Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Social Sciences 127

Fear, Startle, and Fear-PotentiatedStartle

Probing Emotion in the Human Brain

BY

ANNA PISSIOTA

ACTA UNIVERSITATIS UPSALIENSISUPPSALA 2003

Contents

1.Introduction..................................................................................................1 1.1 About this thesis ...................................................................................1 1.2 The emotion of fear ..............................................................................1

1.2.1 Why fear? – The defense system..................................................2 1.2.2 Learning to fear: the behavioral paradigm of fear conditioning ...3 1.2.3 Excessive fear: anxiety disorders..................................................4

Psychopathology of PTSD................................................................5 Psychopathology of specific phobias................................................5

1.2.4 Fear and anxiety: a matter of definitions ......................................6 1.3 The startle response..............................................................................6

1.3.1 Affective startle modulation .........................................................7 1.3.2 Anxiety disorders and affective startle modulation ......................7

Enhanced startle in PTSD .................................................................8 Startle potentiation in specific phobia...............................................8

1.3.3 What can startle tell us about emotional processing? ...................9 1.4 Fear circuits in the brain.....................................................................10

1.4.1 Amygdala: a prime structure for fear..........................................12 1.4.2 Hippocampus: emotional memories ...........................................13 1.4.3 Prefrontal cortex: emotional regulation ......................................14 1.4.4 Cingulate cortex..........................................................................14 1.4.5 Downstream from the amygdala: automated, reflexive defense behavior ...............................................................................................16 1.4.6 The neural circuitry of the primary startle response ...................16 1.4.7 Neuronal pathways in startle modulation ...................................17

1.5 Studying fear circuits in the brain with functional neuroimaging......18 1.5.1 Designing an imaging study .......................................................18 1.5.2 Positron emission tomography (PET).........................................20 1.5.3 Pre-processing and analyzing PET images.................................21

1.6 Objectives...........................................................................................23

2. The empirical studies ................................................................................24 2.1 Paper I ................................................................................................24

2.1.1 Aim and background ..................................................................24 2.1.2 Methods ......................................................................................25 2.1.3 Results ........................................................................................26

2.1.4 Conclusions ................................................................................28 2.2 Paper II ...............................................................................................29

2.2.1 Aim and background ..................................................................29 2.2.2 Methods ......................................................................................29 2.2.3 Results ........................................................................................30 2.2.4 Conclusions ................................................................................31

2.3 Paper III..............................................................................................32 2.3.1 Aim and background ..................................................................32 2.3.2 Methods ......................................................................................33 2.3.3 Results ........................................................................................34 2.3.4 Conclusions ................................................................................36

3. General discussion ....................................................................................38 3.1 Discussion of the empirical studies ....................................................38

3.1.1 Discussion of Paper I ..................................................................38 3.1.2 Discussion of Paper II.................................................................43 3.1.3 Discussion of Paper III ...............................................................45

3.2 Bringing it together ............................................................................47 3.3 Future directions.................................................................................48

4. Conclusions...............................................................................................50

5. References.................................................................................................51

Appendix A...................................................................................................61

Appendix B...................................................................................................63

Acknowledgements.......................................................................................65

Abbreviations

15O oxygen-15 radioactive isotope ACad ACC affective division ACC Anterior cingulate cortex ANOVA Analysis of variance BA Brodmann area bpm Beats per minute CAPS Clinician-Administrated PTSD Scale DSM-IV Diagnostic and Statistical Manual of Mental Disorders, fourth

edition EMG Electromyography fMRI Functional magnetic resonance imaging NSF Non-specific fluctuations PAG Periaqueductal gray matter PAS Panic Anxiety Scale PCC Posterior cingulate cortex PCL PTSD-checklist PET Positron emission tomography PnC Nucleus reticularis pontis caudalis PTSD Posttraumatic stress disorder rCBF Regional cerebral blood flow ROI Region of interest SCID Structured Clinical Interview for DSM-IV SNAQ Snake Anxiety Questionnaire SPQ Spider Phobia Questionnaire STAI-S Spielberger State Trait Anxiety Inventory SUDs Subjective Units of Distress

1

1.Introduction

1.1 About this thesis This thesis is about brain mechanisms of emotion. In particular, it is about the defensive behaviors of fear, startle and fear-potentiated startle. The thesis is based on three empirical studies, each addressing one of these behaviors. The method used to measure brain activity is positron emission tomography (PET), a functional neuroimaging technique.

The first study, Paper I, examines the neurofunctional correlates of fear in a group of patients suffering from combat-related posttraumatic stress disorder (PTSD), using an auditory symptom provocation paradigm. The second study, Paper II, explores the neural circuitry of the human startle response, its initiation as well as its habituation, in a group of healthy volunteers. The last study, Paper III, investigates neural activity during fear-potentiated startle in a group of subjects with specific snake or spider phobia, using a visual fear provocation paradigm.

In the following, a general background on emotion and its circuits in the brain is outlined, the emphasis being on those of fear and startle. The PET methodology is briefly presented, followed by a summary of the three empirical studies included in this thesis. Finally, a general discussion of all three papers is given, and an attempt to integrate the results and view them in the more general context of affective neuroscience.

1.2 The emotion of fear Do we run from a bear because we are afraid or are we afraid because we run? This has been a core question about the nature of emotion ever since William James posed it in 1884 in his article “What is an emotion?”. Contrary to the common sense notion that we meet a bear, are frightened, and run, James argued “that the bodily changes follow directly the perception of the exciting fact, and that our feeling of the same changes as they occur IS the emotion”. Thus, the experience of a particular emotion

2

results from the brain's perception of the body's reactions, particularly the pattern of autonomic nervous system activation. This was the first physiological theory of emotion, known as the James-Lange theory since it was proposed independently by James and Lange. A modern version along this line of thought is Damasio's somatic marker hypothesis which views emotions as resulting from somatic feedback to the brain (e.g. Damasio, 1994).

In early 1900s, this view was challenged by the Cannon-Bard theory, which holds that emotional experience and emotional expression are parallel processes that have no direct causal relation. Thus, emotional stimuli have two separate effects: they excite the feeling of emotion in the cortex and the expression of emotion in the autonomic and somatic nervous systems (Cannon, 1915). Today most researchers hold the view that each of the three components that orchestrate an emotional happening, i.e. the perception of the emotional stimulus (threat), the conscious experience of the emotion, and the expression of emotion (fear reaction), influences the other two.

Some decades later, James Papez, in search of a structural basis of emotion, asked “Is emotion a magic product, or is it a physiologic process which depends on an anatomic mechanism?”. He proposed that the mammillary bodies of the hypothalamus, the anterior thalamic nuclei, the cingulate gyrus, the hippocampus, and their interconnections constitute a circuit or mechanism through which emotion is expressed and controlled (Papez, 1937). The structures he described belonged to what, at that time, was referred to anatomically as the limbic lobe (from the Latin “limbus”, meaning border), and usually thought to have some olfactory function. Since 1937, a few nuclei and tracts have been added to the basic Papez circuit to produce what is called the limbic system.

Even though the limbic system concept remains the predominant view about how the brain processes emotions, today's view is a little more complicated and much less unproblematic. Functionally, this system is not dedicated to emotion exclusively nor is it the only brain system involved in emotional processing. Anatomically, there are yet no defined and agreed upon criteria for inclusion of brain areas in the limbic system (cf. LeDoux, 2000). This issue is further discussed elsewhere in this thesis (section 1.4).

1.2.1 Why fear? – The defense system In 1872 Darwin published his book “The Expression of Emotions in Man and Animals”. He argued that human expressions of emotion were universal, in other words innate and the product of evolution. As such, they are seen as adaptive behaviors that promote the survival of the organism (and the species). Those individuals that most successfully detect and respond to

3

danger are the ones with the best chances of survival. In mid 1900s, MacLean put forth the concept of the limbic system in the context of a general evolutionary explanation of emotion, arguing that emotions originate in the old cortex, i.e. the limbic system, in contrast to cognitions that are the responsibility of the neocortex (MacLean, 1949).

Along this line, emotions can be thought of as driven by the brain's primary motivational systems. Laid down early in our evolutionary history, these systems mediate behaviors that are basic to the survival of individuals and species. Unpleasant emotions, such as fear and anxiety, are generated by the defense system and are associated primarily with withdrawal, escape from pain, and defensive aggression1 (Davidson, 2002; Davidson, & Irwin, 1999; Lang, Davis, & Öhman, 2000; Öhman, 1996). Consider the emotion of fear in James' bear episode: the individual that detects the animal fastest and responds most adequately, by “fight or flight” (defensive action) or maybe “freezing” (defensive immobility), has the best survival options.

Thus, within an evolutionary perspective, the function of fear is to protect the organism from potential harm.

1.2.2 Learning to fear: the behavioral paradigm of fear conditioning As discussed above, fear is an evolutionarily adaptive response in that it protects the individual from harm. By remembering which situations are potentially harmful, we do not have to learn about the same dangers over and over again. After being attacked by James' bear in the forest, one might either enter the forest armed next time or avoid that forest in the future (some might even avoid all forests in the future). Further, by learning to discriminate between dangerous and safe situations, we are able to respond in flexible and appropriate ways. One will probably both feel and act differently when encountering a bear depending on whether it happens in a forest or in a zoo. How then do we learn to fear potentially dangerous situations?

Classical or Pavlovian conditioning (Pavlov, 1927) is at hand when an initially neutral stimulus after being repeatedly paired with a biologically significant stimulus, called the unconditioned stimulus (UCS), becomes a conditioned stimulus (CS) that later, in the absence of the UCS, elicits the full or part of the original response, now called the conditioned response. In the case of fear conditioning, after being associated with an aversive or threatening stimulus, the CS acquires affective properties and has the ability

1 The counterpoint of the defensive system is the appetitive system, associated with pleasant emotions and behaviors like approach, hunger, and nurturing.

4

to evoke behavioral and physiological fear responses that characteristically occur when the organism encounters a threatening or fear arousing stimulus. Examples of such defensive responses include autonomic nervous system activation (e.g. sweating, heart rate and blood pressure elevation), stress hormone release into the bloodstream, changes in pain sensitivity (hypoalgesia), and reflex potentiation (e.g. fear-potentiated startle).

Much of what we know about how the brain processes emotional, fear-relevant information and controls defense responses comes from the experimental study of fear conditioning. The technique has several advantages: it can be applied throughout the species, learning takes place very quickly and is relatively permanent, the stimuli involved can be specified and controlled, and the sensory system that processes the CS can be used as starting point for tracing the fear pathways through the brain (cf. LeDoux, 1996).

Not all fear learning takes place after direct exposure to or encounter with intrinsically dangerous stimuli, of course. We learn to fear many situations through so called vicarious conditioning (Bandura & Rosenthal, 1966; Craig & Weinstein, 1965), for example by observing somebody else being afraid, or just by being told that something is dangerous.

1.2.3 Excessive fear: anxiety disorders While fear and anxiety are perfectly normal and desirable reactions in the face of real danger (present or anticipated), they become an unwelcome and distressing burden when experienced in excess under circumstances that are not in fact dangerous. Clinically significant anxiety or fear upon exposure to the feared stimulus, together with avoidance of situations associated with such encounters, are the core characteristics of an anxiety disorder according to the DSM-IV (American Psychiatric Association, 1994).

Although the mechanism of fear conditioning alone cannot account for all aspects of fear and fear-related disorders (e.g. Kendler et al., 2002, for an example of a non-associative model of the etiology of phobias), it is an important factor in the etiology and maintenance of anxiety disorders. Fear conditioning models have been proposed for PTSD (e.g. Pitman, Orr, Shalev, Metzger, & Mellman, 1999), panic disorder (Goddard & Charney, 1997), social phobia (Mineka & Zinbarg, 1996; Öst, 1985), as well as specific (formerly called simple) phobia (Seligman, 1971; Watson & Rayner, 1920). In the present thesis, fear is studied in two anxiety disorders: posttraumatic stress disorder (PTSD) and specific (snake/spider) phobia.

5

Psychopathology of PTSD PTSD is a debilitating anxiety disorder resulting from exposure to an extreme and often life-threatening trauma. It is characterized by intense psychological distress and increased physiological arousal in the presence of trauma reminders and by avoidance of stimuli associated with the trauma (American Psychiatric Association, 1994). The diagnostic criteria for PTSD are listed in Appendix A.

Results of psychophysiological research on PTSD support the presence of a variety of autonomic, sensory, and cognitive processing differences between individuals with and without the disorder. The findings include exaggerated startle, increased conditionability and autonomic responsiveness to aversive, high-intensity stimuli, and elevated tonic or baseline physiologic activity. Increased sensitivity of the central nervous system is suggested by electrophysiological evidence for a failure to habituate to redundant information and over-responsiveness to novel information (for a review see Orr, Metzger, & Pitman, 2002). For example, patients with PTSD generally react with increased startle responses particularly during trauma recollection as compared to non traumatized controls (e.g. Pitman, Orr, Shalev, Metzger, & Mellman, 1999). This suggests an increased sensitivity of the brain's fear network, also presumed to modulate the startle reaction.

Psychopathology of specific phobias A specific phobia is an excessive and persistent fear of a specific object or situation, for example animals, blood/injections, heights, closed spaces, or flying. A person with this disorder experiences intense fear or anxiety when exposed to, or when anticipating an encounter with, the phobic stimulus. The feared situation is usually avoided or, less commonly, endured with great dread. Adults with specific phobia recognize that the phobia is excessive or unreasonable (American Psychiatric Association, 1994). The diagnostic criteria for specific phobia are listed in Appendix B.

It has been proposed that we might be biologically “prepared” to easily learn to fear potentially dangerous stimuli (e.g. snakes, angry faces, closed spaces) in a way that has favored survival throughout human evolution (Öhman, 1986; Seligman, 1971). Thus, part of the fear network might be “hard wired” in order to optimize defensive responding (see also Hettema, Annas, Neale, Kendler, & Fredrikson, in press; Lang, Davis, & Öhman, 2000). This may be one of the reasons why some phobias are so easily acquired and sometimes very difficult to extinguish.

As opposed to PTSD, where the traumatic event is extraordinary and all too well remembered, phobias are not always traceable to a single aversive or traumatic experience (one's own or someone's we observed or heard of). One reason for this seemingly irrational fact may be that conscious

6

perception of a fear-relevant stimulus is not necessary for aversive fear conditioning to take place. For example, Öhman and Soares (1998) demonstrated non-conscious conditioning to fear-relevant (snakes and spiders) but not to fear-irrelevant stimuli (flowers and mushrooms), when the conditioned stimuli were presented briefly (30 ms) and effectively masked by an immediately following masking stimulus.

1.2.4 Fear and anxiety: a matter of definitions Fear and anxiety are often used synonymously (and sometimes loosely) to describe a distressful emotional state in the presence of perceived danger. However, based on theoretical as well as clinical considerations, fear can be distinguished from anxiety in that fear assumes a specific feared object, whereas anxiety is thought of as an objectless, more general and long-lasting emotional state. Put another way, fear is held to be cue-specific, being a reaction to an explicit threatening stimulus, with variations in fight or flight as likely active behavioral outcomes. Anxiety on the other hand is prompted by less specific or more generalized cues and, although accompanied by physiological arousal, is often without organized functional behavior (e.g. Lang, Davis, & Öhman, 2000). Further, while fear is conceived as a reaction to a present danger, anxiety is often viewed as a reaction to an anticipated or imagined threat.

Bearing upon this distinction, one can conceive of PTSD as involving both fear and anxiety; fear as a response to explicit fear cues, for instance during symptom provocation by exposure to trauma reminders, and anxiety as generalized heightened vigilance and chronic increased arousal under baseline conditions. Specific phobias, on the other hand, are primarily associated with cue-specific fear, i.e. fear reactions upon exposure to the feared objects.

1.3 The startle response The startle response is a defensive reflex evoked by unexpected and intense stimuli, for instance a loud noise. It can be elicited by stimulation from various sensory modalities (i.e. acoustic, tactile, or visual stimuli), with the acoustic startle response being the most investigated instance of startle in both humans and animals. The reflex is a fast twitch of facial and body muscle, consisting of coordinated eyelid closure and contraction of facial, neck, and skeletal muscles, and is accompanied by arrest of other ongoing behaviors as well as heart rate acceleration. This pattern suggests that the startle reflex has a protective function against injury from a predator or from

7

a blow (as in the eyeblink reflex), that acts as a behavioral interrupt, clearing processors to deal with possible threat (e.g. Dawson, Schell, & Bohmelt, 1999; Koch, 1999). It is important to note that, although being an aversive reaction, the startle reflex itself is not a specific component of a state of fear or of a fear reaction. Rather, it is a reaction to novel, potentially harmful stimuli that mobilizes attention to new circumstances, and which is, as discussed below, potentiated during aversive experiences (cf. Lang, Bradley, & Cuthbert, 1998; Lang, Davis, & Öhman, 2000).

1.3.1 Affective startle modulation Being a reflex, and as such automatic and obligatory, startle is evoked also in the absence of a state of fear. The strength with which it is expressed, however, is affected by the emotional state of the organism, with larger responses occurring when the organism is in a defensive state (i.e. in a state of fear) and smaller in a pleasant state (cf. Bradley, Cuthbert, & Lang, 1990; Lang, Bradley, & Cuthbert, 1990). Startle magnitude is thus modulated by the affective valence of the foreground stimulus, with larger responses when processing unpleasant than pleasant stimuli.

The phenomenon is called affective startle modulation. This modulation occurs independently of probe modality (i.e. acoustic, tactile, or visual startle probes), or foreground modality (i.e. what is being processed at the time). Thus, affective modulation is found when subjects view aversive pictures or films, listen to distressing sounds, or smell disgusting odors (cf. Bradley, Cuthbert, & Lang, 1999; Codispoti, Bradley, & Lang, 2001; Jansen & Frijda, 1994; Sabatinelli, Bradley, & Lang, 2001; Yartz & Hawk, 2002). Furthermore, the startle reflex is enhanced not only during perception of aversive events but also in anticipation of them, for example unpleasant pictures (Sabatinelli, Bradley, & Lang, 2001). The fear-potentiated startle is actually shown to be maximal when anticipating shock, making it a sensitive measure of anticipatory fear or anxiety (Davis, Walker, & Lee, 1999).

1.3.2 Anxiety disorders and affective startle modulation There is a documented relationship between affective startle modulation and high levels of trait-fearfulness (Cook, 1999) as measured with the Fear Survey Schedule (FSS, by Wolpe & Lang, 1964). In a series of studies Cook and coworkers have demonstrated that high-fear subjects show reliable startle potentiation during aversive compared with pleasant imagery and while viewing aversive as compared with neutral pictures, whereas low-fear subjects do not show this effect. The acoustic startle response is also enhanced in patients suffering from a variety of anxiety disorders. Startle

8

modulation in two such groups, PTSD and specific phobia, is addressed below.

Enhanced startle in PTSD Exaggerated startle response is considered one of the core diagnostic criteria for posttraumatic stress disorder (PTSD) according to the DSM-IV (American Psychiatric Association, 1994). The majority of studies on startle and PTSD points to an enhanced startle in this group both at baseline and during affective startle modulation, i.e. when patients are exposed to trauma-related fear stimuli (e.g. Butler, Braff, Rausch, Jenkins, Sprock, & Geyer, 1990; Morgan, Grillon, Southwick, Davis, & Charney, 1995, 1996; Orr, Lasko, Shalev, & Pitman, 1995; but cf. Grillon, Morgan, Southwick, Davis, & Charney, 1996). Elevated startle at baseline, i.e. in the absence of explicit threat cues or trauma reminders, is seen as a symptom of general anxiety and chronic increased arousal since people with PTSD tend to develop an enduring vigilance for and sensitivity to environmental threat (van der Kolk, 1997). Some authors have suggested that the exaggerated baseline startle in PTSD might actually be due to affective modulation because of the increased frequency of aversive emotional states in this disorder, and/or the fact that aspects of testing situation elicit trauma-related fear, for example by resemblance of the acoustic startle probe to gunfire (cf. Cook, 1999).

Making a similar distinction between contextual and explicit fear cues, Grillon and coworkers (1998) showed that the experimental context potentiated baseline startle in a PTSD group but not in controls, suggesting that increased sensitivity to contextual fear cues may be a distinctive characteristic of PTSD. This distinction is interesting in light of animal research showing that contextual and cue-specific fear conditioning are supported by different neural substrates, i.e. amygdala- versus hippocampus-dependent learning (cf. LeDoux, 2000), and might have implications for understanding the psychopathology of PTSD as well as for developing “targeted” treatments of this disorder.

Startle potentiation in specific phobia Research with a variety of samples and study paradigms suggests that specific phobias are associated with potentiated startle during exposure to the feared objects relative to control stimuli. For example, startle potentiation was found during exposure to live spiders relative to control stimuli in a group of clinically diagnosed spider phobics (cf. Cook, 1999). Comparing subjects with specific phobias (snake/spider and mutilation fears) to a low-fear comparison group, Hamm, Cuthbert, Globisch, and Vaitl, (1997) found that both high-fear groups showed larger startle responses while exposed to pictures of their feared objects relative to neutral pictures,

9

whereas the low-fear group did not show this effect. Interestingly, startle reactions for standard picture material did not differ significantly between these groups. Thus, the intense fear experienced by phobics confronted with their feared object provide a model system for investigating fear-potentiated startle in humans under extreme affective conditions using symptom provocation; it could be predicted that stress-induced startle would be enhanced but baseline startle would not.

1.3.3 What can startle tell us about emotional processing? There are several methodological advantages to using startle modification paradigms to study emotional processing, particularly fear. One is that the startle response is a simple reflex that has a nonzero baseline. Since the basic reflex is elicited even in the absence of fear and is highly susceptible to emotional modulation, it offers an independent and sensitive index of fear. Because startle is an easily measured and nonverbal response, it can be used in a wide number of situations where other tools, such as verbal reports, are inappropriate.

The startle modification paradigm can give valuable information about the nature and timing of processing operations following environmental stimuli. That is, one can use the startle stimulus to “probe” into ongoing emotional (and cognitive) processing initiated by various stimuli. In neuroscience, this offers a unique window into the neurophysiological basis of behaviorally relevant processes. For example, since reflexes generally have short latencies, it is possible to determine the neural pathway that mediates the reflex which can then serve as a starting point to determine the neural pathways involved in fear or anxiety (Lang, Davis, & Öhman, 2000).

For clinical purposes, the startle modification paradigm can provide useful information about basic processes and mechanisms that mediate psychopathology, such as information processing and affective dysfunctions. It can be used, for example, to separately evaluate automatic and controlled cognitive processes, in order to determine whether cognitive dysfunctions in different psychopathologies are specific to automated or controlled processing deficits. Alternatively, it may be used to assess emotional processing deficits and determine if those deficits are specific to processing different emotional contents or are more general processing deficits (e.g Dawson, Schell, & Bohmelt, 1999).

Moreover, startle modification methodology can be used to test theories regarding the etiology of psychopathology in high-risk individuals, to identify subgroups within disorders, and to predict or evaluate treatment outcome. In phobias, for instance, it should be expected that, if fear potentiates startle in the presence of phobic stimuli, then effective treatment

10

of the phobia should reduce or eliminate startle potentiation. There is some evidence that this occurs; treatment has been found to eliminate potentiated startle during fear-relevant as compared to neutral imagery, and also during in vivo exposure to spiders. Moreover, it is possible to predict treatment outcome from pre-treatment startle responses; spider phobics with the most potentiated startle during exposure at pre-treatment showed the least reduction in self-reported fear for spiders at post-treatment and follow-up assessments (see Cook, 1999, for an overview).

Thus, the fear-potentiated startle paradigm is a valuable behavioral tool in the study of mechanisms involved in fear and anxiety, motivating a search for its neurobiological basis.

1.4 Fear circuits in the brain Much of what we know about the relation between fear and brain function comes from animal research, particularly from the neurophysiological study of rodents. Although animals lack the ability to communicate their emotions verbally, they do share with us the observable and measurable components of defensive behavior, such as autonomic fear signs, the startle response and fear-potentiated startle. Within the animal groups that have a backbone and a brain (amphibians, fish, reptiles, birds, and mammals, including humans), it seems that the neural organization of basic emotional behavioral systems, like the system underlying fearful behaviors, is very similar across species (LeDoux, 1996). Thus, models of defensive behavior in animals have been used as models of human fear as well (Charney & Deutch, 1996; Charney, Grillon, & Bremner, 1998; Davis, 1998; Lang, Bradley, & Cuthbert, 1998; Lang, Davis, & Öhman, 2000; LeDoux, 1996). These pinpoint subcortical brain structures, and suggest a dedicated fear circuit centered around the amygdala.

Figure 1 is a proposed neuroanatomical model of the key brain structures involved in afferent and efferent processing of fear- and anxiety inducing stimuli. The model is largely schematic in that some important reciprocal connections from the amygdala to key structures projecting to it, e.g. the hippocampus and the anterior cingulate cortex, are not shown. In the following, some key structures' functioning will be highlighted, either because they are of particular relevance for this thesis or because they are too central to be omitted.

11

Figure 1. Schematic diagram of the brain’s fear and anxiety circuit showing the main structuresinvolved in the afferent input of threat signifying stimuli to the amygdala, as well as the outputsfrom the amygdala to various target structures and possible functions of these connections.

THREAT STIMULUS

FEAR RESPONSES

Sensory input

SENSORY THALAMUS

SENSORY CORTEX

PERIAQUEDUCTAL GRAY HYPOTHALAMUS

CINGULATE CORTEX

PREFRONTAL CORTEX

HIPPOCAMPUS

AMYGDALA

NUCLEUS RETICULARISPONTIS CAUDALIS

Fight or flight response

Behavioral “freezing”

Potentiated startle reflex Hormonal stress response

Autonomic nervoussystem activation, e.g.tachychardia, sweatingblood pressure increase

12

1.4.1 Amygdala: a prime structure for fear The amygdala receives inputs from brain structures involved at various levels of cognitive processing. Auditory, visual, and somatosensory information is transmitted from peripheral receptor cells to the dorsal thalamus. The thalamus then relays sensory information to primary sensory areas of the cortex, which in turn projects through neighboring association areas to the amygdala. Inputs received this way, especially from later stages of processing in the cortical sensory systems, allow for more complex aspects of stimuli (i.e. objects and events) to activate the amygdala. The amygdala also receives crude sensory information directly from the thalamus, allowing the amygdala to be triggered by low-level stimulus features as well. Since it bypasses the cortex, the direct thalamic pathway is shorter and thus faster than the thalamo-cortical one, but also unable to benefit from cortical processing. It provides a “quick and dirty” pathway that allows us to begin to respond to potentially dangerous stimuli before we are fully aware of what the stimulus is, which can be very useful in dangerous situations (e.g. LeDoux, 1996). Accordingly, activation of the amygdala has been found even when fear-relevant conditioned emotional stimuli (i.e. angry faces) ere presented outside conscious awareness by means of backward masking (Morris, Öhman, & Dolan, 1998).

Studies of fear conditioning have made a strong case for the involvement of the amygdala in the genesis of fear and anxiety, and the neural changes mediating fear acquisition are thought to be located within the amygdala (cf. Davis, 2000; LaBar, Gatenby, Gore, LeDoux, & Phelps, 1998; LaBar, LeDoux, Spencer, & Phelps, 1995; LeDoux, 2000; Morris et al., 1996). It has been demonstrated (Amorapanth, LeDoux, & Nader, 2000) that the lateral nucleus in the amygdala links the conditioned and unconditioned stimulus to support the acquisition of memories. A wide range of stimuli can signal danger by triggering a response in the amygdala, including incoming sensory information, such as the sight of phobic objects, interoceptive information about bodily responses or states, as well as cognitions, such as retrieval of traumatic memories. For instance, increased neural activity in the human amygdala has been found during symptom provocation in patients suffering from PTSD, using either perceptual trauma reminders, e.g. audiotapes of combat sounds (Liberzon et al., 1999), or cognitive stimuli, e.g. personal scripts of the traumatic event (Rauch et al., 1996) and personally generated mental images of combat-related pictures (Shin et al., 1997).

Chua and Dolan (2000) have proposed that the functional role of the amygdala in response to these inputs is to link them to a central representation of fear. When the linkage is to specific objects or contexts, the resulting psychological state will have strong phobic elements, whereas

13

when the linkage is to interoceptive or cognitive stimuli, the psychological state will be characterized by strong, non-stimulus-bound anxiety elements.

Individual differences in amygdala activation are implicated in dispositional affective styles (Fischer, Tillfors, Furmark, & Fredrikson, 2001) and increased reactivity to negative incentives (Davidson & Irwin, 1999). The amygdala together with the anterior cingulate cortex, the hippocampus, the orbital, ventromedial, and dorsolateral prefrontal cortex, are considered key structures in the circuitry underlying emotion regulation. Emotion regulation includes processes that amplify, attenuate, or maintain an emotional state. Each of these interconnected structures plays a role in different aspects of emotion regulation, and abnormalities in one or more of these regions and/or in the interconnections among them are associated with failure of emotion regulation (Davidson, Putnam, & Larson, 2000).

Recent findings suggest that genetic differences in amygdala responsiveness may underlie differences in the vulnerability to fear and anxiety disorders. Specifically, individuals with one or two copies of the short allele of the serotonin transporter (5-HTT) gene exhibit greater amygdala neuronal activity in response to fearful stimuli compared with individuals homozygous for the long allele, implying that the “carriers” of this gene have an overactive or too easily triggered defense system (Hariri et al., 2002). This claim was substantiated by Garpenstrand, Annas, Ekblom, Oreland, and Fredrikson (2001) demonstrating an increased fear-conditionability in individuals with the short allele present.

1.4.2 Hippocampus: emotional memories Inputs from the hippocampus, which is involved in the formation and retrieval of explicit memories, allows the amygdala to be triggered by emotionally toned memories. Conditioning to fear contexts, but not to specific cues, involves hippocampal projections to the basal and accessory basal nuclei of the amygdala (cf. LeDoux, 2000). In turn, the amygdala is believed to exert a modulatory effect on hippocampal-dependent memory systems, via direct projections from the basolateral nucleus of the amygdala to the hippocampus (not shown in the figure), thus permitting the experience of fear to influence the consolidation of emotional memories (Charney & Deutch, 1996; Davis & Whalen, 2001). It has been suggested that this neuronal loop among the amygdala, hippocampus, and cortical/entorhinal projections relaying sensory information to the hippocampus may serve as a neural network that attaches cognitive significance to fear-inducing events and facilitates memory traces that enable the individual to rapidly initiate adaptive behavioral responses (Charney & Deutch, 1996).

14

Evidence from neuroimaging studies has suggested that the hippocampus may be adversely affected by stress following psychological trauma. Findings indicate decreased hippocampal function (hypoperfusion) during symptom provocation in patients with PTSD as well as changes in the structure (i.e. smaller volume) of the hippocampus in PTSD (for recent reviews see Bremner, 2002; Hull, 2002). Hippocampal damage together with high levels of emotional arousal (excessive amygdala activity) may prevent the proper evaluation and categorization of experience, accounting for the observation that patients with PTSD have difficulties processing arousing information, thereby making them vulnerable to reacting to new arousing stimuli as threats (van der Kolk, 1997).

1.4.3 Prefrontal cortex: emotional regulation The prefrontal cortex is a heterogeneous region with different anatomical and functional subdivisions (for a recent review see Wood & Grafman, 2003). It seems involved in affective working memory (Davidson & Irwin, 1999), the assessment of the significance of fear-relevant stimuli, the choice and implementation of coping behaviors, as well as the process of extinction of conditioned fear (Charney & Deutch, 1996). By direct and indirect projections to the amygdala, this region can modulate emotional processing in the amygdala, possibly by inhibiting activity within the amygdala (Hariri, Bookheimer, & Mazziotta, 2000).

Deficits in the ability of the prefrontal cortex to effectively inhibit the activity of the amygdala have been hypothesized to be involved in the development of anxiety disorders, such as specific phobia and PTSD, as well as the failure of extinction of fear responses in these disorders (Charney, Grillon, & Bremner, 1998). For example, reduced neural activity in the prefrontal cortex is evident during phobic stimulation in subjects with simple animal phobias (Fredrikson et a., 1993; Fredrikson, Wik, Annas, Ericson , & Stone-Elander, 1995), as well as during symptom provocation in patients with PTSD (e.g. Bremner, 2002; Fernandez et al., 2001). Moreover, neural activity in this brain area is shown to normalize after treatment with a selective serotonin reuptake inhibitor (SSRI), possibly reflecting extinction of fear and/or the increased emotional control (Fernandez et al., 2001). Also, lesions in the medial prefrontal cortex have been shown to interfere with extinction of conditioned fear in animals (cf. LeDoux, 2000).

1.4.4 Cingulate cortex The cingulate cortex is traditionally viewed as a major part of the “anatomical limbic system” and, as such, involved in emotion. More recent

15

clinical and experimental findings suggest that this is a somewhat oversimplified view since the cingulate cortex is a functionally heterogeneous area, participating also in sensory, motor, and cognitive processes. A functional dichotomy has been proposed between the anterior and the posterior part of the cingulate cortex by Vogt, Finch, and Olson (1992). They suggest that the anterior cingulate cortex (ACC) subserves primarily executive functions related to emotional control of visceral, skeletal, and endocrine outflow, while the posterior cingulate cortex (PCC) subserves evaluative functions in the organism's orientation within and interpretation of the environment, as well as memory. Close connections between the PCC and the parahippocampal cortex probably contributes to these processes.

The anterior, executive region has further been subdivided into affect and cognition components (for reviews see Bush, Luu, & Posner, 2000; Devinsky, Morrell, & Vogt, 1995). The affective division of the ACC (ACad) has extensive connections with the amygdala and the periaqueductal gray, and parts of it projects to autonomic brainstem nuclei. It is involved in conditioned emotional learning, assessments of motivational content, assigning emotional valence to internal and external stimuli, as well as the regulation of autonomic and endocrine functions (cf. Devinsky et al., 1995). Numerous studies have found involvement of the ACad in affective processing. In a meta-analysis of imaging studies reporting altered activity in the ACC during emotional and cognitive tasks, Bush et al. (2000) reported that the affective division has been activated by a variety of affect-related tasks, including emotional processing in healthy volunteers, symptom provocation in patients suffering from anxiety disorders, and during induced sadness both in healthy and depressed individuals. For example, activation within the ACad has been reported for simple phobia (Rauch et al., 1995), posttraumatic stress disorder (Rauch et al., 1996), and obsessive-compulsive disorder (Rauch et al., 1994) during provoked versus control conditions using PET.

A role for a caudal subdivision of the PCC, the retrosplenial cortex, in emotional processing has recently been proposed by Maddock (1999). Based on evidence from clinical, anatomical, as well as functional neuroimaging studies, he suggests that because this region has been shown to have intrinsic episodic-memory related functions and to increase its activity during processing of emotionally salient stimuli, it might be involved in the interaction between memory and emotion. Anatomically, the retrosplenial cortex is well situated to have such a role, given its connections with the anterior cingulate, prefrontal, parahippocampal and entorhinal cortices. This interaction could involve influences on memory encoding, memory retrieval, or other memory processes. For example, since the evaluation of an

16

emotionally significant stimulus often engages processes that rely on episodic-memory retrieval, the activation of the retrosplenial cortex by emotionally salient stimuli could reflect the retrieval of episodic memories.

1.4.5 Downstream from the amygdala: automated, reflexive defense behavior Structures “downstream” from the amygdala are implicated in the expression of different types of defensive behaviors. Behavioral, autonomic, and endocrine responses to danger-related stimuli are controlled by way of outputs from the amygdala. Direct projections to the nucleus reticularis pontis caudalis (PnC) in the brain stem modulates the startle response, potentiating the reflex in the context of aversive stimuli (for a more detailed description of the neural circuitry involved in startle modulation see below, section 1.4.7). Projections from the lateral hypothalamus control sympathetic nervous system activation, producing increases in blood pressure and heart rate, sweating, piloerection, and pupil dilation. Projections to the paraventricular nucleus of the hypothalamus mediate the release of corticosteroid stress hormones in response to danger. Amygdala projections to the periaqueductal gray (PAG) of the midbrain mediate coping behaviors, with the ventral part of this area being responsible for “freezing” and the dorsal for active fight/flight responses (for overviews see Charney et al., 1998; Davis, 2000; Lang et al., 2000).

1.4.6 The neural circuitry of the primary startle response The startle reflex has an extraordinarily short latency, e.g. 8 ms in rats, measured electromyographically in the hind leg. This means that the neural circuit of the primary acoustic startle response must be mediated by a simple neural pathway. This pathway is well documented in animals and involves three synapses located on (1) the cochlear root neurons, (2) neurons in the nucleus reticularis pontis caudalis (PnC), and (3) motoneurons in the facial motor nucleus and spinal cord (e.g. Davis, Walker, & Lee, 1999; Lee, Lopez, Meloni, & Davis, 1996). When auditory input enters the ear, auditory nerve fibers synapse onto cochlear root neurons embedded in the auditory nerve. Axons from the cochlear root neurons project then through the ventral acoustic stria directly to the PnC, a nucleus in the caudal part of the reticular formation in the pons. Projections of cells in the PnC form the reticulospinal tract, which make monosynaptic and polysynaptic synapses on spinal motoneurons that control whole-body startle. The PnC also projects to the facial motor nucleus that controls the pinna reflex in rats and, probably, the eyeblink reflex in humans. Lesions of the cochlear root neurons, the ventral

17

acoustic stria, the PnC, or the reticulospinal tract eliminate the acoustic startle response (Lee et al., 1996).

In adult humans the acoustic startle blink reflex typically consists of a burst of orbicularis oculi electromyaographic activity at a latency of about 30-50 ms (Berg & Balaban, 1999). The afferent pathway begins with sound transduced in the ear and transmitted along the cochlear nerve. The intervening steps between the cochlear output and the blink premotor area are unclear. The neural circuit might include the ventral cochlear nucleus, the superior olive, the lateral lemniscus, and the inferior colliculus (Tackmann, Ettlin, & Barth 1982). The final motor pathway for the blink reflex stems from the facial motor nucleus located at the pontine level of the brain stem, that innervate the orbicularis oculi muscle by the facial nerve. Thus, areas in the pontine tegmentum that project to the intermediate facial nucleus are presumed to be involved in the premotor control of the blink reflex (Holstege, Tan, van Ham, & Graveland, 1986). The underlying neural pathways of the human startle reflex remain to be clarified.

1.4.7 Neuronal pathways in startle modulation What is known about the neural correlates of fear-potentiated startle stems mainly from studies in rats, where startle is typically potentiated by eliciting the reflex in the presence of a cue that has previously been paired with shock, i.e. after fear conditioning. These studies suggest that the amygdala and its projections have an important modulatory role in startle modification by fear (cf. Davis, Walker, & Lee, 1999). The amygdala receives sensory information through its lateral and basolateral nuclei, which project to the central amygdaloid nucleus. Via the caudal part of the ventral amygdalofugal pathway the central nucleus then projects directly to the acoustic startle pathway, specifically to the PnC which appears to be the point where fear modulates startle. Lesions of the central nucleus or the lateral and basolateral nuclei of the amygdala block the expression of fear-potentiated startle, and electrical stimulation of the central nucleus increases startle amplitude (cf. Davis, Falls, Campeau, & Kim, 1993). The fear-potentiation effect is also blocked by chemical lesions or chemical inactivation of the amygdala (Davis, Walker, & Lee, 1999).

In humans, lesion of the amygdala results in failure to show the typical startle potentiation induced by aversive emotive foreground (Angrilli et al., 1996). Also, patients with right unilateral temporal lobectomy failed to show potentiated startle while viewing negative pictures, as compared to left temporal lobectomy patients and control subjects (Funayama, Grillon, Davis, & Phelps, 2001). In contrast, when presented with a stimulus they had been told signaled the possibility of a shock, the left temporal lobectomy patients

18

failed to show potentiated startle while both the right temporal lobectomy patients and the controls showed normal startle potentiation. Neuroimaging studies on startle modulation in humans are very few. Using functional magnetic resonance imaging (fMRI) and the prepulse inhibition (PPI) startle modification paradigm Hazlett et al. (2001) found greater responding in the anterior cingulate cortex during startle-alone relative to startle inhibition conditions, which might imply a role for this area in startle modulation. In a recent PET study, Frings and coworkers (2002) compared acoustic startle stimuli paired with fear-conditioned light stimuli to fear-conditioned light stimuli only and found increased rCBF in the medial cerebellum during the fear-potentiated startle condition.

1.5 Studying fear circuits in the brain with functional neuroimaging Before the neuroimaging age our understanding of the relationship between the workings of the brain and behavior relied largely on knowledge from more or less invasive techniques and manipulations in live subjects, and/or post-mortem observations. From experimental lesion studies in animals and spontaneous lesions in humans, for instance, one can infer brain function by observing the behavioral effects of lesions in specific areas. Stimulation studies can give us similar clues about region-behavior “matching”, i.e. by stimulating different brain regions and observing the behavioral consequences. Although valuable, those approaches have obvious limitations, especially when it comes to human subjects. Functional neuroimaging techniques such as positron emission tomography (PET) or functional magnetic resonance imaging (fMRI), on the other hand, provide a unique window into the active, living human brain, for example while performing a particular task or being in a state of fear. The interested reader is recommended Posner and Raichle (1994) for an excellent introduction into the field of human brain imaging; for a more comprehensive account see Toga and Mazziotta (1997).

1.5.1 Designing an imaging study The aim of an imaging experiment is to identify brain activity associated with specific behaviors or mental tasks. Since the brain performs numerous activities at all times (just by being active) this technique requires measurements in two states that differ from each other only in the aspect that one wishes to examine. Typically, a baseline condition (the reference task) is subtracted from the experimental condition of interest (the target task)

19

resulting in regionally specific differences in brain activity that identify the corresponding functionally specialized area. A textbook example illustrating the subtraction method would be a subject scanned in the camera once with eyes closed (baseline condition) and once with eyes open (the experimental condition of interest). By subtracting the eyes closed from the eyes open image, one gets a new image representing the changes in blood flow during visual input.

The problem of how to design two tasks such that they differ only in the mental process that one is interested in is often a tricky one. In contrast to the eyes-open versus eyes-closed example, which seems fairly straightforward, the study of emotion doesn't always render itself to a strict separation of conditions, nor is it self-evident what the baseline or control condition should be. When comparing emotion conditions in order to isolate the effects of a specific emotion one has to control for stimulus content as much as possible. By carefully matching the stimuli across conditions one controls for other characteristics that might result in differential responding, for example physical properties such as color, the presence of living objects, human vocalizations etc.

The subtractive paradigm builds upon the assumption of “pure insertion”, i.e. the assumption that cognitive states differ in components that can be purely inserted or removed with no interaction between them, both at the level of function and at the level of its neural representation. In other words, adding a cognitive component evokes an “extra” physiological activation that is the same irrespective of the cognitive or physiological context, and one cognitive component does not affect the effect of another. At the level of brain’s physiological implementation level, however, this assumption may not be true since the dynamics of neuronal networks are known to be nonlinear and there are numerous examples of modulatory interactions between different cortical areas. Moreover, at a conceptual level, the idea of pure insertion does not apply when one wants to study effects that are by definition modulatory, as for example in affective startle modulation where identical startle probes elicit different responses depending on the emotional and physiological context. One way to remedy this potential weakness is to use factorial designs where two factors are combined within the same experiment in order to study their integration, or interaction. The interaction term represents the effect of one factor on the effect of another, i.e. the expression of one cognitive process in the context of another (for an overview see Friston & Price, 2001; Friston, Price, Buchel, & Frackowiak, 1997).

Even though single-subject studies are conducted, PET experiments usually involve a group or several groups of subjects and the collection of one or more functional scans for each subject and condition. Both “within-

20

subject” designs (one group of individuals being scanned under different conditions) and “between-subject” designs (two or more groups scanned under the same condition), as well as combinations of the two, are common.

1.5.2 Positron emission tomography (PET) PET is a non-invasive technique by which “live” human brain function can be studied. It is based on the principle of positron emission, described in the following. Positrons (β+) are emitted from the nuclei of some elements that are unstable because they have a lower number of neutrons than the stable elements. Positron emission stabilizes the nucleus by removing a positive charge through the conversion of a proton into a neutron.

During this process, one element is converted into another, the latter having an atomic number less than the former. After traveling a few millimeters in tissue, the positron emitted from the decaying nucleus will interact with a negatively charged electron (e-) resulting in the annihilation of the particles. In the annihilation process two photons (511 keV gamma-rays) are ejected at 180 degrees from each other, and can be recorded by external detectors.

A PET scanner contains multiple radiation detectors (typically scintillation crystals, for example composed of bismuth germanate), placed in several rings around the object. Each one of these detectors is coupled in coincidence with opposite detectors. If two such detectors are hit simultaneously (i.e. within 10 ns) the PET camera will assume that it has been hit by two photons originating from the same decay, located somewhere along a straight line between the two detectors (called the coincidence line). Coincidence events are sampled along all possible lines between opposing detectors (at angular as well as linear positions) over a period of time. The tomograph's reconstruction software then takes the all the coincidence events to reconstruct an image that depicts the localization and concentration of the positron-emitting radioisotope within the organ that was scanned.

Common PET applications in neuroscience are the so called brain “activation” studies, where changes in regional cerebral blood flow (rCBF) are measured in order to infer brain function. The underlying assumption is that if energy demands of the brain increases locally then so does the blood flow. When an area of the brain is engaged in a particular task more energy is needed in that area, which is followed by an increase in rCBF2. Changes in

2 A change in neural activity is associated with changes in several physiological variables such as: energy consumption, glucose utilization, oxygen consumption, regional cerebral blood flow (rCBF), and regional cerebral blood volume (rCBV). In neuroimaging experiments using for instance PET and fMRI, the signals measured are not directly related to

21

neuronal activity can be traced by following the distribution of radionuclides in the brain. In a typical PET study, the subject is first positioned in the camera where the head is fixated inside the detector rings. A small amount of a positron-emitting radioisotope attached to a biologically relevant tracer, for example water labeled with the positron-emitting radioactive isotope oxygen-15 (15O), is injected into a vein in the arm. The isotope then circulates through the bloodstream and reaches the tissue of the brain where it accumulates in direct proportion to the local blood flow; the greater the blood flow, the greater the amount of radioactivity registered by the PET camera. The measurements of blood flow with PET are obtained during the period of time of one to two minutes following tracer injection, since the isotope of 15O has a very short half-life (about two minutes3). The final reconstructed data is in the form of an image in which the voxel4 values represent local radioactivity concentration in the whole scanned time window.

1.5.3 Pre-processing and analyzing PET images Before proceeding to the analysis of PET data, several image pre-processing steps have to be performed, briefly described in the following. In order to compensate for head movements between different scans, images from each subject are first realigned to each other, usually by estimating movement relative to the first scan in the series and adjusting the other scan/scans to the first scan, using automated image-matching algorithms. Secondly, individual differences in brain anatomy are cancelled out by applying spatial transformations that move and “warp” the images such that they all conform to some standard anatomical space, for example the Montreal brain atlas (Collins, Neelin, Peters, & Evans, 1994) or the Greitz atlas (Greitz, Bohm, Holte, & Eriksson, 1991) used in the studies in this thesis. Spatial normalization (also called anatomical normalization) of all images to standard space is necessary in order to implement voxel-based analyses on imaging data since data from different subjects must be derived from homologous parts of the brain. Finally, images are smoothed, which refers to applying a smoothing kernel to the data (a Hanning or Gaussian filter, for example). There are several reasons for smoothing; it generally increases signal relative to noise, it conditions the data in the sense that they conform

the neuronal activity but reflect, to various extents, the above variables (for a detailed account on brain function measurement issues see Ledberg, 2001). 3 More precisely, the half-life of 15O is 122.24 ± 0.16 s (from the WWW Table of Radioactive Isotopes at LBNL, Berkeley: http://ie.lbl.gov/toi/). 4 A voxel, abbr. for volume element, is defined as the smallest discrete spatial component of a digital volume.

22

more closely to a Gaussian field model (which is an important requirement if one wants to use the Gaussian field theory to make statistical inferences, i.e. assigning p-values), and it facilitates intersubject averaging by ensuring that hemodynamic changes from subject to subject are assessed on a spatial scale at which homologies in functional anatomy are typically expressed.

When using a subtractive design, differences in blood flow are identified by voxel–by–voxel comparisons of the whole image volumes between conditions. The assessment is in terms of an F or t value for each and every voxel in the brain, which can be transformed into the unit normal distribution to give a Gaussian field or z-value. The result is in the form of an activation map (z-score map), where the value of each voxel represents the change in activation (rCBF) in response to the experimental manipulation, an index at each voxel of how the brain has responded. Brain areas with higher rCBF during the experimental condition as compared to the baseline condition are said to be “activated”, and areas with lower rCBF are said to be “deactivated”. Finally, if one has several subjects, the individual images are averaged producing a mean difference activation image across subjects.

With an anatomically constrained hypothesis about effects in a particular brain region a priori the z-value in that region in the z-score map can be used to test the hypothesis. With an anatomically open hypothesis (that is, a null hypothesis that there are no effects anywhere in the brain) a correction for multiple non-independent comparisons is required, and the theory of Gaussian fields is used to compute the corrected p-values (for a comprehensive overview of the methodology of preprocessing and statistical analysis of imaging data see Frackowiak, Friston, Frith, Dolan, & Mazziota, 1997).

The location of the activations can then be indicated in terms of stereotaxic coordinates (x, y, z) given in millimeters from a specified location within a standard coordinate system, such as the Talairach and Tournoux brain atlas (1988). Some atlases (e.g. the Greitz atlas used in this thesis) provide more extensive anatomical information, such as macroanatomical (e.g. cortical sulci/gyri, deep brain nuclei) and cytooarchitectonical labels (e.g. Brodmann areas, BA). By using a standard metric to localize brain functions one can compare results across individuals and studies, independently of the method or neuroimaging software used. Even though the majority of laboratories report activation localizations in terms of Talairach coordinates, the issue of functional localization in the brain is not entirely unproblematic (e.g. Brett, Johnsrude, & Owen, 2002). One reason is the fact that different templates (brain atlases) are used to spatially normalize the images before converting them into Talairach space. The correspondence is often imperfect because of the differences in the

23

templates' size and shape, which results in a mismatch of reported activation areas, as well as confusion among neuroimagers.

There are several tools available for data pre-processing and analysis of neuroimaging data. The software used in the present thesis is a combination of the CBA (Computerized Brain Atlas, by Applied Medical Imaging, Uppsala, Sweden) together with additional in-house programs for data pre-processing (Andersson, 1995) and statistical analysis (Andersson, 1997; Friston et al., 1994, 1995).

1.6 Objectives The general aim of this thesis is to study neural correlates of defense behavior in humans. There are three specific objectives, each addressed by one empirical study (Papers I–III):

The first objective is to examine neurofunctional correlates of fear in posttraumatic stress disorder, an anxiety disorder associated with elevated startle, using a perceptual (auditory) symptom provocation paradigm (Paper I).

The second is to define the neural circuitry mediating the primary acoustic startle response in a group of healthy volunteers, by comparing neural activity during startle stimulation with activity during a resting state (Paper II).

The third objective is to study neural activity during affective startle modulation in a group of high-fear subjects (snake and spider phobics), using a perceptual (visual) fear provocation paradigm (Paper III).

24

2. The empirical studies

2.1 Paper I Neurofunctional correlates of posttraumatic stress disorder: a PET symptom provocation study

2.1.1 Aim and background Intense anxiety accompanied by increased physiological reactivity, e.g. exaggerated startle response, during trauma recollection are key features of PTSD symptomatology. This would suggest an involvement of the brain’s fear network, and especially of the amygdala being a central structure in the network and a known modulator of the startle response. To date, neuroimaging findings in PTSD seem to converge on brain areas associated with sensorimotor and emotional processing. Accordingly, altered neural activity in response to symptom provocation has been observed in the motor cortex, the amygdala, the hippocampus, the anterior and posterior cingulate cortex as well as regions in the prefrontal cortex (for some recent reviews see Bremner, 2002; Grossman, Monte, Buchsbaum, & Yehuda, 2002; Hull, 2002).

There are, however, inconsistencies across studies as to (1) the direction of change, e.g. the cingulate cortex is reported both to increase and decrease its activity during symptom provocation, and (2) the affected brain regions, e.g. an increase in amygdala activity is not invariably demonstrated in response to symptom provocation. Moreover, earlier imaging studies of symptom provocation in combat-related PTSD have exclusively involved Vietnam veterans, a group with high comorbidity of other psychiatric disorders as well as drug abuse. Because this can be a possible confound, it would be informative to study a more recently traumatized group. The aim of the first study was to examine brain activity associated with auditory symptom provocation in a group of patients with combat-related PTSD of relatively recent origin, using 15O-water PET.

25

2.1.2 Methods Participants. Seven male subjects (mean age = 37.7, range 28–52) fulfilling the DSM-IV criteria for chronic severe PTSD with a history of heavy combat and, in some cases, torture experience, were recruited through the University Hospital’s Unit of Transcultural Psychiatry in Uppsala. Screening included the Structured Clinical Interview (SCID) (First, Spitzer, Gibbon, & Williams, 1995), the Clinician-Administrated PTSD Scale (CAPS) (Blake et al., 1990), the PTSD-checklist (PCL) (Blanchard, Jones-Alexander, Buckley, & Forneris, 1996; Weathers, Litz, Herman, Huska, & Keane, 1993), as well as a psychophysiological assessment of responsivity to auditory war reminders. Patients could not meet criteria for previous or present alcohol or other drug abuse and were free of psychoactive medication at least two weeks prior to participation in this study. Four patients had comorbid mild depression and/or chronic pain disorder. All patients gave informed consent.

Experimental procedures. Subjects underwent 15O-water PET while listening to loudness-matched audiotapes of neutral5 (a repetitive sequence of a seven-tone melody) and trauma-related war sounds (e.g., machine gun fire, explosions, helicopter sounds). Each condition was presented twice and the order of scans was fixed for all subjects (i.e. neutral-traumatic-traumatic-neutral). The results presented here are limited to the contrast of the first traumatic to the first neutral condition only, because anxiety elicited by the traumatic stimulation persisted into the subsequent neutral condition, producing an asymmetric bias. Subjects had their eyes closed during scanning.

Psychophysiological and subjective anxiety measures. Heart rate, calculated from the Inter Beat Interval and expressed in beats per minute (bpm), and electrodermal activity, expressed as non-specific fluctuations per minute (NSF/min.), were obtained during all scans using the PSYLAB6 integrated system for psychophysiology (Contact Precision Instruments, London, UK). Subjective ratings of anxiety were obtained immediately after each scan using the state part of the Spielberger State Trait Anxiety Inventory (STAI-S 20–80) (Spielberger, Gorsuch, & Lushene, 1983) and Subjective Units of Distress (SUDs 0–100). In addition, subjects were given the Panic Anxiety Scale (PAS) (Wik et al., 1993), which is a visual analogue scale (VAS; 0–100) that contains all items from the DSM-IV definition of a panic attack. Rating four or more items as 51 (above the median) or higher

5 In the Symptom provocation procedures section in the published article, it is wrongly stated that the neutral stimulation consisted of simple 1000 Hz tones. Instead, it consisted of a repetitive sequence of a seven-tone melody. This does not, however, affect the results or the conclusions. Erratum has been submitted to the European Archives of Psychiatry and Clinical Neuroscience.

26

defined a panic attack. Because of an a priori hypothesis of increased responsiveness as a function of trauma reminders, all behavioral data were analyzed with directional t-tests of significance, using the StatView5 for Windows (SAS Institute, Inc).

PET measures. PET scans were obtained using a GEMS PC2048-15B eight-ring brain PET scanner (General Electric Medical Systems, Milwaukee, WI, USA) with a 6 mm axial and transaxial resolution, a 100 mm axial field of view and a 6.5 mm slice spacing. Following an initial transmission scan, 700–1300 MBq (approximately 15 MBq/kg body weight) of 15O labeled water was injected during each emission scan. Data were collected in fifteen 10-s frames following tracer injection, reconstructed to a 128 X 128 matrix and filtered with a 15 mm Hanning. The first seven frames after bolus arrival to the brain were summed producing a 70-s image for each scan. Individual images were realigned and spatially normalized to the Greitz computerized brain atlas (Greitz et al., 1991), using the CBA 3.0 software package (Applied Medical Imaging AB, Uppsala, Sweden).

Data were corrected for variations in global flow through linear scaling, and analyzed as a one way blocked analysis of variance (ANOVA) using an in-house program based on the general linear model (Friston et al., 1994, 1995). Local changes were evaluated by using the spatial extent of connected clusters of voxels with a z-score above 2.6, corresponding to a p < .01 corrected for multiple comparisons (Friston et al., 1994). In addition to the whole-brain search, a directed region of interest (ROI) search was performed on the amygdala, using directional t-tests of significance and the value of the central voxel in the right and left amygdala (at the CBA-predefined amygdala ROIs) for each subject and condition. Finally, amygdala activity at these voxels and subjective ratings of anxiety during traumatic stimulation were correlated.

2.1.3 Results Both self-reported anxiety measures were significantly higher during exposure to auditive war reminders as compared to neutral stimulation [STAI-S: t(6) = 4.09, p <.01; SUDs: t(6) = 2.03, p < .05]. Six out of seven patients had a full-blown panic attack in the scanner during traumatic stimulation as assessed using the PAS scale. However, two of those patients also experienced panic during the neutral condition, possibly due to anticipatory anxiety. For the psychophysiological reactivity measures, heart rate was significantly higher during traumatic stimulation as compared to neutral [t(6) = 2.10, p <.05], but electrodermal activity was not.

Analysis of the PET data showed altered brain activity in three main clusters as a function of traumatic provocation. Firstly, rCBF increased in the

27

right premotor (BA 6) extending into the primary motor (BA 4) and sensory (BA 1/2/3) cortices during the war related condition as compared to the neutral. The second cluster with increased activity due to symptom provocation was observed medially in the cerebellar vermis and extended into an area of the periaqueductal gray matter (PAG) adjacent to the pons. A third cluster located in the right retrosplenial cortex (BA 26/29/30 extending into BA 23) showed decreased rCBF as a function of combat sound provocation (see Figure 2).

Figure 2. Transaxial PET images, superimposed on a magnetic resonance (MR) template, showing significantly altered neural activity as a function of symptom provocation, in a group of seven patients with posttraumatic stress disorder. Left: Increased regional cerebral blood flow (rCBF) in the right supplementary motor cortex. Middle: Increased rCBF in the cerebellar vermis, extending into the periaqueductal gray matter. Right: Decreased rCBF in the right retrosplenial cortex. Images are displayed in radiological convention; the left side of the image corresponds to the right side of the brain.

The ROI analysis revealed significantly higher rCBF (expressed in radioactive counts) in the right [t(6) = 2.05, p < .05] but not the left amygdala [t(6) < 1, n.s.] during the traumatic as compared to neutral sound stimulation. Also, rCBF in the right amygdala correlated significantly with self reported anxiety (SUDs: rxy = .79, p < .05 and STAI-S: rxy = .78, p<.05) during the trauma-related condition (see Figure 3), whereas left amygdala activity was not correlated with either SUDs (rxy = -.19, n.s.) or STAI-S scores (rxy = -.38, n.s.).

28

Figure 3. Correlations between subjective measures of anxiety (STAI-S and SUDs) and regional cerebral blood flow (rCBF) (expressed in radioactive counts) in the right amygdala during exposure to auditory war and combat reminders, in a group of patients with post traumatic stress disorder.

2.1.4 Conclusions This study demonstrated that exposure to auditory trauma reminders relative to neutral sounds in a group of patients with combat-related PTSD was associated with increased neural activity in sensorimotor areas, the cerebellar vermis, the PAG, and the right amygdala. The amygdala and the PAG are parts of the fear network of the brain, known to be involved in /to mediate fear behaviors (e.g. the fight/flight reaction and fear conditioning). Co-activation of the amygdala, PAG, and motor/premotor areas suggests that symptom provocation in PTSD is associated with an emotionally driven motor preparation. It is thus possible that increased activity in this network also mediates the potentiation of startle that is characteristic for PTSD symptomatology.

In the present study, decreased activity in the retrosplenial cortex was observed. This area has previously been implied in memory processing of emotional material, although mainly by increasing its activity, for example during retrieval of emotionally toned autobiographical memories in non-clinical samples. It is conceivable that retrosplenial deactivation observed in this study reflects that aspects of flashback experiences in our patients during exposure to trauma-related cues represent subcortically driven unintegrated memory fragments rather than cortically stored integrated episodic memories, as suggested by van der Kolk (e.g. van der Kolk & Fisler 1997).

-20

0

20

40

60

80

100

120

SUDs

850 860 870 880 890 900 910 920 930 940

Right amygdala rCBF

40

45

50

55

60

65

70

75

80

STAI

-S

850 860 870 880 890 900 910 920 930 940

Right amygdala rCBF

29

2.2 Paper II The human startle reflex and pons activation: a regional cerebral blood flow study

2.2.1 Aim and background The neural circuit mediating the primary acoustic startle is well documented in animals and involves three synapses located on (1) the cochlear root neurons, (2) neurons in the nucleus reticularis pontis caudalis (PnC), and (3) motoneurons in the facial motor nucleus or spinal cord (Davis et al., 1999). The neural network underlying the human startle response, however, is not fully characterized since questions concerning the intervening steps between cochlear output and premotor control of the blink reflex are still unanswered. The primary aim of the second study was to determine the neural circuit of the human startle response by comparing rCBF measured during acoustic startle stimulation with rCBF during a resting condition. In humans it has recently been observed that startle repetition results in a decreased blood flow in the cerebellum (Timmann et al., 1998). An additional aim was to study neural correlates of the habituation of the startle response by comparing rCBF during the initial startle scans with rCBF during subsequent startle scans.

2.2.2 Methods Participants and experimental procedures. Eight healthy subjects (4 men and 4 women, mean age = 32, range = 26–43) were scanned with 15O-water PET during acoustic startle stimulation and rest. Startle probes consisted of eight presentations of a 500 ms white noise, delivered with inter-stimulus intervals ranging between 5 and 25 seconds (mean = 15 s), and an individually adjusted sound level (mean ≈ 95 dB; range 80–100). Each condition was scanned twice, in a counterbalanced order, and the average time between successive scans was 30–35 minutes.

Electromyographic (EMG) measures. The eyeblink component of the startle reflex was measured as electromyographic (EMG) activity from the orbicularis oculi beneath the eye, using two miniature electrodes (4-mm-diameter) placed approximately 25 mm apart. The EMG signal was band-pass filtered at 30–500 Hz, rectified/integrated and sampled at a rate of 100 Hz. Startle stimulus control and physiological data acquisition was accomplished using a standard setup of PSYLAB6 integrated system for psychophysiology (Contact Precision Instruments, London, UK). Startle amplitude was scored off-line as the maximum deviation from the mean

30

electromyographic level recorded for the last 200 ms before white noise onset, in the period from 20 to 120 ms after noise onset, using an in-house scoring program, and analyzed as a repeated measures analysis of variance (ANOVA) with the StatView 5.0 for Windows (SAS Institute, Inc.).

PET measures. PET scans were performed on an ECAT EXACT HR+ tomograph (CIT, Knoxville, TN, USA) with a 155 mm axial field of view and an axial/transaxial resolution of approximately 5 mm, and 2.5 mm plane spacing. During each scan 600–800 MBq of 15O-water was injected (approximately 10 MBq/kg body weight) and data collected in fifteen 10-s frames. Individual frames were corrected for attenuation and scatter and reconstructed to a 128 X 128 matrix using back projection and an 8 mm Hanning filter. Data from twelve frames after bolus arrival to the brain were summed, producing a 120-s image for each scan. Individual images were realigned and anatomically adapted to the standard stereotactic space of the Greitz brain atlas (Greitz et al, 1991), using the CBA 3.0 software package (Applied Medical Imaging AB, Uppsala, Sweden).

Data were normalized for global flow using linear scaling, and analysed as a one way blocked ANOVA using multiple linear regression (Friston et al, 1994). A t-map was created for each contrast, and subsequently converted to a z-score map. Two contrasts were performed: (1) average startle vs. average rest, and (2) startle session 1 vs. startle session 2. Significance was assessed at a z-score level of 1.96 because of the anatomically constrained a priori hypothesis about effects in the PnC and the cerebellum. Also, relative rCBF for each startle condition and each region of interest was correlated with startle EMG magnitude during that condition, in order to study covariation between central neural activity and peripheral startle measures between sessions.

2.2.3 Results EMG recordings confirmed the elicitation of the startle reaction and its habituation. Startle habituated both within-session (F(7,49) = 18.53, p < .0001) and between sessions (session by trial interaction, F(7,49) = 3.53, p < .005).

Primary analysis of PET data revealed increased rCBF in a medial posterior area of the pons with a maximum corresponding to the PnC [Talairach coordinates (8 -36 -18), z-score = 2.65, p < .01] during acoustic startle stimulation as compared to rest. The activation extended dorsally into the periaqueductal gray of the midbrain. Every individual had a higher rCBF in the pons area during startle stimulation as compared to rest (see Figure 4).

31

Figure 4. Left: Transaxial PET image, superimposed on a magnetic resonance (MR) template, showing significantly increased regional cerebral blood flow (rCBF) as a function of startle stimulation, in a group of eight healthy subjects. The activated pons area corresponds anatomically to the nucleus reticularis pontis caudalis (PnC). The image is in radiological convention; the left side of the image corresponds to the right side of the brain. Right: Average individual rCBF (expressed in radioactive counts) in the PnC during startle stimulation and rest.

Comparison of the first to the second startle scan (the between-session habituation contrast) showed decreased rCBF in the ventral cerebellum (z-score = 3.53, p < .001), and in the ventral pons (z-score = 2.69, p < .01) during the second scan. Increased rCBF was found in the medial cerebellum as a consequence of startle repetition (z-score = 2.59, p < .01). Also, there was a significant positive correlation between rCBF and EMG startle magnitude in the ventral cerebellum (maximum z-score = 3.47), but not in the PnC (z-score < 1.65). Startle magnitude was also positively correlated with neural activity in the ventral cerebellum (z-score = 3.42), but this area was located 21 mm dorsally to the area where the reduced rCBF was evident. Negative correlations between rCBF and startle magnitude were observed in the medial cerebellum (z-score = 2.95).

2.2.4 Conclusions This study demonstrated that the human startle response is associated with increased neural activity in a posterior medial pons area corresponding to the PnC. Thus, there is now evidence that the PnC is part of the neural circuit underlying the primary acoustic startle pathway also in humans. Long-term habituation (i.e. between-session habituation) was associated with decreased neural activity in the ventral cerebellum and the ventral pons, as well as increased activity in the medial cerebellum. Also, both positive and negative correlations were observed between EMG magnitude and rCBF in the

Individual observations440

460

480

500

520

540

560

580

600

Aver

age

PnC

rCBF

Startle Rest

32

cerebellum. EMG magnitude and rCBF was positively correlated in the ventral cerebellum (but not with the same focus as the deactivation), and in the ventral pons (approximately at the same coordinates as that deactivated as a result of startle repetition). Neural activity in the medial cerebellum was negatively correlated with EMG magnitude, suggesting that medial cerebellum controls long-term habituation also in humans, possibly by inhibiting PnC-associated activity.

Taken together, these results indicate that cerebellar activity might have a role in long-term startle habituation, but it can not be concluded if regulatory processes are initiated in the cerebellum or elsewhere. The negative correlation between medial cereballar rCBF and EMG magnitude would support the former, while decreased rCBF in the ventral cerebellum and ventral pons associated with between-session habituation together with the positive correlations between EMG and rCBF in these areas would support the latter proposition. In conclusion, it appears that a similar neurocircuitry underlies startle initiation and startle habituation in humans and animals.

2.3 Paper III Amygdala and anterior cingulate activation during affective startle modulation: a PET study of fear

2.3.1 Aim and background The amygdala is thought to modulate startle responses, since amygdala lesions have been shown to block fear-potentiated startle both in animals (e.g. Davis et al., 1999) and in human subjects (Angrilli et al., 1996; Funayama et al., 2001). The anterior cingulate cortex (BA 32) has been implied in a fMRI study of startle prepulse inhibition (Hazlett et al., 2001). The third study sought to investigate neural activity associated with affective startle modulation in a group of subjects with specific phobia by measuring rCBF during exposure to fear-relevant and fear-irrelevant pictures, paired and unpaired with acoustic startle stimuli. A factorial design was employed because startle modulation is conceptualized as the interaction between startle and affect. Thus, startle modulation was evaluated by the double subtraction (interaction) contrast [phobic startle vs. phobic alone] vs. [non-phobic startle vs. non-phobic alone].

33

2.3.2 Methods Participants. Sixteen female subjects (mean age = 22.75, range = 19–34) fulfilling the DSM-IV criteria for specific phobia of snakes or spiders, but not both, were recruited through advertisements at Uppsala University's major campus areas and in the local newspaper. Screening included web-based versions of the Snake Anxiety Questionnaire (SNAQ, 0–30) and Spider Phobia Questionnaire (SPQ, 0–31) (Klorman, Weerts, Hastings, Melamed, & Lang, 1974), telephone questioning, clinical interview, and a snake- and spider-fear provocation behavioral test. Exclusion criteria were current psychiatric disorder (other than specific snake or spider phobia), organic brain disorder, prescribed medication against those conditions, somatic disease, left-handedness, substance abuse, or pregnancy. Eight spider- and eight snake-fearful subjects that met the above criteria were recruited (spider phobics: mean SPQ-score = 23.25, mean SNAQ-score = 2.37; snake phobics: mean SNAQ-score = 23.87, mean SPQ-score = 3.13). Written informed consent was obtained from all subjects.

Symptom provocation procedures. Subjects underwent PET with 15O-water during four visual stimulus conditions: phobic alone, phobic startle (phobic paired with startle stimuli), non-phobic alone and non-phobic startle (non-phobic paired with startle stimuli). Subjects with snake phobia watched pictures of snakes as phobic stimuli and spiders as non-phobic, while the reverse was true for the spider phobics. Phobic and non-phobic stimulation was administered on separate days and the order of scans was counterbalanced across conditions. During startle sessions, eight startle probes were delivered concurrent with the visual stimulation in a pseudorandom order, with inter-stimulus intervals ranging from 5 to 17 s (mean = 10 s). Each startle stimulus consisted of a 500 ms white noise burst, with instantaneous rise time and individually adjusted sound level (mean = 95 dB, range = 80–100 dB).

Psychophysiological and subjective anxiety measures. The startle reflex was measured as EMG activity beneath the right eye, with procedures identical to those described in Paper II. Heart rate expressed in beats per minute (bpm), and skin conductance activity, expressed as non-specific fluctuations per minute (NSF/min), were recorded during all scans. Startle stimulus control and physiological data acquisition were accomplished using the PSYLAB6 system for psychophysiology (Contact Precision Instruments, London, UK). Subjective ratings of anxiety were obtained immediately after each scan using the state part of the Spielberger State Trait Anxiety Inventory (STAI-S, 20–80) (Spielberger et al., 1983) and Subjective Units of Distress (SUDs, 0–100). Self-reported anxiety measures, heart rate, and skin conductance activity were analyzed as 2 X 2 ANOVAs, i.e. Affective

34

Condition (phobic vs. non-phobic) X Startle Condition (startle vs. no startle), using the StatView 5.0 for Windows (SAS Institute, Inc).

PET measures. Scans were performed on the ECAT EXACT HR+ camera (CIT, Knoxville, TN, USA) with a 155-mm axial field of view and an axial/transaxial resolution of approximately 5 mm, producing 63 planes with 2.5-mm plane spacing. Approximately 10 MBq/kg bodyweight of 15O-water was injected during each scan. Data were collected in three 30-s frames, starting at bolus arrival to brain, and reconstructed to a 128 X 128 matrix using back projection and an 8-mm Hanning filter. The three 30-s frames of each scan were then summed, producing one 90-s image for each person and condition. All individual images were realigned and anatomically normalized into standard stereotactic space of the Greitz brain atlas (Greitz et al., 1991), using the CBA 3.0 software package (Applied Medical Imaging AB, Uppsala, Sweden).

Data were corrected for variations in global flow through global scaling to overall mean, and startle modulation was evaluated by the double subtraction contrast ([phobic startle vs. phobic alone] vs. [non-phobic startle vs. non-phobic alone]). A blocked analysis of variance (ANOVA) was performed using in-house software based on the general linear model (Friston et al., 1994, 1995). A t-map was created for the contrast, and subsequently converted to a z-score map. Because of the theoretically motivated directed search for predefined limbic and paralimbic brain regions, the critical z-value was set to 2.6 (corresponding to a p-value of < .01, uncorrected for multiple comparisons).

A note on the factorial design: Because startle modulation is conceptualized as the interaction between startle and affective conditions, this effect was evaluated by the double subtraction contrast [phobic startle vs. phobic alone] vs. [non-phobic startle vs. non-phobic alone]. This is necessary in order to study the specific effect of fear-potentiated startle that cannot be captured by the simple contrasts of (1) phobic startle vs. non-phobic startle, being the combined effect of phobic stimulation and startle potentiation, and (2) phobic startle vs. phobic alone, being the effect of startle during the phobic condition but providing no information as to whether this is specific for startle potentiation (contextual specificity) or if it is a general response to startle stimulation.

2.3.3 Results Analysis of EMG startle data confirmed startle potentiation during the phobic as compared to non-phobic condition [F(1,15) = 7.424, p = .0157, mean phobic = 470.16 vs. mean non-phobic = 307.46], as well as within-session habituation [F(7,105) = 7.505, p < .0001] (see Figure 5).

35

Figure 5. Fear-potentiated startle. Enhanced mean electromyographic (EMG) startle magnitude, in analog-to-digital (A/D) units, to eight acoustic startle stimuli presented during exposure to phobic as compared to non-phobic pictures, in a group of sixteen subjects with either snake or spider phobia.

The habituation data, not included in the original Paper III, are presented here because they illustrate that habituation of the startle response occurs also in a state of fear, and that in terms of short-term habituation phobic individuals seem to be comparable to the non-phobic volunteers described in Paper II.

The two subjective anxiety measures were higher during phobic stimulation as compared to the non-phobic, as indicated by significant main effects for the affective, but not the startle, condition [STAI-S: F(1,15) = 137.901, p < .0001; SUDs: F(1,15) = 125.280, p < .0001]. Similarly, heart rate and skin conductance activity were significantly elevated during fear-related exposure [F(1,15) = 18.243, p = .0007, and F(1,15) = 5.832, p = .0290, respectively], but for skin conductance a significant main effect of startle condition was also observed, indicating more NSF/min during startle stimulations [F(1,15) = 18.732, p = .0006].

Analysis of PET data revealed that startle potentiation during phobic fear was associated with significantly increased rCBF in the left amygdaloid-hippocampal area, encompassing the periamygdaloid cortex (BA34), the hippocampus and the amygdala [with the maximum activation located at Talairach coordinates (-30, -8, -17), z-value = 2.603]. Significantly increased activity was also found in the medial anterior cingulate cortex [BA32, Talairach coordinates (-4, 52, -1), z-value = 2.869] (see Figure 6).

0

100

200

300

400

500

600

700

1 2 3 4 5 6 7 8

Startle probes

Star

tle a

mpl

itude

(A/D

uni

ts)

PhobicNeutral

36

Figure 6. PET images, superimposed on a magnetic resonance (MR) template, showing altered neural activity during affective startle modulation. The images are based on a group of sixteen subjects with either snake or spider phobia during exposure to phobic and non-phobic pictures, paired and unpaired with acoustic startle stimuli. Transaxial, sagital, and coronal slices are displayed at the pixel of maximal activation. Top panel: Increased regional cerebral blood flow (rCBF) in the left amygdaloid-hippocampal area. Bottom panel: Increased rCBF in the left medial anterior cingulate cortex. Images are displayed in radiological convention; the left side of the image corresponds to the right side of the brain.

2.3.4 Conclusions The third study demonstrated that fear-potentiated startle was associated with increased neural activity in the left amygdaloid-hippocampal region and the anterior cingulate cortex (ACC). This particular region of the ACC is part of the affective division of the ACC (e.g. Devinsky et al., 1995), and is often found to increase its activity during various aspects of emotional processing. The amygdala has a well documented role in emotion in general (e.g. Aggleton, 2000), and in attention to and detection of threat in particular (e.g. Davis & Whalen, 2001; LeDoux, 1996). Given the close neuronal connections between the two regions, the present results suggest that a neural network dedicated to attention and orientation toward danger might

37

be activated, and that this network also is responsible for the potentiation of startle by fear, exerting a modulatory effect on the startle circuit.

38

3. General discussion

3.1 Discussion of the empirical studies

3.1.1 Discussion of Paper I The first study demonstrated that fear provocation by exposure to trauma reminders in patients with PTSD resulted in activation of parts of the network traditionally known to be involved in fear behaviors (i.e. the amygdala and PAG), and motor control (motor/premotor areas, cerebellum), as well as in emotional memory (amygdala). It is possible that increased activity in parts of this network mediates the potentiation of startle that is common in PTSD symptomatology. The deactivation observed in the retrosplenial cortex may reflect its role in compromised processing of autobiographical emotional memories. Each of these issues is discussed at some length below.

Amygdala activation and motor preparation. The increased neural activity observed in the amygdala during symptom provocation replicates previous findings in PTSD (Liberzon et al., 1999; Rauch et al., 1996; Shin et al., 1997). Amygdala activation probably reflects its central role in the processing of fear-related stimuli, from threat detection to generation and expression of fear, as further indicated by the significant correlation between amygdala rCBF and measures of self-reported anxiety.

Activation of motor cortex is also in line with previous results during symptom provocation in this group (Bremner, Narayan, et al., 1999; Liberzon et al., 1999; Rauch et al., 1996). In the present study, the “motor” cluster included the primary (BA 4), and supplementary motor cortex (BA 6) as well as the primary somatosensory cortex (BA 1/2/3), with the maximum voxel value located in BA 6. Neuroanatomically, the amygdala has little or no direct access to motor and premotor systems. The premotor cortex (BA 6) is one of the areas of the prefrontal cortex that receive only insubstantial amygdaloid input (Amaral, Price, Pitkänen, & Carmichael, 1992), and there are no direct projections from the amygdala to the motor cortex (Gloor,

39

1992). Thus, amygdala activation has to influence these systems indirectly, probably through connections to the nucleus accumbens of the ventral striatum and the brain stem. Also, by projecting to non-specific systems involved in the regulation of cortical arousal (e.g. the nucleus basalis of the forebrain), amygdala activity can arouse cortical networks indirectly (cf. LeDoux, 2000).

In addition to the primary and secondary motor cortex, increased activity was observed in the cerebellar vermis and the periaqueductal gray matter (PAG). PAG activation is associated with active escape behavior, i.e. the fight/flight response, but also with the freezing response. The observed activation during symptom provocation in the PTSD sample roughly corresponds to the dorsal part of the PAG, that is the part associated with fight/flight. These areas have close structural and functional interconnections, since the central nucleus of the amygdala projects to the PAG (cf. Swanson & Petrovich, 1998) and both structures seem involved in the fight/flight reaction (Bandler & Shipley, 1994; Behbehani, 1995; Gray, 1987) as well as fear conditioning in animals (cf. LeDoux, 1996) and humans (Büchel, Morris, Dolan, & Friston, 1998; Fredrikson, Wik, Fischer, & Andersson, 1995; Furmark, Fischer, Wik, Larsson, & Fredrikson, 1997; LaBar et al., 1998). The activated midbrain area adjacent to the pons may support aspects of the acoustic startle reflex in particular (Paper II, this thesis), while the supplementary motor cortex may carry motor aspects of the memory necessary for the fight or flight preparation (Squire & Zola-Morgan, 1991).

The cerebellum is now known to be involved in a wide range of behaviors beside that of motor control and coordination, such as classical conditioning, language, memory, and affect (for reviews see Desmond & Fiez, 1998; Schmahmann, 1998), supporting a broader concept of cerebellar function. For example, recent neuroimaging studies have established a role for the cerebellum in coordination and direction of selective attention in order to anticipate action (Allen, Buxton, Wong, & Courchesne, 1997; Courchesne & Allen, 1997), and in attention and memory (Cabeza & Nyberg, 2000). Thus, it is possible that the cerabellar activation observed here reflects emotionally driven attentional processes associated with motor activity (see also Damasio et al., 2000). An alternative view (Schmahmann, 1998) holds that the primary function of the cerebellum is that of a general neuromodulator, serving as an “oscillation dampener” to maintain function steadily around a homeostatic baseline by adjusting responsiveness in a variety of brain systems. Through its reciprocal connections to cortical and paralimbic regions (parahippocampal and cingulate cortex), it is well situated to modulate functions subserved by other systems. In this view, a plausible interpretation of the cerebellar activation observed during symptom

40

provocation would be that activity in the cerebellum inhibits activity of other (excessively) activated areas, maybe in a way similar to its involvement in the startle response habituation (Paper II).

In conclusion, the activation pattern observed in Paper I may reflect an emotionally determined motor preparation in a state of fear. The co-activation of the amygdala and motor areas most likely reflects the neural response to a threatening situation and the preparation for action (fight/flight) in order to cope with the threat. It is suggested that the increased neural activity in the primary and secondary motor cortex, the cerebellum, the amygdala, and the periaqueductal gray matter might represent a functional network supporting emotionally determined motor preparation.

Amygdala and affective memories. Traumatic memory recollection is a central component in PTSD. Presumably, these memories become especially intrusive and vivid during symptom provocation. Evidence from PET and fMRI studies converge with neuropsychological data and animal findings in that the amygdala is involved in the long-term storage of explicit affective memories, i.e. the modulation of memories that occur when people experience emotionally arousing events (for a review see Cahill, 2000). Administration of adrenergic agonists, at both pre and post-learning, has been shown to enhance memory for emotional material, whereas blockade of β-adrenergic receptors impairs memory for emotional material (Cahill, 2000). Also, amygdala activity has been shown to correlate with long-term recall for highly arousing material (Cahill et al., 1996). Hyperperfusion of limbic and paralimbic areas may be a result of long-term potentiation of the NMDA-mediated pathway between the amygdala and the periaqueductal gray (Adamec, 1997; Davidson & Sutton, 1995). The NMDA receptors are thought to be activated to produce long-term memories when sufficient glutamate is released as a result of stress. There is evidence from animal data that inputs from the basolateral amygdala (BLA) influences the processing of hippocampally dependent memory (McGaugh, Ferry, Vazdarjanova, & Roozendaal, 2000). Except from the hippocampal formation, the basal forebrain, the orbitofrontal cortex, and the retrosplenial cortex have been proposed as possible sites of amygdala modulation of memory (Cahill, 2000). Overactivation of this memory-modulating system may underlie the formation of excessively strong memories that are characteristic for PTSD.

Although the amygdala is clearly implicated in the encoding of traumatic events, its role in retrieval processes is less clear. It is also difficult to separate effects attributable to retrieval of aversive traumatic memories from the effects of the distress caused by symptom provocation. There are findings that support the participation of the amygdala in retrieval of memories for traumatic events (e.g. Rauch et al., 1996), and conscious

41

retrieval of episodic emotional memories (Dolan, Lane, Chua, & Fletcher, 2000). Moreover, the amygdala is involved in implicit memory processes associated with fear conditioning. For example, experimental studies on fear conditioning in humans have supported a role for the amygdala in the retrieval of conditioned fear responses (Büchel et al., 1998; Furmark et al., 1997; LaBar et al., 1998). The neural changes mediating fear conditioning are thought to be located within the amygdala (cf. Davis, 2000; LeDoux, 2000; but see also Cahill, Vazdarjanova, & Setlow, 2000). It has been demonstrated (Amorapanth et al., 2000) that the lateral nucleus in the amygdala links the conditioned and unconditioned stimuli to support the acquisition of memories. In the present study the experience of fear was strongly correlated with amygdala activation at coordinates consistent with the location of the lateral nucleus (Mai, Assheuer, & Paxinos, 1997), consistent with subcortically initiated memory recollection.

Amygdala activation in the present study could reflect its participation in retrieval of traumatic memories, implicit, explicit, or both. Moreover, hippocampally dependent retrieval of traumatic memories probably serves as fear-related input that further triggers the amygdala. It is also possible that activation is attributable to the amygdala's modulatory role in encoding of traumatic memories de novo during the experimental trauma-related exposure, rather than retrieval of past memories (associated with the difficulty to separate encoding from retrieval in a symptom provocation paradigm). Thus, it might be argued that the fear network centered around the amygdala is essential for forming emotionally relevant memories and that recollection of such traumatic memories reactivate the fear network or, conversely, that fear network activation facilitates memory retrieval.

Retrosplenial cortex. It has been proposed that the retrosplenial cortex might be the site where the interaction between memory and emotion takes place. Its involvement seems to be more related to perception and evaluation than to generation of emotional responses, and PET studies suggest that the retrosplenial cortex has intrinsic episodic-memory related functions (for a review see Maddock, 1999). Retrosplenial activation has been reported during retrieval of autobiographical episodic memories, as compared to retrieval of impersonal episodic information (Fink et al., 1996). In clinical populations, decreased activity in the posterior cingulate and retrosplenial cortices has been observed in Alzheimer's disease. For example, reduction of the metabolic rate in the retrosplenial cortex has been found in patients with very early Alzheimer, when memory impairment is the only symptom expressed, and in completely asymptomatic subjects at risk. In contrast, activity in this area has been reported to increase in schizophrenia and major depression, and to correlate with the severity of anxiety symptoms in major

42

depression, obsessive-compulsive disorder, and social phobia (reviewed by Maddock, 1999).

The present study showed decreased rCBF in the retrosplenial area of the posterior cingulate as a function of symptom provocation. This is in line with observations made by Liberzon et al. (1999) who also used a combat sound provocation paradigm. Contrarily, Bremner, Staib, et al. (1999) found increased rCBF in the posterior cingulate cortex (area 23) following exposure to traumatic pictures and sounds in Vietnam combat veterans with PTSD as compared to controls, as did Fischer and coworkers (Fischer, Wik, & Fredrikson, 1996) in a study on traumatized bank officials without a PTSD diagnose watching scenes from a video uptake of the actual robbery. These inconsistencies may be due to the differences in stimulation modality, i.e. auditory versus visual, or auditory only versus visual concurrent with auditory (Bremner, Staib, et al., 1999). The presence of visual stimulation might recruit active attentional processes and prompt monitoring and interpretation of the environment that activate the retrosplenial area in a way consistent with its role in visuospatial awareness (e.g. Vogt et al., 1992). The present study involved auditory stimulation only and the subjects had their eyes closed during scanning, which may promote intrusive phenomena (e.g. flashbacks) that lead to decreased neural activity, rather than active attentional processes that lead to increases in neural activity.

If retrosplenial activation is associated with processing of autobiographical memories, it is possible that the decreased activity observed in the present study reflects flashback experiences, i.e. the processing of extremely aversive autobiographical memories, a failure of effective processing of those memories during extreme stress, and/or defective memory integration. Given the retrosplenial cortex's close connections to the hippocampus, it is also possible that the “shutdown” of the retrosplenial cortex is associated with stress-induced altered function in the hippocampus during the course of PTSD, and that the reduced activity in the retrosplenial cortex reflects a possible flaw in the memory processing chain. Although speculative, this deactivation pattern might parallel the decreased activity in prefrontal areas often found during fear provocation in phobic subjects, and associated with the perseverative nature of those fear behaviors as well as their resistance to extinction (e.g. LeDoux, 1996). If flashbacks are conceptualized as perseverative memories, the retrosplenial deactivation might reflect the persisting and intrusive nature of these memories. Thus, the decreased activity in this area as a function of perceptually induced symptomatic responding in traumatized subjects with a PTSD diagnose, but not in those without (Fischer et al., 1996), might suggest that normal and pathological emotions affect similar cortical territory but in an opposite manner.

43

Finally, an obvious limitation of the PTSD study is its lack of control group, meaning that one cannot determine the specificity of the observed brain activation pattern. Although the combination of amygdala and motor cortex activation seem to be specific rather than non-specific signs of PTSD, the inclusion of control groups (healthy non-traumatized as well as traumatized volunteers without a PTSD diagnosis) is highly recommended. Another drawback of this study was the administration of neutral and traumatic stimulation on the same day. The results presented here are limited to the contrast of the first traumatic to the first neutral condition only, because anxiety elicited by the traumatic stimulation persisted into the subsequent neutral condition, producing an asymmetric bias. Administration of neutral and trauma-related (and generally fear-related) stimulation could be done on separate days, in order to avoid carry-over effects of the fear-related stimulation, as well as anticipation of it during the neutral condition. The potential gain from administration on separate days outweighs its practical inconvenience.

3.1.2 Discussion of Paper II The second study demonstrated for the first time in humans increased neural activity in a posterior medial pons area corresponding to the reticularis pontis caudalis (PnC) during startle stimulation. Even though the PnC is not delineated in the computerized brain atlas used in this study (Greitz et al., 1991) the location of the activated pons area is entirely consistent with the anatomical location of this nucleus (c.f. Hendelman, 1994). Activation of other nuclei like the nucleus facialis, the motor trigeminal nucleus, or possibly the nucleus coeruleus cannot with complete certainty be excluded but data suggest that beyond reasonable doubt PnC activation is at hand.

It has been suggested (Brown et al., 1991) that blinks elicited by repeated acoustic stimulation are instances of the blink reflex, and not instances of acoustic startle, and that different central/premotor areas might mediate the acoustic startle depending on whether the blink reflexes occur with or without other startle flexor reactions. Reflexes accompanied by other startle responses are controlled at the bulbopontine level by the brainstem startle center, whereas the blink alone response is controlled at the mesencephalic level by the midbrain reticular formation, the nucleus ruber and the inferior colliculus (Brown et al., 1991; Holstege et al., 1986). In the present study, however, there was no evidence of inferior colliculus or nucleus ruber activation, supporting the view that brain activity reflects the startle reflex.

An additional aim was to study neural correlates of the habituation of the startle response by comparing rCBF during the initial startle scans with rCBF during subsequent startle scans. In humans it has been observed that

44

startle repetition results in a decreased blood flow in the cerebellum (Timmann et al., 1998). In the present study, long-term habituation (i.e. between-session habituation) was associated with decreased neural activity in the ventral cerebellum and the ventral pons. Moreover, EMG magnitude and rCBF was positively correlated in the ventral pons approximately at the same coordinates as the deactivated area, as well as in the ventral cerebellum (but not with the same focus as the deactivation above). The two areas where neural activity simultaneously decreased are interconnected; the anatomical location of the ventral pons area that habituated with stimulus repetition is generally consistent with ventral pontine nuclei with known excitatory projections, via the cerebellar peduncle, to the granule cell layer in the cerebellar cortex and to deep cerebellar nuclear cells (Stein, 1986). Decreased activity in the ventral pons and ventral cerebellum indicate that this network is modulated during startle habituation.

Neural activity in the medial cerebellum increased as a consequence of startle repetition and was negatively correlated with EMG magnitude, supporting the idea that medial cerebellum controls long-term habituation also in humans by inhibiting PnC-associated activity (speculatively in pre-motor or motor areas). This is in line with animal studies that show that lesions in the medial cerebellum abolish long-term startle habituation (Leaton & Supple, 1991; Lopiano, DeSperati, & Montarolo, 1990). It has been suggested that short-term habituation involves inhibitory processes, either intrinsic or extrinsic to the startle circuit (reviewed by Koch, 1999). It is conceivable that activity in the medial cerebellum modulates PnC activation and associated startle EMG responses, as well as activity in the ventral cerebellum and the ventral pons, by inhibiting those areas during startle repetition also during short-term habituation. This would be consistent with the view of the cerebellum as general modulator of neurological function, regulating excessive activity in other brain areas, and/or involved in the motor control of eye movements (Schmahmann, 1998). From the present data, however, no conclusions regarding the neural correlates of short-term habituation can be drawn.

To conclude, this study provides evidence that the PnC is part of the neural circuit underlying the primary acoustic startle pathway also in humans. The between-session rCBF data together with the correlational data indicate that cerebellar activity is involved in long-term startle habituation. Thus, it appears that a similar neurocircuitry underlies startle initiation and habituation in humans and animals.

45

3.1.3 Discussion of Paper III The third study demonstrates that startle modulation by fear is associated with increased neural activity in the left amygdaloid-hippocampal region (encompassing the amygdala, the hippocamus, and the periamygdaloid cortex), and medially in the affective division of the anterior cingulate cortex (ACC). The amygdala has a well documented involvement in affect (e.g. Aggleton, 2000; LeDoux, 1996). Activation within the affective part of the anterior cingulate cortex has been reported for a variety of affect-related tasks (e.g. Bush et al., 2000; Devinsky et al., 1995), including symptom provocation in patients suffering from anxiety disorders (e.g. Rauch et al., 1995, 1996), as well as emotional processing in healthy volunteers (Elliott, Rubinsztein, Sahakian, & Dolan, 2000; Whalen et al., 1998). The amygdala and the anterior cingulate cortex are closely interconnected, and it has been suggested that they operate together to produce affective behaviors (cf. Vogt et al., 1992). The present results suggest that this network is involved in the modulation of human startle responses as well.

In animals, the amygdala has a known neuromodulatory role on brainstem activity associated with startle potentiation through its direct projections to the PnC (Davis et al., 1993; Groenink, Joordens, Hijzen, Dirks, & Olivier, 2000; Walker & Davis, 2002). The increased activity in the amygdaloid cortex observed in the present study most likely reflects this modulatory effect on the startle circuit, although we could not show increased activity in the PnC itself during startle potentiation. Involvement of ACC BA32 in startle modification have previously been reported in an imaging study using the prepulse inhibition (PPI) paradigm by Hazlett and coworkers (2001). They found that the startle-alone condition elicited the greatest response, the ignored-PPI condition an intermediate, and the attended-PPI condition the smallest response. The differential responding of area 32, with relatively increased activity during the startle-alone condition in relation to the startle inhibition conditions (i.e. the attended- and ignored-PPI) in Hazlett's study, and increased activity during startle potentiation in the present study, strongly points to its involvement in startle modulation. It is possible to view these altered neural responses, i.e. varying from smaller to greater, as corresponding to the startle phenomena that accompany them, i.e. from inhibited via “normal” to enhanced startle responses. This would suggest that this area may be crucial for the processing and modulation of startle responses both by attention and affect. It should also be noted that the observed activation pattern does not represent phobic fear per se, since the effect of fear is canceled out by the use of a factorial design. Thus, the observed activations correspond to the interaction between fear and startle, that is, the context-specific effect of fear-potentiated startle.

46

Although the affective significance of the stimulus (i.e. the fact that snake and spider stimuli signify danger) is probably determined by the amygdala, it is possible that changes in neuronal responsiveness to new environmental cues (i.e. the startle stimuli) may be guided by the ACC. From the present results one can not determine the exact point at which ACC exerts its modulatory effects on the startle pathway. It could be by projections via the amygdala (basolateral nucleus) or by direct or indirect projections to the PnC. The following “dynamics” are conceivable: the amygdala detects threat and arouses the defense system to begin with, but once the fear-network is activated the ACC helps direct and lock attention to the fear-relevant stimuli, further increasing vigilance in a loop-like fashion. There is of course the possibility that the ACC itself does not participate in modulation of the startle response but rather “mirrors” activation elsewhere; activation in the ACC could be caused by direct inputs from the amygdala as well as from the thalamus, or both.

There are, however, findings that support an active modulatory role for this particular ACC region. Area 32 has recently been implied also in pain modulation by anxiety (Ploghaus et al., 2001), with increased activity observed during exacerbation of pain by anxiety. Thus, the affective ACC may have a general regulatory role in the processing of emotionally aversive experiences, for example during fear, pain, and anxiety, with an activation in this area associated with heightened attention to unpleasant and threatening stimuli.

Also, the co-activation of amygdala and ACC found here has previously been reported by a number of studies dealing with affective processing (e.g. Blair, Morris, Frith, Perrett, & Dolan, 1999; LaBar et al., 1998; Vuilleumier, Armony, Driver, & Dolan, 2001). For example, co-activation of ACC and amygdala during processing of fearful vs. happy faces has been interpreted as the engagement of a neural system coordinating behavioral responses to threatening stimuli (Morris, Friston, et al., 1998). A similar central integrative role for the human “rostral limbic system” in the processing of sensory stimuli and the formation of context-dependent adaptive behavioral responses has been suggested by Büchel et al. (1998), who showed enhanced responses in the ACC together with differential responding in the amygdala during aversive conditioning. The results from the present study demonstrate that it is possible that the motor output, i.e. startle, is modulated by co-activation of the amygdala and the ACC.

To conclude, the third study provides evidence from functional brain imaging for a modulatory role of the amygdala on startle reactions in humans. They also point to an involvement of the affective ACC in the processing of startle stimuli during emotionally aversive experiences. The co-activation of these areas may reflect increased attention to fear-relevant

47

stimuli. It is suggested that the amygdala and the ACC form part of a neural system dedicated to attention and orientation to danger, and that this network is involved in startle potentiation driven by negative affect.

3.2 Bringing it together This thesis has highlighted the role of some of the brain's structures involved in fear and defense behavior. Two of the studies employed emotion manipulation (Paper I and III) and demonstrated activation of previously identified fear-structures, such as the amygdala, premotor/motor areas, and the anterior cingulate cortex, in a manner consistent with existing findings. A somewhat surprising finding was the decreased neural activity in posterior cingulate cortex during trauma reminders in the PTSD group. The startle “elicitation” study (Paper II) demonstrated, for the first time in humans, that activation of the PnC is at hand during startle, pointing to a similar primary startle circuit for humans and other mammals. The startle modulation study (Paper III) demonstrated amygdala involvement during fear-potentiated startle also in human subjects, further adding to the body of evidence for a shared fear network. The recruitment of the affective ACC during startle modulation suggests that this region together with the amygdala may form, or be part of, a general network specialized in the detection of and attention to threat.

The amygdala and the PAG are parts of the fear network of the brain, and are known to mediate fear behaviors (e.g. the fight/flight reaction and fear conditioning). Co-activation of the amygdala, PAG, and motor/premotor areas suggests that symptom provocation in PTSD is associated with an emotionally driven motor preparation. It is thus possible that increased activity in this network also mediates the potentiation of startle that is characteristic for PTSD symptomatology. Although it can not be held that the startle itself is a fear response, the startle circuit is part of the fear network of the brain inasmuch as it is under its modulatory control. Since startle in humans is associated with activity in the PnC, modulation within the startle circuit during a state of fear is probably mediated through direct or indirect projections from the amygdala to the PnC. The startle response can be considered a premotor behavior in that it functions as an “interrupt” of ongoing behavior in preparation to deal with a new setting, possibly a threatening one.

The amygdala is involved in coding fear cues from different modalities. The studies included in this thesis documented amygdala activation following visual as well as auditory fear provocation. A neuromodulatory role for the amygdala has been proposed by many researchers and amygdala

48

activation has been associated with increased vigilance and attention to threat, as well as enhanced detection of threatening stimuli. The activation in the amygdaloid region found during startle modulation by fear may reflect such an increase in vigilance. Moreover, since this was accompanied by an increase in the affective division of the anterior cingulate cortex, an extended network dedicated to threat detection and orientation to danger (including both structures) is proposed. Hypervigilance is one of the main PTSD symptoms associated with increased arousal. It is possible that the increased activity in the right amygdala observed in the PTSD patients during traumatic exposure reflects aspects of heightened vigilance, and that it also accounts for the exaggerated startle often observed in this group. Amygdala “hyperarousability”, i.e. when amygdala activity (too) easily passes above activation threshold, could lead to a state of constant apprehension; with the fear system more or less constantly activated, being on-guard, or in “threat-detection mode”, could result in activation of motor/premotor areas as well as potentiated startle.

Finally, in the PTSD study (Paper I) amygdala activation was right lateralized, whereas in the startle modulation study (Paper II) increased activity was observed in the left amygdala. One explanation for this may be the notion that the right amygdala is part of a dynamic emotional stimulus detection system, while the left is specialized for sustained stimulus evaluations, as suggested by Wright and coworkers (2001), or/and that the left amygdala activates preferentially during anticipation of aversive events, as suggested by Phelps et al. (2001). In the present thesis, thus, right amygdala activation would reflect its threat detection properties during symptom provocation in PTSD, while left amygdala activation would reflect its role in more sustained, evaluative processing or/and aspects of anticipation during affective startle modulation. Also, in the startle modulation study the left-lateralized activation pattern does not represent phobic fear per se, but the interaction effect that corresponds to startle potentiation by fear.

3.3 Future directions This thesis has answered some important questions, left many unanswered, and posed even more new ones. Invariably, limitations and drawbacks have surfaced during the course of the work, to be remedied in the future.

First, the present work has demonstrated the successful use of startle probe methodology in conjunction with functional neuroimaging. Using startle as a probe of ongoing emotional processes, several future directions are possible. It would be very informative to study the neural correlates

49

associated with the inhibition of the startle reflex. This might help clarify the possible brain mechanisms underlying emotional control, extinction, and/or symptom reduction, in clinical as well as in non-clinical samples. Also, using startle probe methodology one could map fear versus anxiety in terms of shared versus unique brain activation territory. Are there, for example, differences in the activation pattern between anxiety disorders considered highly cue-specific, such as specific phobias, and those considered more generalized and prone to high levels of baseline stress, such as generalized anxiety disorder? Are there differences in brain function in chronic as compared to acute PTSD? Also, bearing upon the distinction between fear and anxiety, are the amygdala and the bed nucleus of the stria terminalis differentially activated during startle potentiation in cue-specific versus contextual fear conditioning?

Secondly, inherent in the PET methodology is the problem of poor temporal resolution. As opposed to event-related designs using fMRI where measurement can be tightly locked on specific stimuli in time, PET data has to be sampled in “blocks” that correspond a scan's time window, i.e. about 60 seconds. This makes it difficult to study short-latency phenomena and within-session variables, such as short-term startle habituation. With the utilization of event-related fMRI methodology, however, “locking on” to the startle probe time window and assessing each response separately to study within-session (i.e. short-term) habituation would be possible. Provided, that is, that the noisy circumstances of the fMRI environment are overcome practically, or that a non-acoustic (e.g. tactile) startle paradigm is used. Because the use of fMRI is not subject to radiation protection restrictions as is PET, it also makes it possible to obtain multiple scans for the same subject, e.g. measures during multiple baselines together with startle potentiation as well as startle inhibition conditions.

50

4. Conclusions

The conclusions of this thesis can be summarized as follows:

Perceptually induced symptom provocation in posttraumatic stress disorder (PTSD) is associated with increased neural activity in the primary and secondary motor cortex, the cerebellum, the amygdala, and the periaqueductal gray matter. It is suggested that this activation pattern represents a functional network that supports emotionally determined motor preparation in a state of fear.

For the first time in humans, it is demonstrated that acoustic startle stimulation is associated with increased neural activity in a posterior medial pons area corresponding to the reticularis pontis caudalis (PnC). Altered activity in the cerebellum as a result of startle repetition points to its involvement in startle habituation. It is, thus, suggested that a similar neurocircuitry underlies startle initiation and habituation in humans and animals.

Startle modulation by phobic fear is associated with activation in the affective part of the anterior cingulate and the amygdaloid-hippocampal area. It is suggested that these areas form part of a network involved in regulation of attention driven by fear, possibly a neural system dedicated to attention and orientation to danger.

51

5. References

Adamec, R. (2001). Does long term potentiation in periacqueductal gray (PAG)

mediate lasting changes in rodent anxiety-like behavior (ALB) produced by predator stress? - Effects of low frequency stimulation (LFS) of PAG on place preference and changes in ALB produced by predator stress. Behavioural Brain Research, 120, 111-135.

Aggleton, J. P. (2000). The amygdala: A functional analysis. New York: Oxford University Press.

Allen, G., Buxton, R. B., Wong, E. C., & Courchesne, E. (1997). Attentional activation of the cerebellum independent of motor involvement. Science, 275, 1940-1943.

Amaral, D. G., Price, J. L., Pitkänen, A., & Carmichael, S. T. (1992). Anatomical organization of the primate amygdaloid complex. In J.P. Aggleton (Ed.), The amygdala: Neurobiological aspects of emotion, memory, and mental dysfunction (pp. 1-66). New York: Wiley-Liss.

American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: American Psychiatric Press.

Amorapanth, P., LeDoux, J. E., & Nader, K. (2000). Different lateral amygdala outputs mediate reactions and actions elicited by a fear-arousing stimulus. Nature Neuroscience, 3, 74-79.

Andersson, J. L. (1995). A rapid and accurate method to realign PET scans utilizing image edge information. Journal of Nuclear Medicine, 36, 657-669.

Andersson, J. L. (1997). How to estimate global activity independent of changes in local activity. Neuroimage, 6, 237-244.

Angrilli, A., Mauri, A., Palomba, D., Flor, H., Birbaumer, N., Sartori, G., & di Paola, F. (1996). Startle reflex and emotion modulation impairment after a right amygdala lesion. Brain, 119, 1991-2000.

Bandler, R., & Shipley, M. T. (1994). Columnar organization in the midbrain periaqueductal gray: modules for emotional expression? Trends in Neurosciences, 17, 379-389.

Bandura, A., & Rosenthal, T. L. (1966). Vicarious classical conditioning as a function of arousal level. Journal of Personality and Social Psychology, 3, 54-62.

Behbehani, M. M. (1995). Functional characteristics of the midbrain periaqueductal gray. Progress in Neurobiology, 46, 575-605.

Berg, W. K, & Balaban, M. T. (1999). Startle Elicitation: Stimulus parameters, recording techniques, and quantification. In M. E. Dawson, A. M. Schell, & A. H. Bohmelt (Eds.), Startle modification: Implications for neuroscience, cognitive science, and clinical science (pp. 21-50). New York: Cambridge University Press.

52

Blair, R. J., Morris, J. S., Frith, C. D., Perrett, D. I., & Dolan, R. J. (1999). Dissociable neural responses to facial expressions of sadness and anger. Brain, 122, 883-893.

Blake, D., Weathers, F., Nagy, D., Kaloupek, G., Klauminzer, D., Charney, D., & Keane, T. (1990). Clinician-Administered PTSD Scale (CAPS). Boston, MA: National Center for Post-Traumatic Stress Disorder, Behavioral Science Division Boston-VA.

Blanchard, E. B., Jones-Alexander, J., Buckley, T. C., Forneris, C. A. (1996). Psychometric properties of the PTSD Checklist (PCL). Behaviour Research and Therapy, 34, 669-673.

Bradley, M. M., Cuthbert, B. N., & Lang, P. J. (1990). Startle reflex modification: emotion or attention? Psychophysiology, 27, 513-522.

Bradley, M. M., Cuthbert, B. N., & Lang, P. J. (1999). Affect and the startle reflex. In M. E. Dawson, A. M. Schell, & A. H. Bohmelt (Eds.), Startle modification: Implications for neuroscience, cognitive science, and clinical science (pp. 157-183). New York: Cambridge University Press.

Bremner, J. D. (2002). Neuroimaging studies in post-traumatic stress disorder. Current Psychiatry Reports, 4, 254-63.

Bremner, J. D., Narayan, M., Staib, L. H., Southwick, S. M., McGlashan, T., & Charney, D. S. (1999). Neural correlates of memories of childhood sexual abuse in women with and without post traumatic stress disorder. American Journal of Psychiatry, 156, 1787-1795.

Bremner, J. D., Staib, L. H., Kaloupek, D., Southwick, S. M., Soufer, R., & Charney, D. S. (1999). Neural correlates of exposure to traumatic pictures and sound in Vietnam combat veterans with and without post traumatic stress disorder: A positron emission tomography study. Biological Psychiatry, 45, 806-816.

Brett, M., Johnsrude, I. S., & Owen, A. M. (2002). The problem of functional localization in the human brain. Nature Reviews Neuroscience, 3, 243-249.

Brown, P., Rothwell, J. C., Thompson, P. D., Britton, T. C., Day, B. L., & Marsden, C. D. (1991). New observations on the normal auditory startle reflex in man. Brain, 114, 1891-1902.

Buchel, C., Morris, J., Dolan, R. J., & Friston, K. J. (1998). Brain systems mediating aversive conditioning: an event-related fMRI study. Neuron, 20, 947-957.

Bush, G., Luu, P., & Posner, M. I. (2000). Cognitive and emotional influences in anterior cingulate cortex. Trends in Cognitive Sciences, 4, 215-222.

Butler, R. W., Braff, D. L., Rausch, J. L., Jenkins, M. A., Sprock, J., & Geyer, M. A. (1990). Physiological evidence of exaggerated startle response in a subgroup of Vietnam veterans with combat-related PTSD. American Journal of Psychiatry, 147, 1308-1312.

Cabeza, R., & Nyberg, L. (2000). Imaging cognition II: An empirical review of 275 PET and fMRI studies. Journal of Cognitive Neuroscience, 12, 1-47.

Cahill, L. (2000). Modulation of long-term memory storage in humans by emotional arousal: adrenergic activation and the amygdala. In J. P. Aggleton (Ed.), The amygdala: Neurobiological aspects of emotion, memory, and mental dysfunction (pp. 425-446). New York: Wiley-Liss.

Cahill, L., Haier, R. J., Fallon, J., Alkire, M. T., Tang, C., Keator, D., Wu, J., & McGaugh, J. L. (1996). Amygdala activity at encoding correlated with long-term, free recall of emotional information. Proceedings of the National Academy of Science, USA, 93, 8016-8021.

53

Cahill, L., Vazdarjanova, A., & Setlow, B. (2000). The basolateral amygdala complex is involved with, but is not necessary for, rapid acquisition of Pavlovian 'fear conditioning'. European Journal of Neuroscience, 12, 3044-3050.

Cannon, W. B. (1915). Bodily changes in pain, hunger, fear and rage. New York: Appleton and Company.

Charney, D. S., & Deutch, A. (1996). A functional neuroanatomy of anxiety and fear: implications for the pathophysiology and treatment of anxiety disorders. Critical Reviews in Neurobiology, 10, 419-446.

Charney, D. S., Grillon, C., & Bremner, J. D. (1998). The neurobiological basis of anxiety and fear: Circuits, mechanisms, and neurochemical interactions (part I). Neuroscientist, 4, 35-44.

Chua, P., & Dolan, R. J. (2000). The neurobiology of anxiety and anxiety-related disorders: A functional neuroimaging perspective. In J. C. Mazziota, A. W. Toga, & R. J. F. Frackowiak (Eds.), Brain mapping: The disorders (pp. 509-522). San Diego, CA: Academic Press.

Codispoti, M., Bradley, M. M., & Lang, P. J. (2001). Affective reactions to briefly presented pictures. Psychophysiology, 38, 474-478.

Collins, D. L., Neelin, P., Peters, T. M., & Evans, A. C. (1994). Automatic 3D intersubject registration of MR volumetric data in standardized Talairach space. Journal of Computer Assisted Tomography, 18, 192-205.

Cook, E. W., III (1999). Affective individual differences, psychopathology, and startle reflex modification. In M. E. Dawson, A. M. Schell, & A. H. Bohmelt (Eds.), Startle modification: Implications for neuroscience, cognitive science, and clinical science (pp. 187-208). New York: Cambridge University Press.

Courchesne, E., & Allen, G. (1997). Prediction and preparation, fundamental functions of the cerebellum. Learning and Memory, 4, 1-35.

Craig, K. D., & Weinstein, M. S. (1965). Conditioning vicarious affective arousal. Psychological Reports, 17, 955-963.

Damasio, A. R. (1994). Descartes' error: emotion, reason, and the human brain. New York: G.P. Putnam.

Damasio, A. R., Grabowski, T. J., Bechara, A., Damasio, H., Ponto, L. L., Parvizi, J., & Hichwa, R. D. (2000). Subcortical and cortical brain activity during the feeling of self-generated emotions. Nature Neuroscience, 3, 1049-1056.

Darwin, C. (1872). The expression of the emotions in man and animals. London: John Murray.

Davidson, R. J. (2002). Anxiety and affective style: role of prefrontal cortex and amygdala. Biological Psychiatry, 51, 68-80.

Davidson, R. J., & Irwin, W. (1999). The functional neuroanatomy of emotion and affective style. Trends in Cognitive Sciences, 3, 11-21.

Davidson, R. J., Putnam, K. M., & Larson, C. L. (2000). Dysfunction in the neural circuitry of emotion regulation: A possible prelude to violence. Science, 289, 591-594.

Davidson, R. J., & Sutton, S. K. (1995). Affective neuroscience: the emergence of a discipline. Current Opinion in Neurobiology, 5, 217-224.

Davis, M. (1998). Anatomic and physiologic substrates of emotion in an animal model. Journal of Clinical Neurophysiology, 15, 378-387.

Davis, M. (2000). The role of the amygdala in conditioned and unconditioned fear and anxiety. In J. P. Aggleton (Ed.), The amygdala: A functional analysis (pp. 213-287). New York: Oxford University Press.

54

Davis, M., Falls, W. A., Campeau, S., & Kim, M. (1993). Fear-potentiated startle: a neural and pharmacological analysis. Behavioural Brain Research, 58, 175-198.

Davis, M., Walker, D. L., & Lee, Y. (1999). Neurophysiology and neuropharmacology of startle and its affective modification. In M. E. Dawson, M. E. Schell, & A. H. Bohmelt (Eds.), Startle modification: Implications for neuroscience, cognitive science, and clinical science (pp. 95-113). New York: Cambridge University Press.

Davis, M., & Whalen, P. J. (2001). The amygdala: vigilance and emotion. Molecular Psychiatry, 6, 13-34.

Dawson, M. E., Schell, A. M., & Bohmelt, A. H. (Eds.). (1999). Startle modification: Implications for neuroscience, cognitive science, and clinical science. New York: Cambridge University Press.

Desmond, J. E., & Fiez, J. A. (1998). Neuroimaging studies of the cerebellum: Language, learning and memory. Trends in Cognitive Sciences, 2, 355-262.

Devinsky, O., Morrell, M. J., & Vogt, B. A. (1995). Contributions of anterior cingulate cortex to behaviour. Brain, 118, 279-306.

Dolan, R. J., Lane, R., Chua, P., & Fletcher, P. (2000). Dissociable temporal lobe activations during emotional episodic memory retrieval. Neuroimage, 11, 203-209.

Elliott, R., Rubinsztein, J. S., Sahakian, B. J., & Dolan, R. J. (2000). Selective attention to emotional stimuli in a verbal go/no-go task: an fMRI study. Neuroreport, 11, 1739-1744.

Fernandez, M., Pissiota, A., Frans, O., von Knorring, L., Fischer, H., & Fredrikson, M. (2001). Brain function in a patient with torture related post-traumatic stress disorder before and after fluoxetine treatment: a positron emission tomography provocation study. Neuroscience Letters, 297, 101-104.

Fink, G. R., Markowitsch, H. J., Reinkemeier, M., Bruckbauer, T., Kessler, J., & Heiss, W. D. (1996). Cerebral representation of one's own past: neural networks involved in autobiographical memory. Journal of Neuroscience, 16, 4275-4282.

First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1995). Structured clinical interview for DSM-IV. New York: New York State Psychiatric Institute, Biometrics Research Department.

Fischer, H., Tillfors, M., Furmark, T., & Fredrikson, M. (2001). Dispositional pessimism and amygdala activity: a PET study in healthy volunteers. Neuroreport, 12, 1635-1638.

Fischer, H., Wik, G., & Fredrikson, M. (1996). Functional neuroanatomy of robbery re-experience: Affective memories studied with PET. Neuroreport, 7, 2081-2086.

Frackowiak, R. S. J., Friston, K. J., Frith, C. D., Dolan, R. J., & Mazziota, J. C. (Eds.). (1997). Human brain function. London: Academic Press.

Fredrikson, M., Fischer, H., & Wik, G. (1997). Cerebral blood flow during anxiety provocation. Journal of Clinical Psychiatry, 58 (Suppl. 16), 16-21.

Fredrikson, M., Wik, G., Annas, P., Ericson, K., & Stone-Elander, S. (1995). Functional neuroanatomy of visually elicited simple phobic fear: additional data and theoretical analysis. Psychophysiology, 32, 43-48.

Fredrikson, M., Wik, G., Fischer, H., & Andersson, J. (1995). Affective and attentive neural networks in humans: A PET study of Pavlovian conditioning. Neuroreport, 7, 97–101.

55

Fredrikson, M., Wik, G., Greitz, T., Eriksson, L., Stone-Elander, S., Ericson, K., & Sedvall, G. (1993). Regional cerebral blood flow during experimental phobic fear. Psychophysiology, 30, 126-130.

Frings, M., Maschke, M., Erichsen, M., Jentzen, W., Muller, S. P., Kolb, F. P., Diener, H. C., & Timmann, D. (2002). Involvement of the human cerebellum in fear-conditioned potentiation of theacoustic startle response: a PET study. Neuroreport, 13, 1275-1278.

Friston, K. J., Holmes, A. P., Worsley, K. J., Poline, J. B., Frith, C. D., & Frackowiak, R. S. J. (1995). Statistical parametric maps in functional imaging: A general linear approach. Human Brain Mapping, 2, 189-210.

Friston, K. J., & Price, C. J. (2001). Generative models, brain function and neuroimaging. Scandinavian Journal of Psychology, 42, 167-177.

Friston, K. J., Price, C. J., Buchel, C., & Frackowiak, R. S. J. (1997). A taxonomy of study designs. In R. S. J. Frackowiak, K. J. Friston, C. D. Frith, R. J. Dolan, & J. C. Mazziota (Eds.), Human brain function (pp. 141-159). London: Academic Press.

Friston, K. J., Worsley, K. J., Frackowiak, R. S. J., Mazziotta, J. C., & Evans, A. C. (1994). Assessing the significance of focal activations using their spatial extent. Human Brain Mapping, 1, 210-220.

Funayama, E. S., Grillon, C., Davis, M., & Phelps, E. A. (2001). A double dissociation in the affective modulation of startle in humans: effects of unilateral temporal lobectomy. Journal of Cognitive Neuroscience, 13, 721-729.

Furmark, T., Fischer, H., Wik, G., Larsson, M., & Fredrikson, M. (1997). The amygdala and individual differences in human fear conditioning. Neuroreport, 8, 3957-3960.

Garpenstrand, H., Annas, P., Ekblom, J., Oreland, L., & Fredrikson, M. (2001). Human fear conditioning is related to dopaminergic and serotonergic biological markers. Behavioral Neuroscience, 115, 358-364.

Gloor, P. (1992). Role of the amygdala in temporal lobe epilepsy. In J. P. Aggleton (Ed.), The amygdala: Neurobiological aspects of emotion, memory, and mental dysfunction (pp. 505-538). New York: Wiley-Liss.

Goddard, A. W., & Charney, D. S. (1997). Toward an integrated neurobiology of panic disorder. Journal of Clinical Psychiatry, 58 (Suppl 2.), 4-11.

Gray, J. A. (1987). The psychology of fear and stress (2nd ed.). Cambridge: Cambridge University Press.

Greitz, T., Bohm, C., Holte, S., & Eriksson, L. (1991). A computerized brain atlas: construction, anatomical content, and some applications. Journal of Computer Assisted Tomography, 15, 26-38.

Grillon, C., Morgan, C. A., 3rd, Davis, M., & Southwick, S. M. (1998). Effects of experimental context and explicit threat cues on acoustic startle in Vietnam veterans with posttraumatic stress disorder. Biological Psychiatry, 44, 1027-1036.

Grillon, C., Morgan, C. A., Southwick, S. M., Davis, M., & Charney, D. S. (1996). Baseline startle amplitude and prepulse inhibition in Vietnam veterans with posttraumatic stress disorder. Psychiatry Research, 64, 169-178.

Groenink, L., Joordens, R. J., Hijzen, T. H., Dirks, A., & Olivier, B. (2000). Infusion of flesinoxan into the amygdala blocks the fear-potentiated startle. Neuroreport, 11, 2285-2288.

56

Hamm, A. O., Cuthbert, B. N., Globisch, J., & Vaitl, D. (1997). Fear and the startle reflex: blink modulation and autonomic response patterns in animal and mutilation fearful subjects. Psychophysiology, 34, 97-107.

Hariri, A. R., Bookheimer, S. Y., & Mazziotta, J. C. (2000). Modulating emotional responses: effects of a neocortical network on the limbic system. Neuroreport, 11, 43-48.

Hariri, A. R., Mattay, V. S., Tessitore, A., Kolachana, B., Fera, F., Goldman, D., Egan, M. F., & Weinberger, D. R. (2002). Serotonin transporter genetic variation and the response of the human amygdala. Science, 297, 400-403.

Hazlett, E. A., Buchsbaum, M. S., Tang, C. Y., Fleischman, M. B., Wei, T. C., Byne, W., & Haznedar, M. M. (2001). Thalamic activation during an attention-to-prepulse startle modification paradigm: a functional MRI study. Biological Psychiatry, 50, 281-291.

Hendelman, W. J. (1994). Student’s atlas of neuroanatomy. Philadelphia: WB Saunders Co.

Hettema, J. M., Annas, P., Neale, M. C., Kendler, M. D., & Fredrikson, M. (in press). A twin study of the genetics of fear conditioning. Archives of General Psychiatry.

Holstege, G., Tan, J., van Ham, J. J., & Graveland, G. A. (1986). Anatomical observations on the afferent projections to the retractor bulbi motoneuronal cell group and other pathways possibly related to the blink reflex in the cat. Brain Research, 374, 321-334.

Hull, A. M. (2002). Neuroimaging findings in post-traumatic stress disorder. Systematic review. British Journal of Psychiatry, 181, 102-110.

James, W. (1884/1969). What is an emotion? In Collected essays and reviews. New York: Russell and Russell.

Jansen, D. M., & Frijda, N. H. (1994). Modulation of the acoustic startle response by film-induced fear and sexual arousal. Psychophysiology, 31, 565-571.

Kendler, K. S., Myers, J., & Prescott, C. A. (2002). The etiology of phobias: an evaluation of the stress-diathesis model. Archives of General Psychiatry, 59, 242-248.

Klorman, R., Weerts, T. C., Hastings, J. E., Melamed, B. G., & Lang, P. J. (1974). Psychometric description of some specific fear questionnaires. Behavior Therapy, 5, 401-409.

Koch, M. (1999). The neurobiology of startle. Progress in Neurobiology, 59, 107-128.

LaBar, K. S., Gatenby, J. C., Gore, J. C., LeDoux, J. E., & Phelps, E. A. (1998). Human amygdala activation during conditioned fear acquisition and extinction: a mixed-trial fMRI study. Neuron, 20, 937-945.

LaBar, K. S., LeDoux, J. E., Spencer, D. D., & Phelps, E. A. (1995). Impaired fear conditioning following unilateral temporal lobectomy in humans. Journal of Neuroscience, 15, 6846-6855.

Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1990). Emotion, attention, and the startle reflex. Psychological Review, 97, 377-395.

Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1998). Emotion, motivation, and anxiety: brain mechanisms and psychophysiology. Biological Psychiatry, 44, 1248-1263.

Lang, P. J., Davis, M., & Öhman, A. (2000). Fear and anxiety: animal models and human cognitive psychophysiology. Journal of Affective Disorders, 61, 137-159.

57

Leaton, R. N., & Supple, W. F., Jr. (1991). Medial cerebellum and long-term habituation of acoustic startle in rats. Behavioral Neuroscience, 105, 804-816.

Ledberg, A. (2001). Measuring brain functions: Statistical tests for neuroimaging data. Doctoral dissertation, Karolinska Institutet, Stochholm, Sweden.

LeDoux, J. (1996). The emotional brain: The mysterious underpinnings of emotional life. New York: Simon and Schuster.

LeDoux, J. E. (2000). Emotion circuits in the brain. Annual Review of Neuroscience, 23, 155-184.

Lee, Y., Lopez, D. E., Meloni, E. G., & Davis, M. (1996). A primary acoustic startle pathway: obligatory role of cochlear root neurons and the nucleus reticularis pontis caudalis. Journal of Neuroscience, 16, 3775-3789.

Liberzon, I., Taylor, S. F., Amdur, R., Jung, T. D., Chamberlain, K. R., Minoshima, S., Koeppe, R. A., & Fig, L. M. (1999). Brain activation in PTSD in response to trauma-related stimuli. Biological Psychiatry, 45, 817-826.

Lopiano, L., DeSperati, C., & Montarolo, P. G. (1990). Long-term habituation of the acoustic startle response: role of the cerebellar vermis. Neuroscience, 35, 79-84.

Maddock, R. J. (1999). The retrosplenial cortex and emotion: new insights from functional neuroimaging of the human brain. Trends in Neurosciences, 22, 310-316.

Mai, J. K., Assheuer, J., & Paxinos, G. (1997). Atlas of the human brain. London: Academic Press.

McGaugh, J. L., Ferry, B., Vazdarjanova, A., & Roozendaal, B. (2000). Amygdala: Role in modulation of memory storage. In J. P. Aggleton (Ed.), The amygdala: Neurobiological aspects of emotion, memory, and mental dysfunction (pp. 391-424). New York: Wiley-Liss.

MacLean, P. D. (1949). Psychosomatic disease and the “visceral brain”: Recent developments bearing on the Papez theory of emotion. Psychosomatic Medicine, 11, 338-353.

Maddock, R. J. (1999). The retrosplenial cortex and emotion: new insights from functional neuroimaging of the human brain. Trends in Neurosciences, 22, 310-316.

Mineka, S., & Zinbarg, R. (1996). Conditioning and ethological models of anxiety disorders: stress-in-dynamic-context anxiety models. Nebraska Symposium on Motivation, 43, 135-210.

Morgan, C. A., 3rd, Grillon, C., Southwick, S. M., Davis, M., & Charney, D. S. (1995). Fear-potentiated startle in posttraumatic stress disorder. Biological Psychiatry, 38, 378-385.

Morgan, C. A., 3rd, Grillon, C., Southwick, S. M., Davis, M., & Charney, D. S. (1996). Exaggerated acoustic startle reflex in Gulf War veterans with posttraumatic stress disorder. American Journal of Psychiatry, 153, 64-68.

Morris, J. S., Friston, K. J., Buchel, C., Frith, C. D., Young, A. W., Calder, A. J., & Dolan, R. J. (1998). A neuromodulatory role for the human amygdala in processing emotional facial expressions. Brain, 121, 47-57.

Morris, J. S., Frith, C. D., Perrett, D. I., Rowland, D., Young, A. W., Calder, A. J., & Dolan, R. J. (1996). A differential neural response in the human amygdala to fearful and happy facial expressions. Nature, 383, 812-815.

Morris, J. S., Öhman, A., & Dolan, R. J. (1998). Conscious and unconscious emotional learning in the human amygdala. Nature, 393, 467-470.

58

Öhman, A. (1986). Face the beast and fear the face: animal and social fears as prototypes for evolutionary analyses of emotion. Psychophysiology, 23, 123-145.

Öhman, A., & Soares, J. J. (1998). Emotional conditioning to masked stimuli: expectancies for aversive outcomes following nonrecognized fear-relevant stimuli. Journal of Experimental Psychology: General, 127, 69-82.

Orr, S. P., Lasko, N. B., Shalev, A. Y., & Pitman, R. K. (1995). Physiologic responses to loud tones in Vietnam veterans with posttraumatic stress disorder. Journal of Abnormal Psychology, 104, 75-82.

Orr, S. P., Metzger, L. J., & Pitman, R. K. (2002). Psychophysiology of post-traumatic stress disorder. Psychiatric Clinics of North America, 25, 271-293.

Öst, L. G. (1985). Ways of acquiring phobias and outcome of behavioral treatments. Behaviour Research and Therapy, 23, 683-689.

Papez, J. W. (1937). A proposed mechanism of emotion. Archives of Neurology and Psychiatry, 38, 725-743.

Pavlov, I. P. (1927/1960). Conditioned reflexes: An investigation of the physiological activity of the cerebral cortex. New York: Dover Publications.

Phelps, E. A., O’Connor, K. J., Gatenby, J. C., Gore, J. C., Grillon, C., & Davis, M. (2001). Activation of the left amygdala to a cognitive representation of fear. Nature Neuroscience, 4, 437-441.

Pitman, R. K., Orr, S. P., Shalev, A. Y., Metzger, L. J., & Mellman, T. A. (1999). Psychophysiological alterations in post traumatic stress disorder. Seminars in Clinical Neuropsychiatry, 4, 234-241.

Ploghaus, A., Narain, C., Beckmann, C. F., Clare, S., Bantick, S., Wise, R., Matthews, P. M., Rawlins, J. N., & Tracey, I. (2001). Exacerbation of pain by anxiety is associated with activity in a hippocampal network. Journal of Neuroscience, 21, 9896-9903.

Posner, M. I., & Raichle, M. E. (1994). Images of mind. New York: Scientific American Library.

Rauch, S. L., Jenike, M. A., Alpert, N. M., Baer, L., Breiter, H. C., Savage, C. R., & Fischman, A. J. (1994). Regional cerebral blood flow measured during symptom provocation in obsessive-compulsive disorder using oxygen 15-labeled carbon dioxide and positron emission tomography. Archives of General Psychiatry, 51, 62-70.

Rauch, S. L., Savage, C. R., Alpert, N. M., Fischman, A. J., & Jenike, M. A. (1997). The functional neuroanatomy of anxiety: a study of three disorders using positron emission tomography and symptom provocation. Biological Psychiatry, 42, 446-452.

Rauch, S. L., Savage, C. R., Alpert, N. M., Miguel, E. C., Baer, L., Breiter, H. C., Fischman, A. J., Manzo, P. A., Moretti, C., & Jenike, M. A. (1995). A positron emission tomographic study of simple phobic symptom provocation. Archives of General Psychiatry, 52, 20-28.

Rauch, S. L., van der Kolk, B. A., Fisler, R. E., Alpert, N. M., Orr, S. P., Savage, C. R., Fischman, A. J., Jenike, M. A., & Pitman, R. K. (1996). A symptom provocation study of post traumatic stress disorder using positron emission tomography and script-driven imagery. Archives of General Psychiatry, 53, 380-387.

Sabatinelli, D., Bradley, M. M., & Lang, P. J. (2001). Affective startle modulation in anticipation and perception. Psychophysiology, 38, 719-722.

59

Seligman, M. E. P. (1971). Phobias and preparedness. Behavior Therapy, 2, 307-321.

Shin, L. M., Kosslyn, S. M., McNally, R. J., Alpert, N. M., Thompson, W. L., Rauch, S. L., Macklin, M. L., & Pitman, R. K. (1997). Visual imagery and perception in post traumatic stress disorder. A positron emission tomographic investigation. Archives of General Psychiatry, 54, 233-241.

Schmahmann, J. (1998). Dysmetria of thought: Clinical consequences of cerebellar dysfunction on cognition and affect. Trends in Cognitive Sciences, 2, 362-371.

Spielberger, C. D., Gorsuch, R. L., & Lushene, R. (1983). Manual for the State-Trait Anxiety (STAI Form Y). Palo Alto, CA: Consulting Psychologists Press.

Squire, L. R., & Zola-Morgan, S. (1991). The medial temporal lobe memory system. Science, 253, 1380-1386.

Stein, J. F. (1986). Role of the cerebellum in the visual guidance of movement. Nature, 323, 217-222.

Swanson, L. W., & Petrovich, G. D. (1998). What is the amygdala? Trends in Neurosciences, 21, 323-331.

Tackmann, W., Ettlin, T., & Barth, R. (1982). Blink reflexes elicited by electrical, acoustic and visual stimuli. I. Normal values and possible anatomical pathways. European Neurology, 21, 210-216.

Talairach, J., & Tournoux, P. (1988). A co-planar stereotactic atlas of the human brain: 3-dimensional proportional system: An approach to cerebral imaging. Stuttgart, Germany: Thieme Verlag.

Timmann, D., Musso, C., Kolb, F. P., Rijntjes, M., Juptner, M., Muller, S. P., Diener, H. C., & Weiller, C. (1998). Involvement of the human cerebellum during habituation of the acoustic startle response: A PET study. Journal of Neurology, Neurosurgery, and Psychiatry, 65, 771-773.

Toga, A. W., & Mazziotta, J. C. (Eds.). (1996). Brain mapping: The methods. San Diego, CA: Academic Press.

van der Kolk, B. A. (1997). The psychobiology of posttraumatic stress disorder. Journal of Clinical Psychiatry, 58 (Suppl. 9), 16-24.

van der Kolk, B. A., & Fisler, R. (1995). Dissociation and the fragmentary nature of traumatic memories: overview and exploratory study. Journal of Traumatic Stress, 8, 505-525.

Vogt, B. A., Finch, D. M., & Olson, C. R. (1992). Functional heterogeneity in cingulate cortex: the anterior executive and posterior evaluative regions. Cerebral Cortex, 2, 435-443.

Vuilleumier, P., Armony, J. L., Driver, J., & Dolan, R. J. (2001). Effects of attention and emotion on face processing in the human brain: an event-related fMRI study. Neuron, 30, 829-841.

Walker, D. L., & Davis, M. (2002). The role of amygdala glutamate receptors in fear learning, fear-potentiated startle, and extinction. Pharmacology, Biochemistry, and Behavior, 71, 379-392.

Watson, J. B., & Rayner, R. (1920). Conditioned emotional reactions. Journal of Experimental Psychology, 3, 1-4.

Weathers, F. W., Litz, B. T., Herman, D. S., Huska, J. A., & Keane, T. M. (1993). The PTSD checklist: reliability, validity & diagnostic utility. Paper presented at the Annual Meeting of the International Society for Traumatic Stress Studies, San Antonio, TX.

Whalen, P. J., Bush, G., McNally, R. J., Wilhelm, S., McInerney, S. C., Jenike, M. A., & Rauch, S. L. (1998). The emotional counting Stroop paradigm: a

60

functional magnetic resonance imaging probe of the anterior cingulate affective division. Biological Psychiatry, 44, 1219-1228.

Wik, G., Fredrikson, M., Ericson, K., Eriksson, L., Stone-Elander, S., & Greitz, T. (1993). A functional cerebral response to frightening visual stimulation. Psychiatry Research, 50, 15-24.

Wolpe, J., & Lang, P. J. (1964). A fear survey schedule for use in behaviour therapy. Behaviour Research and Therapy, 2, 27-30.

Wood, J. N., & Grafman, J. (2003). Human prefrontal cortex: processing and representational perspectives. Nature Reviews Neuroscience, 4, 139-147.

Wright, C. I., Fischer, H., Whalen, P. J., McInerney, S. C., Shin, L. M., & Rauch, S. L. (2001). Differential prefrontal cortex and amygdala habituation to repeatedly presented emotional stimuli. Neuroreport, 12, 379-383.

Yartz, A. R., & Hawk, L. W, Jr. (2002). Addressing the specificity of affective startle modulation: fear versus disgust. Biological Psychology, 59, 55-68.

61

Appendix A

The DSM-IV diagnostic criteria for Posttraumatic Stress Disorder (American Psychiatric Association, 1994, p. 427).

A. The person has been exposed to a traumatic event in which both of the following were present: (1) the person experienced, witnessed, or was confronted with an event or events that involved actual or threatened death or serious injury, or a threat to the physical integrity of self or others (2) the person's response involved intense fear, helplessness, or horror. Note: In children, this may be expressed instead by disorganized or agitated behavior

B. The traumatic event is persistently reexperienced in one (or more) of the following ways: (1) recurrent and intrusive distressing recollections of the event, including images, thoughts, or perceptions. Note: In young children, repetitive play may occur in which themes or aspects of the trauma are expressed. (2) recurrent distressing dreams of the event. Note: In children, there may be frightening dreams without recognizable content. (3) acting or feeling as if the traumatic event were recurring (includes a sense of reliving the experience, illusions, hallucinations, and dissociative flashback episodes, including those that occur on awakening or when intoxicated). Note: In young children, trauma-specific reenactment may occur. (4) intense psychological distress at exposure to internal or external cues that symbolize or resemble an aspect of the traumatic event physiological reactivity on exposure to internal or external cues that symbolize or resemble an aspect of the traumatic event

C. Persistent avoidance of stimuli associated with the trauma and numbing of general responsiveness (not present before the trauma), as indicated by three (or more) of the following: (1) efforts to avoid thoughts, feelings, or conversations associated with the trauma (2) efforts to avoid activities, places, or people that arouse recollections of the trauma (3) inability to recall an important aspect of the trauma (4) markedly diminished interest or participation in significant activities (5)feeling of detachment or estrangement from others (6) restricted range of affect (e.g., unable to have loving feelings)

62

(7) sense of a foreshortened future (e.g., does not expect to have a career, marriage, children, or a normal life span)

D. Persistent symptoms of increased arousal (not present before the trauma), as indicated by two (or more) of the following: (1) difficulty falling or staying asleep (2) irritability or outbursts of anger (3) difficulty concentrating (4) hypervigilance (5) exaggerated startle response

E. Duration of the disturbance (symptoms in Criteria B, C, and D) is more than 1 month.

F. The disturbance causes clinically significant distress or impairment in social, occupational, or other important areas of functioning.

Specify if: Acute: if duration of symptoms is less than 3 months Chronic: if duration of symptoms is 3 months or more Specify if: With Delayed Onset: if onset of symptoms is at least 6 months after the stressor

63

Appendix B

The DSM-IV diagnostic criteria for Specific Phobia (American Psychiatric Association, 1994, p. 410).

A. Marked and persistent fear that is excessive or unreasonable, cued by the presence or anticipation of a specific object or situation (e.g., flying, heights, animals, receiving an injection, seeing blood).

B. Exposure to the phobic stimulus almost invariably provokes an immediate anxiety response, which may take the form of a situationally bound or situationally predisposed Panic Attack. Note: In children, the anxiety may be expressed by crying, tantrums, freezing, or clinging.

C. The person recognizes that the fear is excessive or unreasonable. Note: In children, this feature may be absent.

D. The phobic situation(s) is avoided or else is endured with intense anxiety or distress.

E. The avoidance, anxious anticipation, or distress in the feared situation(s) interferes significantly with the person’s normal routine, occupational (or academic) functioning, or social activities or relationships, or there is marked distress about having the phobia.

F. In individuals under age 18 years, the duration is at least 6 months.

G. The anxiety, Panic Attacks, or phobic avoidance associated with the specific object or situation are not better accounted for by another mental disorder, such as Obsessive-Compulsive Disorder (e.g., fear of dirt in someone with an obsession about contamination), Posttraumatic Stress Disorder (e.g., avoidance of stimuli associated with a severe stressor), Separation Anxiety Disorder (e.g., avoidance of school), Social Phobia (e.g., avoidance of social situations because of fear of embarrassment), Panic Disorder with Agoraphobia, or Agoraphobia Without History of Panic Disorder.

64

Specify type: Animal Type Natural Environmental Type (e.g. heights, storms, water) Blood-Injection-Injury Type Situational Type (e.g. airplanes, elevators, enclosed places) Other Type (e.g. phobic avoidance of situations that may lead to choking, vomiting, or contracting an illness; in children, avoidance of loud sounds or costumed characters)

65

Acknowledgements

I owe many people a great deal for helping me through this project. First of all, I want to thank my academic supervisor Mats Fredrikson for excellent scientific guidance, support, patience, and for his enthusiastic approach to science, and, well, everything.

Many thanks to all the members of the affective neuropsychology research group. Particular thanks to my senior colleagues, Örjan Frans, Maria Tillfors, and Tomas Furmark, for pleasant collaboration, great fun, and friendship during the past five years. To my junior colleagues and friends, Åsa Michelgård of the Order of Knäckehäxan and Fredrik Åhs, for great company and collaboration – I feel that the psychophysiology lab is safe with you. To Håkan Fischer for introducing me to the realm of PET-data analysis in the first place, for his friendly ways and his good advice. To my friend co-author Manuel Fernandez for stimulating collaboration in the most challenging of studies.

I am grateful to Claes von Hofsten and Kerstin Rosander for valuable comments on a preliminary version of this thesis, and to Andreas Birgegård and Henrik Olsson for proofreading the manuscript and offering useful feedback. I am also thankful to the staff of the Department of Psychology at Uppsala University and of the Uppsala Research Imaging Solutions (former Uppsala University PET Centre) for a supportive and stimulating working environment.

I am greatly indebted to the study participants for their cooperation and their great courage in facing what they most fear.

I wish to thank my parents, Ulla and Georgios, for always caring and for showing me, with their doings, that all is not foreseeable and nothing is impossible. To my sisters, Sofia and Despina, for always being very close to me, in spite of the geographical distance.

Lastly, to Conny, my best friend all categories, thank you for your love and for putting up with me and my projects. I promise that my next enterprise will involve fewer bunkers and more fairways...