conscious effortsto conform. Nevertheless, even the bestadapted were rarely able to be self-reli-ant or to form friendships. The one cir-cumstance that seemed to be helpful inall the cases was an extremely struc-tured environment.

As soon as the work of the pioneersbecame known, every major clinic be-gan to identify autistic children. It was

found that such children, in addition totheir social impairments, have substan-tial intellectual handicaps. Althoughmany of them perform relatively wellon certain tests, such as copying mosaicpatterns with blocks, even the most abletend to do badly on test questions thatcan be answered only by the applica-tion of common sense.

Autism is rare. According to the strictcriteria applied by Kanner, it appears infour of every 10,000 births. With thesomewhat wider criteria used in currentdiagnostic practice, the incidence ismuch higher: one or two in 1,000 births,

ABR

AH

AM

MEN

ASH

E

Copyright 1997 Scientific American, Inc.

Page 74: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

about the same as Down’s syndrome.Two to four times as many boys as girlsare affected.

For many years, autism was thoughtto be a purely psychological disorderwithout an organic basis. At first, no ob-vious neurological problems were found.The autistic children did not necessarilyhave low intellectual ability, and theyoften looked physically normal. For thesereasons, psychogenic theories were pro-

posed and taken seriously for manyyears. They focused on the idea that achild could become autistic because ofsome existentially threatening experi-ence. A lack of maternal bonding or adisastrous experience of rejection, so thetheory went, might drive an infant towithdraw into an inner world of fantasythat the outside world never penetrates.

These theories are unsupported by anyempirical evidence. They are unlikely to

be supported because there are many in-stances of extreme rejection and depri-vation in childhood, none of which haveresulted in autism. Unfortunately, ther-apies vaguely based on such notions arestill putting pressure on parents to ac-cept a burden of guilt for the supposed-ly avoidable and reversible breakdownof interpersonal interactions. In con-trast, well-structured behavior modifica-tion programs have often helped fami-

Autism94 Mysteries of the Mind

ROD

ICA

PR

ATO

Autistic Behavior

The traits most characteristic ofautistic people are aloneness,

an insistence on sameness and aliking for elaborate routines. At thesame time, some autistic individu-als can perform complicated tasks,provided that the activity does notrequire them to judge what someother person might be thinking.These traits lead to characteristicforms of behavior, a number ofwhich are portrayed here. —U.F.

Displays indifference

Indicates needs by using an adult’s hand

Parrots words

Laughs and giggles inappropriately

Does not make eye contact Does not pretend in playing Prefers samenessYet some do certain things well if the task

does not involve social understanding.

Does not play with other children

Handles or spins objects

Behaves in bizarre ways

Talks incessantly about one topic

Is one-sided in interactions

Joins in only if an adult insists and assists

Copyright 1997 Scientific American, Inc.

Page 75: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

lies in the management of autistic chil-dren, especially children with severe be-havior problems. Such programs donot claim to reinstate normal develop-ment in the children.

The insupportability of the psycho-genic explanation of autism led a num-ber of workers to search for a biologicalcause. Their efforts implicate a defectivestructure in the brain, but that structurehas not yet been identified. The defect isbelieved to affect the thinking of autis-tic people, making them unable to eval-uate their own thoughts or to perceiveclearly what might be going on insomeone else’s mind.

Autism appears to be closely associ-ated with several other clinical and med-ical conditions. They include maternalrubella and chromosomal abnormality,as well as early injury to the brain andinfantile seizures. Most impressive, per-haps, are studies showing that autismcan have a genetic basis. Both identicaltwins are much more likely to be autis-tic than are both fraternal twins. More-over, the likelihood that autism will oc-cur twice in the same family is 50 to 100times greater than would be expectedby chance alone.

Defect in Frontal Lobes

Structural abnormalities in the brainsof autistic individuals have turned

up in anatomic studies and brain-imag-ing procedures. Both epidemiologicaland neuropsychological studies havedemonstrated that autism is stronglycorrelated with mental retardation,which is itself clearly linked to physio-logical abnormality. This fact fits wellwith the idea that autism results from adistinct brain abnormality that is oftenpart of more extensive damage. If theabnormality is pervasive, the mental re-tardation will be more severe, and thelikelihood of damage to the critical brainsystem will increase. Conversely, it is pos-sible for the critical system alone to bedamaged. In such cases, autism is notaccompanied by mental retardation.

Neuropsychological testing has alsocontributed evidence for the existenceof a fairly circumscribed brain abnor-mality. Autistic individuals who areotherwise able show specific and exten-sive deficits on certain tests that involveplanning, initiative and spontaneousgeneration of new ideas. The same defi-cits appear in patients who have frontallobe lesions. Therefore, it seems plausi-ble that whatever the defective brain

structure or system is, the frontal lobesare implicated.

Population studies carried out byLorna Wing and her colleagues at theMedical Research Council’s Social Psy-chiatry Unit in London reveal that thedifferent symptoms of autism do notoccur together simply by coincidence.Three core features in particular—im-pairments in communication, imagina-tion and socialization—form a distincttriad. The impairment in communica-tion includes such diverse phenomena asmuteness and delay in learning to talk,

as well as problems in comprehendingor using nonverbal body language. Oth-er autistic individuals speak fluently butare overliteral in their understanding oflanguage. The impairment in imagina-tion appears in young autistic childrenas repetitive play with objects and insome autistic adults as an obsessive in-terest in facts. The impairment in so-cialization includes ineptness and inap-propriate behavior in a wide range of re-ciprocal interactions, such as the abilityto make and keep friends. Nevertheless,many autistic individuals prefer to havecompany and are eager to please.

The question is why these impair-ments, and only these, occur together.The challenge to psychological theoristswas clear: to search for a single cogni-tive component that would explain thedeficits yet still allow for the abilities thatautistic people display in certain aspectsof interpersonal interactions. My col-leagues at the Medical Research Coun-cil’s Cognitive Development Unit inLondon and I think we have identifiedjust such a component. It is a cognitivemechanism of a highly complex andabstract nature that could be describedin computational terms. As a shorthand,one can refer to this component by oneof its main functions, namely, the abilityto think about thoughts or to imagineanother individual’s state of mind. Wepropose that this component is dam-aged in autism. Furthermore, we suggestthat this mental component is innate andhas a unique brain substrate. If it werepossible to pinpoint that substrate—

whether it is in fact an anatomical struc-ture, a physiological system or a chemi-cal pathway—one might be able toidentify the biological origin of autism.

The power of this component in nor-mal development becomes obvious veryearly. From the end of the first year on-ward, infants begin to participate inwhat has been called shared attention.For example, a normal child will pointto something for no reason other thanto share his interest in it with someoneelse. Autistic children do not showshared attention. Indeed, the absence ofthis behavior may well be one of theearliest signs of autism. When an autis-tic child points at an object, it is onlybecause he wants it.

In the second year of life, a particu-larly dramatic manifestation of the crit-ical component can be seen in normalchildren: the emergence of pretense, orthe ability to engage in fantasy and pre-tend play. Autistic children cannot un-

Autism Mysteries of the Mind 95

IQRANGE

PROPORTION AFFECTEDWITH AUTISM

0–19

86 PERCENTOF 44 CHILDREN

42 PERCENTOF 96 CHILDREN

2 PERCENT OF700 CHILDREN

0.013 PERCENT OF34,100 CHILDREN

20–49

50–69

70+

Autism andMental Retardation

SOURCE: Lorna Wing, Medical Research Council, London

CLOSE LINK between autism and men-tal retardation is reflected in this chart.The percentage of children showing thesocial impairments typical of autism ishighest at low levels of intelligence asmeasured by tests in which an intelligencequotient (IQ) below 70 is subnormal. Forexample, 86 percent of 44 children in thelowest IQ range showed the social im-pairments of autism. The data are drawnfrom a population of about 35,000 chil-dren aged under 15 years.

JARE

D S

CH

NEI

DM

AN

DES

IGN

Copyright 1997 Scientific American, Inc.

Page 76: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

derstand pretense and do not pretendwhen they are playing. The differencecan be seen in such a typical nurserygame as “feeding” a teddy bear or a dollwith an empty spoon. The normal childgoes through the appropriate motionsof feeding and accompanies the actionwith appropriate slurping noises. Theautistic child merely twiddles or flicksthe spoon repetitively. It is precisely theabsence of early and simple communi-cative behaviors, such as shared atten-tion and make-believe play, that oftencreates the first nagging doubts in theminds of the parents about the develop-ment of their child. They rightly feelthat they cannot engage the child in theemotional to-and-fro of ordinary life.

My colleague Alan M. Leslie deviseda theoretical model of the cognitivemechanisms underlying the key abilitiesof shared attention and pretense. Hepostulates an innate mechanism whosefunction is to form and use what wemight call second-order representations.The world around us consists not onlyof visible bodies and events, capturedby first-order representations, but alsoof invisible minds and mental events,which require second-order representa-tion. Both types of representation haveto be kept in mind and kept separatefrom each other.

Second-order repre-sentations serve tomake sense of other-wise contradictory orincongruous informa-tion. Suppose a normalchild, Beth, sees hermother holding a ba-nana in such a way asto be pretending that itis a telephone. Beth hasin mind facts about ba-nanas and facts abouttelephones—first-orderrepresentations. Nev-ertheless, Beth is notthe least bit confusedand will not start eat-ing telephones or talk-ing to bananas. Con-fusion is avoided be-cause Beth computesfrom the concept of

pretending (a second-order representa-tion) that her mother is engaging simul-taneously in an imaginary activity anda real one.

As Leslie describes the mental process,pretending should be understood as com-puting a three-term relation between anactual situation, an imaginary situationand an agent who does the pretending.The imaginary situation is then nottreated as the real situation. Believingcan be understood in the same way aspretending. This insight enabled us topredict that autistic children would notbe able to understand that someone canhave a mistaken belief about the world.

Together with our colleague SimonBaron-Cohen, we tested this predictionby adapting an experiment originallydevised by two Austrian developmentalpsychologists, Heinz Wimmer and Jo-sef Perner. The test has become knownas the Sally-Anne task. Sally and Anneare playing together. Sally has a marblethat she puts in a basket before leavingthe room. While she is out, Anne movesthe marble to a box. When Sally returns,wanting to retrieve the marble, she ofcourse looks in the basket. If this sce-nario is presented as, say, a puppet showto normal children who are four yearsof age or more, they understand that Sal-ly will look in the basket even thoughthey know the marble is not there. Inother words, they can represent Sally’serroneous belief as well as the true stateof things. Yet in our test, 16 of 20 autis-tic children with a mean mental age ofnine failed the task—answering thatSally would look in the box—in spite of

being able to answer correctly a varietyof other questions relating to the factsof the episode. They could not concep-tualize the possibility that Sally believedsomething that was not true.

Many comparable experiments havebeen carried out in other laboratories,largely confirming our prediction: au-tistic children are specifically impairedin their understanding of mental states.They appear to lack the innate compo-nent underlying this ability. This com-ponent, when it works normally, hasthe most far-reaching consequences forhigher-order conscious processes. It un-derpins the special feature of the hu-man mind: the ability to reflect on itself.Thus, the triad of impairments in au-tism—in communication, imaginationand socialization—is explained by thefailure of a single cognitive mechanism.In everyday life, even very able autisticindividuals find it hard to keep in mindsimultaneously a reality and the fact thatsomeone else may hold a misconceptionof that reality.

The automatic ability of normal peo-ple to judge mental states enables us tobe, in a sense, mind readers. With suffi-cient experience we can form and use atheory of mind that allows us to specu-late about psychological motives forour behavior and to manipulate otherpeople’s opinions, beliefs and attitudes.Autistic individuals lack the automaticability to represent beliefs, and there-fore they also lack a theory of mind.They cannot understand how behavioris caused by mental states or how be-liefs and attitudes can be manipulated.Hence, they find it difficult to under-stand deception. The psychological un-dercurrents of real life as well as litera-ture—in short, all that gives spice to so-cial relations—for them remain a closedbook. “People talk to each other withtheir eyes,” said one observant autisticyouth. “What is it that they are saying?”

Theory of Mind

Lacking a mechanism for a theory of mind, autistic children develop quite

differently from normal ones. Mostchildren acquire more and more sophis-ticated social and communicative skillsas they develop other cognitive abilities.For example, children learn to be awarethat there are faked and genuine expres-sions of feeling. Similarly, they becomeadept at that essential aspect of humancommunication—reading between thelines. They learn how to produce and

96 Mysteries of the Mind

UNUSUAL BEHAVIOR is often dis-played by autistic individuals. Autisticchildren, for example, tend to fixate

on making toys and other objects spinand to play repetitively.

ABR

AH

AM

MEN

ASH

E

Autism

Copyright 1997 Scientific American, Inc.

Page 77: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Autism Mysteries of the Mind 97

understand humor and irony. In sum,our ability to engage in imaginativeideas, to interpret feelings and to under-stand intentions beyond the literal con-tent of speech are all accomplishments

that depend ultimately on an innate cog-nitive mechanism. Autistic children findit difficult or impossible to achieve anyof these things. We believe this is becausethe mechanism is faulty.

This cognitive explanation for autismis specific. As a result, it enables us todistinguish the types of situations inwhich the autistic person will and willnot have problems. It does not preclude

The astonishing variability in the signs and symptoms of au-tism is only beginning to be fully appreciated. Some autistic

individuals never develop speech or nonverbal communication,whereas others become fluent and can pass for normal in social in-teractions. A screening test that identifies the lack of shared atten-tion, pretend play and eye contact characteristic of autism—devel-oped by Simon Baron-Cohen of the University of Cambridge and hiscolleagues at Guys Hospital in London—appears to be remarkablysuccessful in predicting autism in children as young as 18 months.

The most severe cases of autism are associated with mental re-tardation, but IQ does not consistently correlate with abilities andspecial talents. Some studies report that up to 10 percent of theautistic population has a savant skill—exceptional ability in onearea, such as playing the piano, drawing or mathematics. Signifi-cantly, almost all savants are diagnosed as autistic.

One of the most important advances in the field has been thegrowing recognition of a subgroup of autistic individuals who pos-sess high verbal ability and develop a high degree of social aware-ness by utilizing an acquired, nonintuitive theory of mind. This vari-ant of autism is called Asperger syndrome, and some individualswho exhibit it have successful academic careers in spite of their in-terpersonal communication problems, obsessive tendencies andrestricted interests. Although autistic individuals with normal orhigher IQs can show a high degree of social adaptation, even themost compensated have some difficulty in the give and take of ev-eryday conversation and are unlikely to have intimate friends.

The theory of mind—that autisticindividuals lack the ability to under-stand the role of mental states inothers—proved to be a crucial stepin explaining how the social andcommunication deficits of autismcould coexist with good general abil-ities. This hypothesis also predictsthat there is a specific substrate orpathway in the brain that gives usthe ability to conceive of mentalstates, and recent brain imagingstudies indicate that such an areamay be located in the left medialprefrontal cortex. Yet the theory ofmind is unable to account for all as-pects of autism, such as stereotypedbehavior and the desire for same-ness or the exceptional talents pres-ent in a significant proportion ofautistic individuals. Two additionalhypotheses have been proposed.

Bruce F. Pennington of the Univer-sity of Denver and others in the U.S.,as well as James Russell and his col-leagues at the University of Cam-bridge in the U.K., have put forwardthe executive dysfunction hypothe-

sis, which proposes that autistic individuals have a deficit in execu-tive functions such as planning and working memory, impulse con-trol, and initiation and monitoring of action. The processing of ex-ecutive functions is thought to occur in the prefrontal cortex, andpoor performance of these functions is directly related to repetitivethought and stereotyped, rigid behavior in autistic individuals.

Francesca Happé of London University and I have proposed theweak central coherence hypothesis as an explanation for the ex-ceptional talents and restricted interests displayed by some autisticindividuals. Weak central coherence refers to a preference by autis-

tic individuals for segmental overholistic information processing. Howthe brain integrates information isobscure, but long-range connectionsbetween the hemispheres may wellbe involved. There is some evidencethat people with autism process in-formation in piecemeal fashion—thetotal attention of the autistic in-dividual often is captured by frag-ments or selective features usually oflittle interest to normal persons. Sur-prisingly, autistic persons tend to beless susceptible to visual illusions,perhaps because they are less affect-ed by the context in which the figureis embedded.

Because it provides a model forthe ability to conceive of mentalstates, research into autism is stimu-lating philosophical debate on self-consciousness. Future studies maylead to the identification of subcom-ponents or precursors of conscious-ness in other species, which in turnmight lead to a better understandingof the development of conscious ex-perience in humans. —U.F.

FREEHAND DRAWING by E.C., a male autistic sa-vant, was made spontaneously and without any cor-rections. Although the perspective appears realistic, itis achieved without the “vanishing points” most artistswould need. Studies by Laurent Mottron and SylvieBelleville of the University of Montreal show thatE.C.’s ability to integrate parts of visual patterns is im-paired; he is unable to reproduce anything resemblinga human face but has exceptional ability to rememberand draw individual objects and geometric shapes.

CO

UR

TESY

OF

LAU

REN

T M

OTT

RO

N

BRAIN SCANS show differences in activity between normaland autistic people. In normal persons reading a story that re-quires inferring the mental state of others, the left medial pre-frontal cortex of the brain was active (left). In persons with As-perger syndrome performing the same task, an adjacent lowerarea was active instead (right). The left medial prefrontal cortexmay be a key component of the theory of mind capability.

FRA

NC

ESC

A H

APP

É ET

AL.

Explaining Autism’s Variability

Copyright 1997 Scientific American, Inc.

Page 78: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Autism98 Mysteries of the Mind

the existence of special assets and abili-ties that are independent of the innatemechanism my colleagues and I see asdefective. Thus it is that autistic individ-uals can achieve social skills that do notinvolve an exchange between two minds.They can learn many useful social rou-tines, even to the extent of sometimescamouflaging their problems. The cog-nitive deficit we hypothesize is alsospecific enough not to preclude highachievement by autistic people in suchdiverse activities as musical performance,artistic drawing, mathematics and mem-orization of facts.

It remains to be seen how best to ex-plain the coexistence of excellent andabysmal performance by autistic peoplein abilities that are normally expectedto go together. It is still uncertain wheth-er there may be additional damage toemotions that prevents some autisticchildren from being interested in socialstimuli. We have as yet little idea whatto make of the single-minded, often ob-sessive, pursuit of certain activities. With

the autistic person, it is as if a powerfulintegrating force—the effort to seekmeaning—were missing.

Helping the Handicapped

The old image of the child in the glassshell is misleading in more ways

than one. It is incorrect to think that in-side the glass shell is a normal individu-al waiting to emerge, nor is it true thatautism is a disorder of childhood only.The motion picture Rain Man came atthe right time to suggest a new image toa receptive public. Here we see Ray-mond, a middle-aged man who is un-worldly, egocentric in the extreme andall too amenable to manipulation byothers. He is incapable of understand-ing his brother’s double-dealing pur-suits, transparently obvious though theyare to the cinema audience. Throughvarious experiences it becomes possiblefor the brother to learn from Raymondand to forge an emotional bond withhim. This is not a farfetched story. We

can learn a great deal about ourselvesthrough the phenomenon of autism.

Yet the illness should not be romanti-cized. We must see autism as a devastat-ing handicap without a cure. The autis-tic child has a mind that is unlikely todevelop self-consciousness. But we cannow begin to identify the particular typesof social behavior and emotional re-sponsiveness of which autistic individu-als are capable. Autistic people can learnto express their needs and to anticipatethe behavior of others when it is regulat-ed by external, observable factors rath-er than by mental states. They can formemotional attachments to others. Theyoften strive to please and earnestly wishto be instructed in the rules of person-to-person contact. There is no doubt thatwithin the stark limitations a degree ofsatisfying sociability can be achieved.

Autistic aloneness does not have tomean loneliness. The chilling aloofnessexperienced by many parents is not apermanent feature of their growing au-tistic child. In fact, it often gives way toa preference for company. Just as it ispossible to engineer the environmenttoward a blind person’s needs or towardpeople with other special needs, so theenvironment can be adapted to an au-tistic person’s needs.

On the other hand, one must be re-alistic about the degree of adaptationthat can be made by the limited person.We can hope for some measure of com-pensation and a modest ability to copewith adversity. We cannot expect autis-tic individuals to grow out of the unre-flecting mind they did not choose to beborn with. Autistic people in turn canlook for us to be more sympathetic totheir plight as we better understand howtheir minds are different from ours.

SELF-ABSORPTION displayed by this autistic girl is a common feature of the disor-der. In the motion picture Rain Man, self-absorption was the key trait of the centralcharacter, an autistic adult portrayed by actor Dustin Hoffman.

ABR

AH

AM

MEN

ASH

E

Further Reading

Autism: Explaining the Enigma. UtaFrith. Blackwell Publishers, 1989.

The Cognitive Basis of a BiologicalDisorder: Autism. Uta Frith, John Mor-ton and Alan M. Leslie in Trends in Neu-rosciences, Vol. 14, No. 10, pages 433-438; October 1991.

Autism and Asperger Syndrome. Editedby Uta Frith. Cambridge University Press,1992.

Understanding Other Minds: Perspec-tives from Autism. Edited by SimonBaron-Cohen, Helen Tager-Flusberg andDonald J. Cohen. Oxford University Press,1993.

SA

Copyright 1997 Scientific American, Inc.

Page 79: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

Imagine you are the father of an eight-year-old boy,” says psychologist Adri-an Raine, explaining where he believes his 17 years of research on the bio-logical basis of crime is leading. “The ethical dilemma is this: I could say toyou, ‘Well, we have taken a wide variety of measurements, and we can pre-

dict with 80 percent accuracy that your son is going to become seriously violentwithin 20 years. We can offer you a series of biological, social and cognitive inter-vention programs that will greatly reduce the chance of his becoming a violent of-fender.’ What do you do? Do you place your boy in those programs and risk stig-matizing him as a violent criminal even though there is a real possibility that he isinnocent? Or do you say no to the treatment and run an 80 percent chance thatyour child will grow up to (a) destroy his life, (b) destroy your life, (c) destroy thelives of his brothers and sisters and, most important, (d) destroy the lives of the in-nocent victims who suffer at his hands?”

For now, such a choice is purely hypothetical. Scientists cannot yet predict whichchildren will become dangerously aggressive with anything like 80 percent accura-cy. But increasingly, those who study the causes of criminal and violent behaviorare looking beyond broad demographic charac-teristics such as age, race and income level tofactors in individuals’ personality, history, envi-ronment and physiology that seem to put them—

and society—at risk. As sociologists reap thebenefits of rigorous long-term studies and neu-roscientists tug at the tangled web of relationsbetween behavior and brain chemistry, manyare optimistic that science will identify markersof maleficence. “This research might not payoff for 10 years, but in 10 years it might revolu-tionize our criminal justice system,” assertsRoger D. Masters, a political scientist at Dart-mouth College.

Preventive Intervention

With the expected advances, we’re going tobe able to diagnose many people who

are biologically brain-prone to violence,” claimsStuart C. Yudofsky, chair of the psychiatry de-partment at Baylor College of Medicine and ed-itor of the Journal of Neuropsychiatry andClinical Neurosciences. “I’m not worried aboutthe downside as much as I am encouraged bythe opportunity to prevent tragedies—to screenpeople who might have high risk and to pre-vent them from harming someone else.” Raine,Yudofsky and others argue that in order to con-trol violence, Americans should trade their tra-ditional concept of justice based on guilt andpunishment for a “medical model” based onprevention, diagnosis and treatment.

But many scientists and observers do worryabout a downside. They are concerned thatsome researchers underplay the enormous com-plexity of individual behavior and overstate sci-

102 Mysteries of the Mind Reprinted from the March 1995 issue

BRU

CE

DA

VID

SON

Mag

num

Seekingthe

CriminalElementby W. Wayt Gibbs, staff writer

Scientists are homingin on social and biological risk

factors that they believe predispose

individuals to criminal behavior.

The knowledge could be ripe withpromise—or rife

with danger

TEXAN TEENS playing with gang signs andloaded guns are acting their age—most adoles-cents dabble in delinquency for several years. Buta small fraction grow into the chronic felons thatcommit the majority of violent crimes. Can scien-tists identify the dangerous few before they at-tack—and if so, what then?

Copyright 1995 Scientific American, Inc.

Page 80: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

entists’ ability to understand and pre-dict it. They also fear that a society des-perate to reduce crime might find thetemptation to make premature use ofsuch knowledge irresistible.

Indeed, the history of science’s assaulton crime is blemished by instances inwhich incorrect conclusions were usedto justify cruel and unusual punishments.In the early 1930s, when the homiciderate was even higher than it is today, eu-genics was in fashion. “The eugenicsmovement was based on the idea thatcertain mental illness and criminal traitswere all inherited,” says Ronald L. Ak-ers, director of the Center for Studies in

Criminology and Law at the Universityof Florida. “It was based on bad science,but they thought it was good science atthe time.” By 1931, 27 states had passedlaws allowing compulsory sterilizationof “the feeble-minded,” the insane andthe habitually criminal.

Studies in the late 1960s—when crimewas again high and rising—revealed thatmany violent criminals had an extra Ychromosome and thus an extra set of“male” genes. “It was a dark day forscience in Boston when they startedscreening babies for it,” recalls XandraO. Breakefield, a geneticist at Massa-chusetts General Hospital. Subsequent

studies revealed that although XYY mentend to score lower on IQ tests, they arenot unusually aggressive.

Social science studies on the causes ofcrime have been less controversial, inpart because they have focused moreon populations than on individuals. Butas consensus builds among criminolo-gists on a few key facts, researchers areassembling these into prediction modelsthat try to identify the juveniles mostlikely to lapse into delinquency and theninto violent crime.

Perhaps their most consistent findingis that a very small number of criminalsare responsible for most of the violence.

Copyright 1995 Scientific American, Inc.

Page 81: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

One study, for example, tracked for 27years 10,000 males born in Philadelphiain 1945; it found that just 6 percent ofthem committed 71 percent of the hom-icides, 73 percent of the rapes and 69percent of the aggravated assaults at-tributed to the group.

Preventing just a small fraction of ado-lescent males from degenerating intochronic violent criminals could thus makea sizable impact on the violent crimerate, which has remained persistentlyhigh since 1973 despite a substantialdecline in property crime. (Females ac-counted for only 12.5 percent of violentcrime in 1992.) “For every 1 percent thatwe reduce violence, we save the country$1.2 billion,” Raine asserts.

The problem, says Terrie E. Moffitt,a psychologist at the University of Wis-consin who is conducting long-term de-linquency prediction studies, is that “alot of adolescents participate in antiso-cial behavior”—87 percent, according toa survey of U.S. teens. “The vast major-ity desist by age 21,” she says. The dan-gerous few “are buried within that pop-

ulation of males trying out delinquency.How do you pick them out? Our hy-pothesis is that those who start earliestare at highest risk.”

Marion S. Forgatch of the OregonSocial Learning Center tested that hy-pothesis on 319 boys from high-crimeneighborhoods in Eugene. At the Nov-ember 1994 American Society of Crim-inology meeting, she reported her find-ings: boys who had been arrested byage 14 were 17.9 times more likely tobecome chronic offenders than thosewho had not, and chronic offenderswere 14.3 times more likely to commitviolent offenses. “This is a good way ofpredicting,” she says.

False Positive ID

Good is a relative term. For if onewere to predict that every boy in

her study who was arrested early wouldgo on to commit violent crimes, onewould be wrong more than 65 percentof the time. To statisticians, those somisidentified are known as false posi-

tives. “All of these predictors have a lotof false positives—about 50 percent onaverage,” says Akers, who recentlycompleted a survey of delinquency pre-diction models. Their total accuracy iseven lower, because the models also failto identify some future criminals.

The risk factors that Akers says re-searchers have found to be most closelyassociated with delinquency are hardlysurprising. Drug use tops the list, fol-lowed by family dysfunction, childhoodbehavior problems, deviant peers, poorschool performance, inconsistent par-ental supervision and discipline, separa-tion from parents, and poverty. Numer-ous other controlled studies have foundthat alcoholism, childhood abuse, lowverbal IQ and witnessing violent actsare also significant risk factors. Com-pared with violent behavior, however,all these experiences are exceedinglycommon. The disparity makes it verydifficult to determine which factors arecauses and which merely correlates.

The difference is important, notesMark W. Lipsey of Vanderbilt Universi-ty, because “changing a risk factor if itis not causal may have no impact,” andthe ultimate goal of prediction is to stopviolence by intervening before it begins.Unfortunately, improvements in predic-tive models do not necessarily translateinto effective intervention strategies.Lipsey recently analyzed how well some500 delinquency treatment programsreduced recidivism. “The conventionalwisdom that nothing works is justwrong,” he concludes. But he concedesthat “the net effect is modest”—on av-erage, 45 percent of program partici-pants were rearrested, versus 50 per-cent of those left to their own devices.Half of that small apparent improve-ment, he adds, may be the result of in-consistency in the methods used toevaluate the programs.

Some strategies do work better thanothers, Lipsey discovered. Behavioralprograms that concentrated on teachingjob skills and rewarding prosocial atti-tudes cut rearrest rates to about 35 per-cent. “Scared straight” and boot campprograms, on the other hand, tended toincrease recidivism slightly.

Patrick H. Tolan of the University ofIllinois at Chicago has also recentlypublished an empirical review of delin-quency programs. To Lipsey’s findingshe adds that “family interventions haverepeatedly shown efficacy for reducingantisocial behavior and appear to beamong the most promising interven-

Seeking the Criminal Element104 Mysteries of the Mind

CRIME RATES have not responded consistently to “get tough” approaches to incar-ceration. Since the early 1970s the proportion of Americans behind bars has more thantripled. Property crime (including burglary, robbery and personal larceny) has droppedabout 30 percent, but violent crime remains high.

JOH

NN

Y JO

HN

SON

YEAR1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993

RA

TE

PE

R 1

,000

PE

OP

LE140

130

120

110

100

90

80

70

60

50

40

30

20

10

0

4

3

2

1

0

0.11

0.10

0.09

0.08

0.07

BURGLARY

ALL PERSONAL CRIMES

ASSAULT

ROBBERY

RAPE

SOURCE: Bureau of Justice Statistics, U.S. Department of Justice

MURDER

INCARCERATION

Copyright 1995 Scientific American, Inc.

Page 82: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

tions to date.” According to Forgatch,two experiments in Eugene, Ore.,showed that teaching parents bettermonitoring and more consistent, lesscoercive discipline techniques reducestheir kids’ misbehavior. “We shouldmake parenting skills classes compulso-ry for high school students,” arguesRaine of the University of SouthernCalifornia.

Unfortunately, Tolan observes, familyintervention is difficult and rarely at-tempted. The most common kinds ofprograms—counseling by social work-ers, peer mediation and neighborhoodantiviolence initiatives—are hardly everexamined to see whether they producelasting benefits. “It usually is hard toimagine that a good idea put into actionby well-meaning and enlightened peo-ple cannot help,” he noted in the paper.“It may seem that any effort is betterthan nothing. Yet our review and sever-al of the more long-term and sophisti-cated analyses suggest that both ofthese assumptions can be dangerouslywrong. Not only have programs thathave been earnestly launched been in-effective, but some of our seemingly best

ideas have led to worseningof the behavior of those sub-jected to the intervention.”

Many researchers are thusfrustrated that the ViolentCrime Control and Law En-forcement Act of 1994 putsmost of its $6.1 billion forcrime prevention in untestedand controversial programs,such as “midnight basket-ball” and other after-schoolactivities. “Maybe these pro-grams will help; maybe theywon’t,” Tolan says. “No onehas done a careful evalua-tion.” The Crime Act does notinsist that grant applicantsdemonstrate or even measurethe effectiveness of their approach. Forthese and other reasons, Republicansvowed in their 1994 “Contract withAmerica” to repeal all prevention pro-grams in the Crime Act and to increasefunding for prison construction. Butthat strategy ignored research. “We doknow,” Tolan asserts, “that locking kidsup will not reduce crime and may even-tually make the problem worse.”

The failure of sociology to demon-strate conclusively effective means ofcontrolling violent crime has made someimpatient. “There is a growing recogni-tion that we’re not going to solve anyproblem in society using just one disci-pline,” says Diana Fishbein, a professorof criminology at the University of Bal-timore. “Sociological factors play a role.But they have not been able to explain

Seeking the Criminal Element Mysteries of the Mind 105

POOR PARENTAL SUPERVISION is a major risk factor forlater delinquency. These children in Philadelphia play with emp-

ty crack cocaine vials. Parent training programs have beenamong the most successful in reducing kids’ antisocial behavior.

EUG

ENE

RIC

HA

RDS

Mag

num

BRAIN OF MURDERER (left ) shows less activityin the frontal cortex (top third of image) than thebrain of a nonviolent subject of the same age andsex. In one study of 22 murderers, about 75 percenthad low frontal activity, which is believed to indi-rectly regulate aggressive impulses.

AD

RIA

N R

AIN

E A

ND

MO

NTE

S. B

UC

HSB

AU

M

Copyright 1995 Scientific American, Inc.

Page 83: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

why one person becomes violent andanother doesn’t.”

Some social scientists are looking topsychiatrists, neurologists and geneticiststo provide answers to that question,ready or not. “Science must tell us whatindividuals will or will not become crim-inals, what individuals will or will notbecome victims, and what law enforce-ment strategies will or will not work,”wrote C. Ray Jeffery, a criminologist atFlorida State University, in 1994 in theJournal of Research in Crime andDelinquency.

Biological Factors

As medical researchers have teased out a few tantalizing links between

brain chemistry, heredity, hormones,physiology and assaultive behavior,some have become emboldened. “Re-search in the past 10 years conclusivelydemonstrates that biological factors playsome role in the etiology of violence.That is scientifically beyond doubt,”Raine holds forth. The importance of

that role is still very much in doubt,however.

As with social risk factors, no biolog-ical abnormality has been shown tocause violent aggression—nor is thatlikely except in cases of extreme psychi-atric disorder. But researchers have spot-ted several unusual features, too subtleeven to be considered medical prob-lems, that tend to appear in the bodiesand brains of physically aggressive men.On average, for example, they havehigher levels of testosterone, a sex hor-mone important for building musclemass and strength, among other func-tions. James M. Dabbs, Jr., of GeorgiaState University has found in his exper-

iments with prison inmates that menwith the highest testosterone concentra-tions are more likely to have committedviolent crimes. But Dabbs emphasizesthat the link is indirect and “mediatedby many social factors,” such as higherrates of divorce and substance abuse.

“Low resting heart rate probablyrepresents the best replicated biologicalcorrelate of antisocial behavior,” Raineobserves, pointing to 14 studies thathave found that problem children andpetty criminals tend to have significant-ly lower pulses than do well-behavedcounterparts. A slower heartbeat “prob-ably reflects fearlessness and under-arousal,” Raine theorizes. “If we lack

106 Mysteries of the Mind

SINS OF THE PARENT are often visitedon the child. Delinquents are more likelyto have parents who abuse drugs or alco-hol, commit crimes or beat them. But riskfactors are generally poor predictors:most children of such parents do not be-come chronic criminals.

EUG

ENE

RIC

HA

RDS

Mag

num

The failure of expensive prison booms and welfare pro-grams to beat back the historically high violent crime rates

of the past 20 years has prepared fertile ground for new ap-proaches to crime control. Encouraged by research that tenta-tively links a few instances of antisocial aggression with biologi-cal abnormalities, some politicians and ac-tivists are turning to science, perhaps toohastily, to identify and treat those who arelikely to become dangerous.

Take the case of Everett L. “Red” Hodges, aCalifornia oilman who has spent more than$1 million to support research that impli-cates the trace metal manganese as a mark-er for violent criminal behavior. Hodges wasstruggling to tame a delinquent son in 1984when he came across a Science News storyon a study that had found high levels of lead,cadmium and copper in the head hair of vio-lent felons.

Intrigued, Hodges offered funding toLouis A. Gottschalk, a psychiatrist at the Uni-versity of California at Irvine, to conduct a

controlled study to replicate the results. Analysis of hair clippedfrom convicted and accused felons at a prison and two countyjails in southern California revealed no unusual levels of lead,cadmium or copper. But Gottschalk did find that average levelsof manganese were about 3.6 times higher in the alleged felons

than in men of similar age and race at localbarbershops. “A new paradigm is opening incriminal justice,” Hodges says, beaming. “It’sa marker.”

That judgment may be premature. Criticsof Gottschalk’s research, published in 1991 ina psychiatric (rather than a nutrition) journal,point out that average manganese levelsvaried from 2.2 parts per million in the pris-oners to just 0.71 in one of the groups of jailinmates. Previous studies had found lowermanganese levels in inmates than in controlsubjects. Skeptics also note that Gottschalkthrew a wide net, measuring levels of 23trace metals. “If you look at enough variables,you’re bound to find a statistically significantassociation,” comments Curtiss D. Hunt of

The Tangled Roots of Violence

KELL

Y N

EWS

AN

D E

NTE

RTA

INM

ENT

KCRA

, Sac

ram

ento

EVERETT L. “RED” HODGES dis-cusses on television the results of anutritional study on inmate behavior.

Copyright 1995 Scientific American, Inc.

Page 84: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

the fear of getting hurt, it may lead to apredisposition to engage in violence.”But that hypothesis fails to explain whyat least 15 studies have failed to find ab-normal heart rates in psychopaths.

Jerome Kagan, a Harvard Universitypsychologist, has suggested that an in-hibited “temperament” may explainwhy the great majority of children fromhigh-risk homes grow up to becomelaw-abiding citizens. One study testedpulse, pupil dilation, vocal tension andblood levels of the neurotransmitternorepinephrine and the stress-regulat-ing hormone cortisol to distinguish in-hibited from uninhibited, underarousedtwo-year-olds. An expert panel on “Un-

derstanding and Preventing Violence”convened by the National ResearchCouncil suggested in its 1993 reportthat inhibited children may be protect-ed by their fearfulness from becomingaggressive, whereas uninhibited childrenmay be prone to later violence. The pan-el concluded that “although such factorsin isolation may not be expected to bestrong predictors of violence, in conjunc-tion with other early family and cogni-tive measures, the degree of predictionmay be considerable.”

Perhaps the most frequently cited bi-ological correlate of violent behavior isa low level of serotonin, a chemical thatin the body inhibits the secretion of stom-

ach acid and stimulates smooth muscleand in the brain functions as a neuro-transmitter. A large body of animal evi-dence links low levels of serotonin toimpulsive aggression. Its role in humansis often oversimplified, however. “Sero-tonin has a calming effect on behaviorby reducing the level of violence,” Jef-fery wrote in 1993 in the Journal ofCriminal Justice Education. “Thus, byincreasing the level of serotonin in thebrain, we can reduce the level of vio-lence.” A front-page article in Decem-ber 1993 in the Chicago Tribune ex-plained that “when serotonin declines...impulsive aggression is unleashed.”

Such explanations do violence to thescience. In human experiments, research-ers do not generally have access to theserotonin inside their subject’s brain-case. Instead they tap cerebrospinal fluidfrom the spinal column and measurethe concentration of 5-hydroxyindole-acetic acid (5-HIAA), which is producedwhen serotonin is used up and brokendown by the enzyme monoamine oxi-dase (MAO). Serotonin does its job by

Mysteries of the Mind 107

CHILDHOOD AGGRESSIVENESS, seenin this boy threatening his brother with abroom, is one of the strongest knownpredictors of later violence. Yet in 1990 a17-year study found that of 209 hyper-aggressive preschoolers predicted to de-velop antisocial behavior, 177 did not.

the Grand Forks Human Nutrition ResearchCenter in North Dakota. “But it may be mean-ingless.” Hunt adds that the concentration ofa metal in the hair does not tell one howmuch is in the blood or the brain. “We knowso little about manganese’s role in the bodythat we haven’t even set an RDA [recom-mended daily allowance] for it.”

Hodges remains convinced he is on theright track. “Violence can be detected andtreated,” he argues. In 1987 a mugger frac-tured the skull of another of Hodges’s sons.That year Hodges founded the Violence Re-search Foundation (VRF) to lobby public offi-cials to experiment with treatment programsthat use what he calls “the power of nutrition”to pacify violent criminals.

The VRF found an ally in Senator RobertPresley of California, who pushed through abill in 1989 authorizing a study of male prisoners by StephenSchoenthaler of California State University at Stanislaus. In thefirst part of the study, 402 offenders were divided randomly intothree groups and given vitamin-mineral supplements equivalentto the RDA, three times the RDA or a placebo. Preliminary results

showed that rule violations among the firstgroup dropped 38 percent during the study.Strangely, the behavior of inmates getting thehigher dose did not improve significantly,whereas violations rose 20 percent amongthe placebo group.

Although encouraging, the equivocal re-sults were so inconclusive that Schoenthalerdecided not to publish them until he com-pleted further studies with more controls.Hodges, however, publicized the results wide-ly at conferences and on television talk shows,much to the scientist’s annoyance.

Trace element deficiencies are just one ofmany frequently cited but poorly demonstrat-ed claims that nutritional problems can causecriminal and violent behavior. A 1992 reportby the Federal Bureau of Prisons stated thatcorrectional facilities in 46 states have incor-

porated a wide array of dietary intervention and testing pro-grams, even though “such programs are perceived by manyphysicians, scientific researchers, registered dietitians, and otherhealth care professionals as an incorporation of food faddisminto public policy.” —Steven Vames and W. Wayt Gibbs

Trace element

deficiencies are just

one of many

frequently cited but

poorly demonstrated

claims that nutritional

problems can cause

criminal and violent

behavior.

STEP

HEN

SH

AM

ES M

atri

x

Copyright 1995 Scientific American, Inc.

Page 85: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

binding to any of more than a dozendifferent neural receptors, each of whichseems to perform a distinct function.The low levels of 5-HIAA seen in vio-lent offenders may indicate a shortageof serotonin in the brain or simply adearth of MAO—in which case theirserotonin levels may actually be high.Moreover, serotonin can rise or drop indifferent regions of the brain at differenttimes, with markedly different effects.

Environment, too, plays a role: non-human primate studies show that sero-tonin often fluctuates with pecking or-der, dropping in animals when they arethreatened and rising when they as-sume a dominant status. The numerouspathways through which serotonin caninfluence mood and behavior confoundattempts to simply “reduce the level ofviolence” by administering serotoninboosters such as Prozac, a widely pre-

scribed antidepressant. Nevertheless,the link between 5-HIAA and impulsiveaggression has led to a concerted huntfor the genes that control the produc-tion and activity of serotonin and sever-al other neurotransmitters. “Right nowwe have in our hand many of the genesthat affect brain function,” says DavidGoldman, chief of neurogenetics at theNational Institute on Alcohol Abuseand Alcoholism. Although none has yetbeen shown to presage violence, “I be-lieve the markers are there,” he says.But he warns that “we’re going to haveto understand a whole lot more aboutthe genetic, environmental and devel-opmental origins of personality andpsychiatric disease” before making useof the knowledge.

Yudofsky is less circumspect. “We areon the verge of a revolution in geneticmedicine,” he asserts. “The future will be

to understand the genetics of aggressivedisorders and to identify those who havegreater tendencies to become violent.”

Few researchers believe genetics alonewill ever yield reliable predictors of be-havior as complex and multifarious asharmful aggression. Still, the notion thatbiologists and sociologists might togeth-er be able to assemble a complicatedmodel that can scientifically pick outthose who pose the greatest threat of vi-cious attack seems to be gaining curren-cy. Already some well-respected behav-ioral scientists are advocating a medicalapproach to crime control based onscreening, diagnostic prediction andtreatment. “A future generation will re-conceptualize nontrivial recidivistic crimeas a disorder,” Raine predicted in hisbook, The Psychopathology of Crime.

Compulsory Treatment?

But the medical model of crime maybe fraught with peril. When the

“disease” is intolerable behavior thatthreatens society, will “treatment” nec-essarily be compulsory and indefinite?If, to reexamine Raine’s hypotheticalexample, prediction models are judgedreliable but “biological, social and cog-nitive intervention programs” are not,

Seeking the Criminal Element108 Mysteries of the Mind

SOCIAL RISK FACTORS that seeminglypush youths toward violent behavior in-clude repeatedly witnessing assaults,heavy drinking or drug use (implicated inabout 60 percent of all offenses), associa-tion with deviant peers and gun posses-sion. Shown here are underage drinkingin New York City (above, left), boys inOmaha detained after a drive-by shoot-out (above, right) and the aftermath of ashooting in Houston (left).

KEN

NET

H J

ARE

CKE

Con

tact

Pre

ss Im

ages

AN

DRE

W L

ICH

TEN

STEI

N Im

pact

Vis

uals

STEP

HEN

SH

AM

ES M

atri

x

Copyright 1995 Scientific American, Inc.

Page 86: Mysteries of the Mind - WordPress.com · Mysteries of the Mind 3 M aster detective Hercule Poirot, the hero of many an Agatha Christie novel, boasted re-peatedly about the power of

might eight-year-old boys be judged in-corrigible before they have broken anylaw? Calls for screening are now heardmore often. “There are areas where wecan begin to incorporate biological ap-proaches,” Fishbein argues. “Delin-quents need to be individually assessed.”Masters claims that “we now knowenough about the serotonergic systemso that if we see a kid doing poorly inschool, we ought to look at his seroto-nin levels.”

In his article Jeffery emphasized that“attention must focus on the 5 percentof the delinquent population who com-mit 50 percent of the offenses. . . . Thiseffort must identify high-risk persons atan early age and place them in treat-ment programs before they have com-mitted the 10 to 20 major felonies char-acteristic of the career criminal.”

Yudofsky suggests a concrete methodto do this: “You could ask parentswhether they consider their infant high-strung or hyperactive. Then screen moreclosely by challenging the infants withprovocative situations.” When kids re-spond too aggressively, he suggests “youcould do careful neurologic testing andtrain the family how not to goad andfight them. Teach the children nonvio-lent ways to reduce frustration. Andwhen these things don’t work, considermedical interventions, such as betablockers, anticonvulsants or lithium.

“We haven’t done this research, but Ihave no doubt that it would make anenormous impact and would be imme-

diately cost-effective,” Yudofsky contin-ues. While he bemoans a lack of drugsdesigned specifically to treat aggression,he sees a tremendous “opportunity forthe pharmaceutical industry,” which hemaintains is “finally getting interested.”

But some worry that voluntary screen-ing for the good of the child might leadto mandatory screening for the protec-tion of society. “It is one thing to con-vict someone of an offense and compelthem to do something. It is another thingto go to someone who has not doneanything wrong and say, ‘You look likea high risk, so you have to do this,’”Akers observes. “There is a very clearethical difference, but that is a very thinline that people, especially politicians,might cross over.”

Even compelling convicted criminalsto undergo treatment raises thorny eth-ical issues. Today the standards forproving that an offender is so mentallyill that he poses a danger to himself orothers and thus can be incarcerated in-definitely are quite high. The medicalmodel of violent crime threatens to low-er those standards substantially. Indeed,Jeffery argues that “if we are to followthe medical model, we must use neuro-logical examinations in place of the in-sanity defense and the concept of guilt.Criminals must be placed in medicalclinics, not prisons.” Fishbein says sheis “beginning to think that treatmentshould be mandatory. We don’t ask of-fenders whether they want to be incar-cerated or executed. They should re-

main in a secure facility until they canshow without a doubt that they areself-controlled.” And if no effectivetreatments are available? “They shouldbe held indefinitely,” she says.

Moral Imperative

Unraveling the mystery of humanbehavior, just like untangling the

human genetic code, creates a moralimperative to use that knowledge. Toignore it—to imprison without treat-ment those whom society defines as sickfor the behavioral symptoms of theirillness—is morally indefensible. But toreplace a fixed term of punishment setby the conscience of a society withforced therapy based on the judgmentof scientific experts is to invite evengreater injustice.

Seeking the Criminal Element110 Mysteries of the Mind

Further Reading

The Psychopathology of Crime. Adri-an Raine. Academic Press, 1993.

Understanding and Preventing Vio-lence. Edited by A. J. Reiss, Jr., and J. A.Roth. National Academy Press, 1993.

What Works in Reducing AdolescentViolence. Patrick Tolan and Nancy Guer-ra. Available from the Center for the Studyand Prevention of Violence, University ofColorado, 1994.

Crime statistics and violence prevention pro-gram information are available on theWorld Wide Web at http://www.ojp.usdoj. gov/bjs

Scientists pursuing the role of biology in violent behaviorhave been twice shy since 1992, when shrill public criticism

forced the National Institutes of Health to withdraw financialsupport of a conference on the ethical implications of “GeneticFactors in Crime” and compelled former health secretary LouisSullivan to abort his proposed “Violence Initiative.” Led by fire-brand psychiatrist Peter Breggin, critics charged that in a societywhere blacks account for 12.4 percent of the population but 44.8percent of arrests for violent crimes, such research plays into thehands of racists.

The controversy did little to dissuade scientists from theirstudies, which continue to grow in number. The NIH reinstatedfunding for the genetics conference and increased its budget forviolence-related research to $58 million. Most Violence Initiativeprojects found support in other programs. In December 1994the National Science Foundation began soliciting proposals for a$12-million, five-year violence research consortium.

But the political wrangling seems to have intimidated investi-gators from including minorities in any violence studies with abiological tinge—and from collecting medical data in multiracial

studies. Designers of an 11,000-subject, eight-year study of thecauses of crime in Chicago, for example, decided not to collectblood and urine samples when in 1994 Breggin organized ralliesto block the project, says Felton Earls, a Harvard University pro-fessor and co-director of the study. As a result of such oppositionand pressure, asserts Adrian Raine of the University of SouthernCalifornia, “all the biological and genetic studies conducted todate have been done on whites. Scientifically, we can make nostatements on the biological basis of violence and crime inblacks or Hispanics or Asians.”

There is no reason to suspect that any genetic connectionlinks race to antisocial behavior. But there is reason to be con-cerned that ostensibly objective biological studies, blindly ignor-ing social and cultural differences, could misguidedly reinforceracial stereotypes. Still, Earls, Raine and other researchers empha-size that biological factors, if they exist, are important only inso-far as they protect individuals from—or make them vulnerableto—bad influences in their family, school and neighborhood. Re-search that excludes those who are most burdened by such pres-sures may be most expedient, but is it most useful? —W. W. G.

For Biological Studies, Minorities Need Not Apply

SA

Copyright 1995 Scientific American, Inc.