Anatomicalrelationship between basal Meynert - … · der of the basal ganglia, whereas Alzheimer...

5
Proc. Nati. Acad. Sci. USA Vol. 84, pp. 1408-1412, March 1987 Medical Sciences Anatomical relationship between the basal ganglia and the basal nucleus of Meynert in human and monkey forebrain (enkephalin/acetylcholinesterase/primate/human) SUZANNE HABER Department of Neurobiology and Anatomy, University of Rochester, Rochester, NY 14642 Communicated by Walle J. H. Nauta, October 20, 1986 ABSTRACT Previous immunohistochemical studies have provided evidence that the external segment of the globus pallidus extends ventrally beneath the transverse limb of the anterior commissure into the area of the substantia in- nominata. Enkephalin-positive staining in the form of "woolly fibers" has been used as a marker for the globus pallidus and its ventral extension. Acetylcholinesterase staining of both fibers and cell bodies, frequently used as a marker for the basal nucleus of Meynert, is also found in the area of the substantia innominata. This study describes the differential distribution of enkephalin-positive woolly fibers and acetylcholinesterase staining on adjacent sections in both the monkey and human basal forebrain area in an attempt to define the relationship between the basal ganglia and the basal nucleus of Meynert. The results show that while both occupy large regions of the basal forebrain, they overlap very little. In both species investigated, dense concentrations of acetylcholinesterase-pos- itive neurons lie, for the most part, outside the boundaries of the pallidal fibers. However, some scattered acetylcholines- terase cells do lie within the confines of the dorsal pallidum, and a more prominent group is found in the subcommissural ventral pallidum. These cells may constitute a group separate from the more densely packed acetylcholinesterase-positive cells in woolly fiber-free regions in that they may receive direct striatal input. Parkinson disease and senile dementia of the Alzheimer type are two degenerative disorders that occur in mid to late life. Whereas each disease is associated with a specific symptom- atology and pathology, a number of cases have recently been described in which some overlap of both disorders is evident (1-3). For example, a patient may initially exhibit classical Parkinson-like motor disabilities and later exhibit dementia characteristic of Alzheimer disease. Furthermore, cases have been described in which patients with the symptoms of one disease were found to have the pathology that characterizes the other disease, either in addition to, or in the absence of, the pathology of the diagnosed disease. Parkinson disease has long been known to reflect a disor- der of the basal ganglia, whereas Alzheimer disease has recently become widely attributed, at least in part, to a pathology of the basal nucleus of Meynert (4). The basal ganglia (composed primarily of the striatum and globus pallidus) and the basal nucleus of Meynert are traditionally viewed as separate functional entities-the former involved in somatic movement, the latter associated with cognitive functions. Recent anatorhical findings, however, have raised questions regarding a purely motoric function of the basal ganglia; in particular, the existence of connections linking the striatopallidal complex not only to the motor cortex but also to the prefrontal cortex (5-7) and amygdala has led to the suggestion that the basal ganglia could serve cognitive as well as motor functions (7). Because this notion places the basal ganglia in a functional category comparable, in part at least, to that of the basal nucleus of Meynert, a more detailed description of the relationship of these two structures to each other seemed of interest. The globus pallidus (in particular its most ventral part, the ventral pallidum) and the basal nucleus of Meynert are adjacent structures (Fig. 2 A-C and 3 A and B). The large acetylcholinesterase (AcChoEase)-positive neurons in the substantia innominata (i.e., the infrapallidal region of the basal forebrain) are regarded as a characteristic marker for the basal nucleus of Meynert and are therefore considered to define the nucleus in the monkey and human brain (8, 9). On the other hand, enkephalin-staining in the form of "woolly fibers" [enkephalin-positive striatal efferents ensheathing each of the long dendrites of pallidal neurons individually (Fig. 1)] has become recognized as an equally characteristic marker for the globus pallidus in the rat, monkey, and human forebrain (10-12). Large AcChoEase-positive neurons have been found within the boundaries of the globus pallidus in a distribution continuous with that of the basal nucleus of Meynert. Such cells are particularly numerous in a ventral region of the globus pallidus and the ventral pallidum of Heimer and Wilson (13) but also occur, more sparsely distributed, in more dorsal pallidal regions. It was initially assumed that these intrapallidal elements of the basal nucleus of Meynert are foreign to the globus pallidus and unrelated to the circuitry of the basal ganglia. Recent findings in the rat forebrain, however, suggest that at least some of these intrapallidal cholinergic neurons receive afferents from the striatum (14-16). Such a connection would provide a direct linkage of the basal ganglia to the basal nucleus of Meynert. This study demonstrates the extent to which the striatal projection to the globus pallidus (as defined by enkephalin- positive woolly fibers), and the basal nucleus of Meynert (as defined by its AcChoEase-positive neurons), overlap in the human and nonhuman primate brain. The rationale for the study was the assumption that striatal efferents establish synaptic contact with neurons in the basal nucleus of Meynert only within such areas of overlap. Moreover, it seemed possible that pathology affecting the function of such areas might be reflected in symptoms that represent both Parkinson and Alzheimer disease. EXPERIMENTAL PROCEDURES Five brains from neurologically normal patients were ob- tained between 8 and 10 hr after death. Coronal slabs of approximately 15-mm thickness were cut immediately, fixed by immersion in buffered fornialin for 2 weeks, then trans- ferred from lower to higher concentrations of sucrose (5%, 10%, and 20%), and finally stored in 30% sucrose in 0.1 M Abbreviations: AcChoEase, acetylcholinesterase; ChoAcTase, choline acetyltransferase. 1408 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Transcript of Anatomicalrelationship between basal Meynert - … · der of the basal ganglia, whereas Alzheimer...

Proc. Nati. Acad. Sci. USAVol. 84, pp. 1408-1412, March 1987Medical Sciences

Anatomical relationship between the basal ganglia and the basalnucleus of Meynert in human and monkey forebrain

(enkephalin/acetylcholinesterase/primate/human)

SUZANNE HABERDepartment of Neurobiology and Anatomy, University of Rochester, Rochester, NY 14642

Communicated by Walle J. H. Nauta, October 20, 1986

ABSTRACT Previous immunohistochemical studies haveprovided evidence that the external segment of the globuspallidus extends ventrally beneath the transverse limb of theanterior commissure into the area of the substantia in-nominata. Enkephalin-positive staining in the form of "woollyfibers" has been used as a marker for the globus pallidus andits ventral extension. Acetylcholinesterase staining of bothfibers and cell bodies, frequently used as a marker for the basalnucleus of Meynert, is also found in the area of the substantiainnominata. This study describes the differential distribution ofenkephalin-positive woolly fibers and acetylcholinesterasestaining on adjacent sections in both the monkey and humanbasal forebrain area in an attempt to define the relationshipbetween the basal ganglia and the basal nucleus of Meynert.The results show that while both occupy large regions of thebasal forebrain, they overlap very little. In both speciesinvestigated, dense concentrations of acetylcholinesterase-pos-itive neurons lie, for the most part, outside the boundaries ofthe pallidal fibers. However, some scattered acetylcholines-terase cells do lie within the confines of the dorsal pallidum, anda more prominent group is found in the subcommissuralventral pallidum. These cells may constitute a group separatefrom the more densely packed acetylcholinesterase-positivecells in woolly fiber-free regions in that they may receive directstriatal input.

Parkinson disease and senile dementia of the Alzheimer typeare two degenerative disorders that occur in mid to late life.Whereas each disease is associated with a specific symptom-atology and pathology, a number of cases have recently beendescribed in which some overlap of both disorders is evident(1-3). For example, a patient may initially exhibit classicalParkinson-like motor disabilities and later exhibit dementiacharacteristic of Alzheimer disease. Furthermore, cases havebeen described in which patients with the symptoms of onedisease were found to have the pathology that characterizesthe other disease, either in addition to, or in the absence of,the pathology of the diagnosed disease.

Parkinson disease has long been known to reflect a disor-der of the basal ganglia, whereas Alzheimer disease hasrecently become widely attributed, at least in part, to apathology of the basal nucleus of Meynert (4). The basalganglia (composed primarily of the striatum and globuspallidus) and the basal nucleus of Meynert are traditionallyviewed as separate functional entities-the former involvedin somatic movement, the latter associated with cognitivefunctions. Recent anatorhical findings, however, have raisedquestions regarding a purely motoric function of the basalganglia; in particular, the existence of connections linking thestriatopallidal complex not only to the motor cortex but alsoto the prefrontal cortex (5-7) and amygdala has led to the

suggestion that the basal ganglia could serve cognitive as wellas motor functions (7). Because this notion places the basalganglia in a functional category comparable, in part at least,to that of the basal nucleus of Meynert, a more detaileddescription of the relationship of these two structures to eachother seemed of interest.The globus pallidus (in particular its most ventral part, the

ventral pallidum) and the basal nucleus of Meynert areadjacent structures (Fig. 2 A-C and 3 A and B). The largeacetylcholinesterase (AcChoEase)-positive neurons in thesubstantia innominata (i.e., the infrapallidal region of thebasal forebrain) are regarded as a characteristic marker forthe basal nucleus of Meynert and are therefore considered todefine the nucleus in the monkey and human brain (8, 9). Onthe other hand, enkephalin-staining in the form of "woollyfibers" [enkephalin-positive striatal efferents ensheathingeach of the long dendrites of pallidal neurons individually(Fig. 1)] has become recognized as an equally characteristicmarker for the globus pallidus in the rat, monkey, and humanforebrain (10-12). Large AcChoEase-positive neurons havebeen found within the boundaries of the globus pallidus in adistribution continuous with that of the basal nucleus ofMeynert. Such cells are particularly numerous in a ventralregion of the globus pallidus and the ventral pallidum ofHeimer and Wilson (13) but also occur, more sparselydistributed, in more dorsal pallidal regions. It was initiallyassumed that these intrapallidal elements of the basal nucleusof Meynert are foreign to the globus pallidus and unrelated tothe circuitry of the basal ganglia. Recent findings in the ratforebrain, however, suggest that at least some of theseintrapallidal cholinergic neurons receive afferents from thestriatum (14-16). Such a connection would provide a directlinkage of the basal ganglia to the basal nucleus of Meynert.

This study demonstrates the extent to which the striatalprojection to the globus pallidus (as defined by enkephalin-positive woolly fibers), and the basal nucleus of Meynert (asdefined by its AcChoEase-positive neurons), overlap in thehuman and nonhuman primate brain. The rationale for thestudy was the assumption that striatal efferents establishsynaptic contact with neurons in the basal nucleus ofMeynert only within such areas of overlap. Moreover, itseemed possible that pathology affecting the function of suchareas might be reflected in symptoms that represent bothParkinson and Alzheimer disease.

EXPERIMENTAL PROCEDURESFive brains from neurologically normal patients were ob-tained between 8 and 10 hr after death. Coronal slabs ofapproximately 15-mm thickness were cut immediately, fixedby immersion in buffered fornialin for 2 weeks, then trans-ferred from lower to higher concentrations of sucrose (5%,10%, and 20%), and finally stored in 30% sucrose in 0.1 M

Abbreviations: AcChoEase, acetylcholinesterase; ChoAcTase,choline acetyltransferase.

1408

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Proc. Natl. Acad. Sci. USA 84 (1987) 1409

ill

I.

A'

I

0.

1. t ~

a;*I

FIG. 1. Single enkephalin-positive striatal efferent fiber joiningother fibers wrapping a pallidal dendrite to form a woolly fiber. Openarrow indicates single efferent fiber; filled arrow indicates woollyfiber.

phosphate buffer, pH 7.2 for 1-3 days. Four adult Africangreen monkeys (Cercopithecus aethiops) were anesthetizedand perfused with 4% paraformaldehyde, and their brainswere then transferred through the concentrations of sucroselisted above.

Antisera to enkephalin were provided by R. Elde (Univer-sity of Minnesota). Characterization of these antibodies hasbeen described in detail elsewhere (17). These antisera do notcross-react with either of the other two classes ofendogenousopiates, endorphin or dynorphin (see ref. 12 for this descrip-tion and a comparison between the differential distribution ofenkephalin and dynorphin in human globus pallidus).

Sections were cut at 50-pum thicknesses on a freezingmicrotome and incubated in primary antisera diluted 1:1000in 0.3% Triton X-100/0.1 M phosphate buffer (PB/T) at pH7.2 for 3-5 days at 40C. After being rinsed, the tissue wasincubated in goat anti-rabbit IgG (diluted 1:500 in PB/T)overnight at 40C, and then rinsed well and incubated in rabbitperoxidase-anti-peroxidase antisera diluted 1:50 in phos-phate buffer for 1 hr at room temperature. After rinsing with0.05 M Tris buffer, pH 7.6, the tissue was incubated in3,3'-diaminobenzidine tetrahydrochloride (0.05 mg/ml) inTris buffer/0.01% H202. Sections were rinsed, dehydrated,and coverslipped with Permount (Fisher). Adjacent sectionswere stained for AcChoEase by the Geneser-Jensen andBlackstad (18) technique. Although in brain stem regionscholinesterase-positive staining does not correspond well tothe staining observed using antisera against the acetylcho-line-synthesizing enzyme choline acetyltransferase (ChoAc-Tase), this does not appear to be the case in the basalforebrain. The decision to use the AcChoEase method ofmarking the cholinergic forebrain system was based on theobservation that neurons in the basal forebrain of the rat thatare strongly positive for AcChoEase are also ChoAcTasepositive (19).

Analysis entailed several approaches. For the line draw-ings, individual AcChoEase-positive cells were plotted with

a drawing tube. The drawing was then photographicallyreduced and taped to the wall for projection of the adjacentenkephalin-stained section onto it. To further verify regionsof overlap the two sections were viewed on separate micro-scopes equipped with a Leitz comparator bridge to allowsuperimposition of the images from the two slides. Finally,key sections were double-stained by the fluorescent-antibodymethod and the Geneser-Jensen and Blackstad technique forAcChoEase to verify the observations in the adjacent sec-tions. Due to the weak and fading fluorescence it was difficultto analyze double-stained sections in detail.

RESULTS

Numerous large AcChoEase-positive neurons were found inthe basal forebrain of all brains examined. Their distributionpattern was consistent with that described previously (8, 9).In many areas the individual cell bodies were somewhatobscured by the vast number of cholinergic fibers passingthrough the region; within such areas the actual number ofAcChoEase-positive neurons may have been underestimat-ed. Adjacent Nissl-stained sections confirmed the locationand distribution of these neurons originally described byMeynert in 1872 (20). Enkephalin-like immunoreactivityappeared in its characteristic pattern of woolly fibers in theregion of the globus pallidus.

Fig. 2 A-C illustrates the relationship between the enkeph-alin-positive fibers and the AcChoEase-positive neurons inthe monkey forebrain, and Fig. 3 A-E illustrates this in thehuman forebrain.

In the human brain, large AcChoEase-positive neuronsappear within the fiber laminae forming the boundaries of theglobus pallidus. Such cells are found within the internalcapsule and within both the external medullary lamina thatseparates the putamen from the external segment of theglobus pallidus and the internal medullary lamina that sepa-rates the external and internal pallidal segments. Whereasthese cells are distributed throughout the thickness of thelaminae, most are located directly adjacent to the externalsegment of the globus pallidus (Figs. 2A and 3B). A fewscattered AcChoEase-positive magnocellular neurons arealso seen within the external and internal segment of theglobus pallidus. These appear to be more numerous in theexternal segment than in the internal segment and also appearto be more numerous in monkey than in human. The numberof cells seen within the boundaries of the dorsal globuspallidus represents approximately 6% of the large Ac-ChoEase-positive neurons appearing in these cross-sections(Figs. 2 and 3). This, however, is likely to be an overestimatebecause, in the dorsal pallidum, such neurons are easilyidentified, whereas those lying ventral to the anterior com-missure are often masked by a great crowding ofAcChoEase-positive fiber bundles. In both monkey and human, the mainconcentration of the large AcChoEase-positive neurons ofthe basal forebrain is located ventral to the conventionallydefined globus pallidus and stretches rostralward to the levelof the anterior commissure (Figs. 1-3).

Enkephalin-positive woolly fibers are densely distributedthroughout the external segment of the globus pallidus andextend ventralward to include the ventral pallidum locatedbeneath the anterior commissure. This ventral region ofdense staining extends both ventrolaterally into the amyg-daloid complex (Fig. 2 B and C and Fig. 3 C and E) andmedially into the bed nucleus of the stria terminalis (Fig. 3 Aand B). Most of the internal pallidal segment is much lessdensely innervated with enkephalin-positive woolly fibers,but its medial portion does contain a prominent plexus ofsuch fibers (Fig. 2C and Fig. 3 B and C).Large AcChoEase-positive neurons located within the

fiber laminae bordering the globus pallidus lie entirely outside

Medical Sciences: Haber

r

4., 0

Proc. Natl. Acad. Sci. USA 84 (1987)

A

FIG. 2. Line drawing of coronal section through rhesus monkeybasal ganglia. Large black dots indicate AcChoEase-positive neurons ofthe basal nucleus of Meynert, and stippling indicates enkephalin-positive woolly fibers marking the globus pallidus. Closed arrows

indicate areas where AcChtEase-positive cell bodies lie outside thewoolly-fiber plexus. Open arrows indicate regions of overlap betweenAcChoEase-positive neurons and woolly fiber plexus. A, amygdala;AC, anterior commissure; C, caudate nucleus; GPe, globus pallidus-external segment; GPj, globus pallidus-internal segment; IC, internalcapsule; NB, basal nucleus of Meynert; P, putamen; and VP, ventralpallidum.

the plexus of woolly fibers. Even where such cells appear tolie directly on the border of the globus pallidus (see Figs. 2Aand 3 B-D), closer inspection reveals that they are adjacentto, but not within, the plexus of woolly fibers. Within themain body of the globus pallidus, the scattered AcChoEase-positive cells are all embedded within the enkephalin-positivefiber plexus, but this relationship is most obvious in theexternal segment (Figs. 2B and 3 B, D, and E).At several levels shown here the largest concentrations of

cells belonging to the basal nucleus of Meynert lie outside(although adjacent to) regions identified by enkephalin-positive woolly fibers as part of the ventral pallidum. Exam-ples of such nonoverlapping distribution are indicated in Figs.2 and 3 by solid arrowheads. In several restricted locations acircumscript void in the woolly-fiber plexus coincides pre-cisely with a cluster of basal-nucleus neurons (Fig. 2C).

In contrast, a substantial overlap of the basal nucleus withthe woolly-fiber plexus-and hence, with the distribution ofenkephalin-positive striatofugal fibers-is evident in a regionextending longitudinally underneath the globus pallidus. Theregion is indicated in Figs. 3 B-E by open arrows. Althoughthe number of cells embedded in this zone of enkephalin-positive fibers in any given cross-section is relatively small(18%, computed from the coronal section at the levels shown)in comparison with those lying outside the fiber plexus, theoverlap is consistent throughout the rostrocaudal extent ofthe ventral pallidum. It involves a greater number of basal-nucleus cells at its more caudal levels.

In sections stained for both AcChoEase and enkephalin,those AcChoEase-positive cells that are embedded within thewoolly-fiber plexus appear to be closely surrounded byenkephalin-like immunoreactivity. Whereas it is not possibleto determine at the light microscopic level whether or not theimmunoreactive tissue components include axons that actu-ally are making direct contact with the AcChoEase-positiveneurons, the histological picture suggests the possibility ofsuch a relationship (Fig. 4).

DISCUSSION

This evidence indicates that the basal nucleus of Meynertlargely lies outside the massive efferent fiber system of thestriatum in both monkey and human. In several places theAcChoEase-positive neurons of the basal nucleus lie imme-diately adjacent to, but not directly within, the plexus ofenkephalin-positive striatofugal fibers. Areas that lack thestaining of one substance are often positive for the other. Thisobservation conveys the impression that the two systems donot interact. However, two exceptions to this rule can bementioned.

(i) A few scattered cholinergic cells are located within theenkephalin-positive plexus permeating the dorsal globuspallidus. Whereas these cells form only a very small percent-age of the total population of the nucleus basalis, theirpresence raises questions as to the origin of their afferents. Itseems unlikely that cells so sparsely and widely disseminatedare innervated by an afferent fiber system separate from thecircuitry of the basal ganglia. As for the nucleus basalisneurons lying in the white matter laminae bordering theglobus pallidus-although their cell bodies are outside theintrapallidal plexus of striatofugal fibers-it must be consid-ered that these cells may send one or more of their longdendrites into the globus pallidus. Thus, there may indeed bea dorsal group of AcChoEase-positive nucleus basalis neu-rons (albeit a small one) that are in some way involved in thecircuitry of the basal ganglia.

(it) Within the region of the ventral pallidum, a largernumber of AcChoEase-positive neurons are found embeddedamong enkephalin-positive striatofugal fibers. Although

1410 Medical Sciences: Haber

Proc. Nati. Acad. Sci. USA 84 (1987) 1411

A ~ ~ ~~~I:\

,/ He-\ N

G-CPe- a/

; Ac, A ,f f ;f, J

C

10

FIG. 3. Line drawing of coronal section through human basal ganglia. Legend as for Fig. 2.

these cells, too, compose a relatively modest percentage ofthe basal-nucleus population, they form a rather consistentgroup clustered in a specific region of the enkephalin-positiveplexus of woolly fibers. They may represent a specific groupof cholinergic neurons that either receive input from, or sendprojections to, a particular brain region. The AcChoEase-positive neurons that constitute the basal nucleus of Meynert

have been further subdivided with respect to their afferentand efferent projections. Those neurons that are embedded inenkephalin-positive woolly fibers correspond best, but notcompletely, to the anterior lateral group (Ch4A1) and thedorsal group of the nucleus of the ansa peduncularis (Ch4id)described by Mesulam and co-workers (8, 21).

This group of cells is of particular interest. AcChoEase-

Medical Sciences: Haber

Proc. Natl. Acad. Sci. USA 84 (1987)

embede

"

tm,,

hAbeuaanuceuthhoae-psthaelamuros,teamygdala,amndan edial regionit s t heta ef ero nta corte xs including icthe edill

prfonitale courtex andth anadjoining mediaio stipnothemoeranoto

Studraies byacmbeddinathreion ofhitohemicntal panlduretogade

laeeings methodsi havfersowetatthprojectionsfromthenulsacregionsi questio(2)n thecorextandoramyogdaaarhicseud the

mosapt(7 fromiatheaindwelingijchoiergict cells rgoftherbsanoucleu (23-26),del whreansptostedntolto thethlmsuventralituegmntlaraandsubsathalamgrcnucljeusoaresoappeaatrignatey

neuprons popuatntheetregpionl(6s7e5)gEientlyuasynaticthcontractlothgensrital with,thebaioosal nucleusannaeurons ihabitingr thles rgonofe overlapus enygable,

tegon the hoinergi innrvatinclofn theamygdialaandoffrontal corticaldaresnhadonotio ofiasuchp contactmthus

frhStureste exaclusivelytuonofhsugghesielgtmicaanreroscopicevideince (15,od16a2),bu currentltappartoebejetin freceivin

segona communiction;trhef 28).rte woud thusal seiem concteiv

ablethacts certain ysfunctionofferethestriatum coubasld mnfleust

nern Vnaiigti ein foelpwudeal hstratmoafet hechoinrgc nnevaio o th aygalanoffotlcria ra.Te oino uhcnattu

evidence (A5and 27), butcurrenstlye npeuros tombeddeceivingmorepdefinpoitiveultrastructufrentsuppors. Arrow indicateywooll

psitive nemurnsintealouscation;in.2)I ol thussemorbanceivable thatliertiemeddysfnctheonsio of the venitralcould thates

themselves both in dyskinesia and in cognitive disorder of theAlzheimer type.

I thank Jeanine Schu for her technical assistance and Dr. N.Kowell, Massachusetts General Hospital, and the Brain TissueResource Center, McLean Hospital, Belmont, MA, for makinghuman material available for study. This study was supported byNational Institutes of Health Grants NS 22511 and Research CareerDevelopment Award K04 NS 01071.

1. Chou, H. C., Teng, E. L., Henderson, V. W. & May, A. C.(1985) Neurology 35, 1544-1550.

2. Mayeux, R., Stern, Y. & Spanton, S. (1985) Neurology 35,453-461.

3. Price, D. L., Whitehouse, P. J. & Struble, R. G. (1986) TrendsNeurosci. 9, 29-33.

4. Whitehouse, P. J., Price, D. L., Struble, R. G., Clark, A. W.,Coyle, J. T. & DeLong, M. R. (1982) Science 215, 1237-1239.

5. Goldschmidt, R. B. & Heimer, L. (1980) Soc. Neurosci. Abstr.6, 271.

6. Young, W. S., Alheid, G. F. & Heimer, L. (1984) J. Neurosci.4, 1626-1638.

7. Haber, S. N., Groenewegen, H., Grove, E. A. & Nauta,W. J. H. (1985) J. Comp. Neurol. 235, 322-335.

8. Mesulam, M.-M., Mufson, E. J., Levey, E. J. & Wainer,B. H. (1983) J. Comp. Neurol. 214, 170-197.

9. Saper, C. B. & Chelimsky, T. C. (1984) Neuroscience 13,1023-1038.

10. Haber, S. N. & Nauta, W. J. H. (1983) Neuroscience 9,245-260.

11. Beach, T. G. & McGeer, E. G. (1984) Neuroscience 13, 29-52.12. Haber, S. N. & Watson, S. J. (1985) Neuroscience 14,

1011-1024.13. Heimer, L. & Wilson, R. D. (1975) Golgi Centennial Sympo-

sium, ed. Santini, M. (Raven, New York), pp. 177-193.14. Zaborszky, L., Eckenstein, F., Leranth, C. S., Oertel, W.,

Schmeckel, D., Alones, V. & Heimer, L. (1984) Soc.Neurosci. Abstr. 10, 8.

15. Haber, S. N. (1984) Soc. Neurosci. Abstr. 10, 7.16. Grove, E. A. & Nauta, W. J. H. (1984) Soc. Neurosci. Abstr.

10, 7.17. Haber, S. N. & Elde, R. (1982) Neuroscience 7, 1049-1095.18. Geneser-Jensen, F. A. & Blackstad, T. W. (1971) Z.

Zellforsch. Mikrosk. Anat. 114, 460-481.19. Levey, A. I., Wainer, B. H., Mufson, E. J. & Mesulam,

M.-M. (1983) Neuroscience 9, 9-22.20. Meynert, T. (1872) ed. Stricker, S. A Manual of Histology

(English transl., J. J. Putnam) (Wood, New York), pp.650-766.

21. Mesulam, M.-M. & Mufson, E. J. (1984) Brain Res. 107,253-274.

22. Wilson, R. D. (1972) Dissertation, Massachusetts Institute ofTechnology, Cambridge, MA.

23. Nagai, T., Kimura, H., Maeda, T., McGeer, P. L., Peng, F. &McGeer, E. G. (1982) J. Neurosci. 2, 513-520.

24. Woolf, N. J. & Butcher, L. L. (1982) Brain Res. Bull. 8,751-763.

25. Grove, E. A., Haber, S. N., Domesick, V. B. & Nauta,W. J. H. (1983) Soc. Neurosci. Abstr. 10, 7.

26. Carlsen, J., Zaborszky, L. & Heimer, L. (1985) J. Comp.Neurol. 234, 155-167.

27. Grove, E. A., Domesick, V. B. & Nauta, W. J. H. (1986)Brain Res. 367, 379-384.

28. Grove, E. A. & Ingham, C. A. (1986) Soc. Neurosci. Abstr.12, 1328.

1412 Medical Sciences: Haber