<<

,

The Author and the

JOSEPH E. LEDOUX is in- terested in the neural founda- tion of memory and emotion. Brain He studies the anatomy, physiology and behavioral organization of these aspects of mental functioning. Le- The neural routes underlying Doux, the Henry and Lucy Moses Professor of Science at the formation of about New York University, is the recipient of two National In- primitive emotional experiences, stitute of Mental Health dis- tinctions: a Merit Award and such as , have been traced a Research Scientist Develop- ment Award. He has also re- ceived an Established Investi- by Joseph E. LeDoux gator Award from the Ameri- can Heart Association.

espite millennia of preoccupation with every of emotion, we are still far from explaining in a rigorous physiological sense this part of our mental experience. Neuroscientists have, in modern times, been especially concerned with the neural basis of such cognitive pro- Dcesses as and memory. They have for the most part ignored the brain’s role in emotion. Yet in recent years, in this mysterious mental terrain has surged. Catalyzed by breakthroughs in understanding the neural basis of and by an increasingly sophisticated knowledge of the anatomical organization and physiology of the brain, investigators have begun to tackle the problem of emotion. One quite rewarding area of research has been the inquiry into the relation between memory and emotion. Much of this examination has involved studies of one particu- lar emotion—fear—and the manner in which specific events or stimuli come, through individual experiences, to evoke this state. Scientists, myself included, have Memories of disturbing been able to determine the way in which the brain shapes how we form memories experiences form deep within about this basic, but significant, emotional event. We call this process “emotional our brains. memory.” By uncovering the neural pathways through which a situation causes a creature to learn about fear, we to elucidate the general mechanisms of this form of memo- ry. Because many human mental disorders—including , , post-traumat- ic stress syndrome and attack—involve malfunctions in the brain’s ability to control fear, studies of the neural basis of this emotion may help us further understand and treat these disturbances. Most of our knowledge about how the brain links memory and emotion has been gleaned through the study of so-called classical . In this process the subject, usually a rat, hears a noise or sees a flashing light that is paired with a brief, mild electric shock to its feet. After a few such experiences, the rat responds automat- ically to the sound or light, even in the absence of the shock. Its reactions are typical to any threatening situation: the animal freezes, its blood pressure and heart rate in-

SPL/Photo Researchers, Inc. Inc. SPL/Photo Researchers, crease, and it startles easily. In the language of such experiments, the noise or flash is a conditioned , the foot shock is an unconditioned stimulus, and the rat’s reac- tion is a conditioned response, which consists of readily measured behavioral and

MEHAM KULYK MEHAM KULYK physiological changes.

68 Mysteries of the Mind Reprinted from the June 1994 issue Copyright 1997 Scientific American, Inc. Conditioning of this kind happens quickly in rats—indeed, it takes place as rapidly as it does in . A single pairing of the shock to the sound or sight can bring on the conditioned ef- fect. Once established, the fearful reac- tion is relatively permanent. If the noise or light is administered many times with- out an accompanying electric shock, the rat’s response diminishes. This change is called . But considerable evi- dence suggests that this behavioral al- teration is the result of the brain’s con- trolling the fear response rather than the elimination of the emotional memo- ry. For example, an apparently extin- guished fear response can recover spon- taneously or can be reinstated by an ir- relevant stressful experience. Similarly, stress can cause the reappearance of in people who have been suc-

AUDITORY CORTEX O OSTI T OBER R

AUDITORY HIPPOCAMPUS

ANATOMY OF EMOTION includes several regions of the brain. Shown here in the rat (above), parts of the amygdala, the thalamus and the cortex interact to create memories about fearful experiences opic

associated, in this case, with sound. Re- osc cent work has located precise areas where icr fear is learned and remembered: certain APL M parts of the thalamus (light pink at top right) communicate with areas in the amygdala (light yellow at bottom right) W LEONARD that process the fear-causing sound stim- uli. Because these neural mechanisms are HS BY ANDRE thought to be similar in humans, the study AP OGR of emotional memory in rodents may illu- T HO minate aspects of fear disorders in people. P

Emotion, Memory and the Brain Mysteries of the Mind 69 Copyright 1997 Scientific American, Inc. SOUND

ELECTRICITY

20 10 20 10 20 10 ) ORPOLE IAN W 8 8 8 15 15 15

6 6 6

10 10 10 MILLIMETERS OF MERCURY 4 4 4 BLOOD PRESSURE BLOOD PRESSURE CESSATION OF MOVEMENT 5 5 CESSATION OF MOVEMENT 5 2 2 2 BLOOD PRESSURE (

0 CESSATION OF MOVEMENT (CUMULATIVE SECONDS) 0 0 0 0 0 0 10 0 10 0 10 0 10 0 10 0 10 TIME (SECONDS) CLASSICAL FEAR CONDITIONING can be brought about shock (center). After several such pairings, the rat’s blood pres- by pairing a sound and a mild electric shock to the foot of a rat. sure rises at the same time that the animal holds still for an ex- In one set of experiments, the rat hears a sound (left), which has tended period when it hears the sound. The rat has been fear- little effect on the animal’s blood pressure or patterns of move- conditioned (right): sound alone achieves the same physiological ment. Next, the rat hears the same sound, coupled with a foot changes as did sound and shock together. cessfully treated. This resurrection dem- or humans, for that matter—encounter because the shock modifies the way in onstrates that the emotional memory un- in their daily lives. The novelty and ir- which neurons in certain important re- derlying the phobia was rendered dor- relevance of these lights and sounds help gions of the brain interpret the sound mant rather than erased by treatment. to ensure that the animals have not al- stimulus. These critical neurons are ready developed strong emotional reac- thought to be located in the neural path- Fear and Emotional Memory tions to them. So researchers are clearly way through which the sound elicits the observing learning and memory at conditioned response. ear conditioning has proved an ide- work. At the same time, such cues do During the past 10 years, researchers Fal starting point for studies of emo- not require complicated cognitive pro- in my laboratory, as well as in others, tional memory for several reasons. First, cessing from the brain. Consequently, have identified major components of this it occurs in nearly every animal group the stimuli permit us to study emotional system. Our study began at Cornell Uni- in which it has been examined: fruit mechanisms relatively directly. Finally, versity Medical College, where I worked flies, snails, birds, lizards, fish, rabbits, our extensive knowledge of the neural several years ago, when my colleagues rats, monkeys and people. Although no pathways involved in processing acous- and I asked a simple question: Is the au- one claims that the mechanisms are pre- tic and visual information serves as an ditory cortex required for auditory fear cisely the same in all these creatures, it excellent starting point for examining conditioning? seems clear from studies to date that the the neurological foundations of fear In the auditory pathway, as in other pathways are very similar in elicited by such stimuli. sensory systems, the cortex is the high- and possibly in all vertebrates. We there- My work has focused on the cerebral est level of processing; it is the culmina- fore are confident in believing that many roots of learning fear, specifically fear tion of a sequence of neural steps that of the findings in animals apply to hu- that has been induced in the rat by as- starts with the peripheral sensory recep- mans. In addition, the kinds of stimuli sociating sounds with foot shock. As do tors, located, in this case, in the ear. If most commonly used in this type of most other investigators in the field, I lesions in (or surgical removal of) parts conditioning are not signals that rats— assume that fear conditioning occurs of the auditory cortex interfered with

70 Mysteries of the Mind Emotion, Memory and the Brain Copyright 1997 Scientific American, Inc. fear conditioning, we could conclude J. Reis and I looked again and found work suggested that the central nucleus O OSTI T that the region is indeed necessary for that, in fact, cells in some regions of the was a crucial part of the system through OBER

R this activity. We could also deduce that auditory thalamus also give rise to fibers which autonomic conditioned respons- the next step in the conditioning path- that reach several subcortical locations. es are expressed. way would be an output from the audi- Could these neural projections be the In a similar vein, we found that lesions tory cortex. But our lesion experiments connections through which the stimu- of this nucleus prevented a rat’s blood in rats confirmed what a series of other lus elicits the response we identify with pressure from rising and limited its abil- studies had already suggested: the audi- fear? We tested this hypothesis by mak- ity to freeze in the presence of a fear- tory cortex is not needed in order to ing lesions in each one of the subcorti- causing stimulus. We also demonstrat- learn many things about simple acous- cal regions with which these fibers con- ed, in turn, that lesions of areas to which tic stimuli. nect. The damage had an effect in only the central nucleus connects eliminated We then went on to make lesions in one area: the amygdala. one or the other of the two responses. the auditory thalamus and the auditory Michael Davis and his associates at Yale midbrain, sites lying immediately below Filling in the Picture University determined that lesions of the the auditory cortex. Both these areas central nucleus, as well as lesions of an- process auditory signals: the midbrain hat observation suddenly created a other brain stem area to which the cen- provides the major input to the thala- Tplace for our findings in an already tral nucleus projects, diminished yet an- mus; the thalamus supplies the major accepted picture of emotional process- other conditioned response: the increased input to the cortex. Lesions in both re- ing. For a long time, the amygdala has startle reaction that occurs when an an- gions completely eliminated the rat’s been considered an important brain re- imal is afraid. susceptibility to conditioning. This dis- gion in various forms of emotional be- The findings from various laborato- covery suggested that a sound stimulus havior. In 1979 Bruce S. Kapp and his ries studying different species and mea- is transmitted through the auditory sys- colleagues at the University of Vermont suring fear in different ways all impli- tem to the level of the auditory thalamus reported that lesions in the amygdala’s cated the central nucleus as a pivotal but that it does not have to reach the central nucleus interfered with a rab- component of fear-conditioning circuit- cortex for fear conditioning to occur. bit’s conditioned heart rate response ry. It provides connections to the vari- This possibility was somewhat puz- once the animal had been given a shock ous brain stem areas involved in the con- zling. We knew that the primary nerve paired with a sound. The central nucle- trol of a spectrum of responses. fibers that carry signals from the audi- us connects with areas in the brain stem Despite our deeper understanding of tory thalamus extend to the auditory involved in the control of heart rate, this site in the amygdala, many details cortex. So David A. Ruggiero, Donald respiration and vasodilation. Kapp’s of the pathway remained hidden. Does

AUDITORY PATHWAY EFFECTS OF LESIONS IMPLICATION OF LESIONS IN THE RAT BRAIN ON FEAR CONDITIONING FOR FEAR-CONDITIONING PATHWAY EAR AUDITORY NERVE

SOUND

CONDITIONING DISRUPTED

AUDITORY LESION MIDBRAIN

CONDITIONING DISRUPTED

AUDITORY THALAMUS

NO EFFECT

UNKNOWN LOCATION

ORPOLE AUDITORY CORTEX IAN W

BRAIN LESIONS have been crucial to pinpointing the sites in- necessary for fear conditioning (center). Only damage to the volved in experiencing and learning about fear. When a sound is cortex does not disrupt the fear response, which suggests that processed by the rat brain, it follows a pathway from ear to mid- some other areas of the brain receive the output of the thalamus brain to thalamus to cortex (left). Lesions can be made in vari- and are involved in establishing memories about experiences ous sites in the auditory pathway to determine which areas are that stimulate fear (right).

Emotion, Memory and the Brain Mysteries of the Mind 71 Copyright 1997 Scientific American, Inc. declarative memory—interfered only Structure of the Amygdala with response to the chamber, not the tone. (Declarative memory involves ex- he amygdala plays an important role in emotional behavior. Experiments in plicit, consciously accessible informa- Trodents have elucidated the structures of various regions of the amygdala tion, as well as spatial memory.) At about and their role in learning about and remembering fear. The lateral nucleus re- the same time, Michael S. Fanselow and ceives inputs from sensory regions of the brain and transmits these signals to the Jeansok J. Kim of the University of Cal- basal, the accessory basal and the central nuclei. The central nucleus connects to ifornia at Los Angeles discovered that the brain stem, bringing about physiological changes. —J.E.LeD. hippocampal lesions made after fear conditioning had taken place also pre- vented the expression of responses to the surroundings. These findings were consistent with CORTEX the generally accepted view that the hip- pocampus plays an important role in processing complex information, such LATERAL as details about the spatial environment NUCLEUS where activity is taking place. Phillips THALAMUS HIPPOCAMPUS and I also demonstrated that the subicu- ACCESSORY BASAL lum, a region of the hippocampus that BASAL NUCLEUS projects to other areas of the brain, NUCLEUS communicated with the lateral nucleus of the amygdala. This connection sug- gests that contextual information may CENTRAL NUCLEUS acquire emotional significance in the same way that other events do—via STIMULUS AMYGDALA transmission to the lateral nucleus. Although our experiments had iden- tified a subcortical sensory pathway that sound, for example, reach the central the basal or basolateral nucleus, which, gave rise to fear conditioning, we did not nucleus directly from the auditory thal- in turn, projects to the central nucleus. dismiss the importance of the cortex. amus? We found that it does not. The Collaborating with Lisa Stefanacci The interaction of subcortical and cor- central nucleus receives projections from and other members of the Salk team, tical mechanisms in emotion remains a thalamic areas next to, but not in, the Claudia R. Farb and C. Genevieve Go hotly debated topic. Some researchers auditory part of the thalamus. Indeed, in my laboratory at New York Univer- believe cognition is a vital precursor to an entirely different area of the amyg- sity found the same connections in the emotional experience; others think that dala, the lateral nucleus, receives inputs rat. We then showed that these connec- cognition—which is presumably a corti- from the auditory thalamus. Lesions of tions form synaptic contacts—in other cal function—is necessary to initiate the lateral nucleus prevented fear con- words, they communicate directly, neu- emotion or that emotional processing is ditioning. Because this site gets informa- ron to neuron. Such contacts indicate a type of cognitive processing. Still oth- tion directly from the sensory system, that information reaching the lateral ers question whether cognition is neces- we have come to think of it as the senso- nucleus can influence the central nucleus sary for emotional processing. ry interface of the amygdala in fear con- via the basolateral nucleus. The lateral It became apparent to us that the au- ditioning. In contrast, the central nucle- nucleus can also influence the central ditory cortex is involved in, though not us appears to be the interface with the nucleus by way of the accessory basal crucial to, establishing the fear response, systems that control responses. or basomedial nucleus. Clearly, ample at least when simple auditory stimuli are opportunities exist for the lateral nucle- applied. Norman M. Weinberger and Mapping the Mechanism us to communicate with the central nu- his colleagues at the University of Cali- cleus once a stimulus has been received. fornia at Irvine have performed elegant hese findings seemed to place us on The emotional significance of such a studies showing that neurons in the au- Tthe threshold of being able to map stimulus is determined not only by the ditory cortex undergo specific physiolog- the entire stimulus response pathway. sound itself but by the environment in ical changes in their reaction to sounds But we still did not know how informa- which it occurs. Rats must therefore as a result of conditioning. This finding tion received by the lateral nucleus ar- learn not only that a sound or visual cue indicates that the cortex is establishing rived at the central nucleus. Earlier stud- is dangerous, but under what conditions its own record of the event. ies had suggested that the lateral nucleus it is so. Russell G. Phillips and I exam- Experiments by Lizabeth M. Roman- projects directly to the central nucleus, ined the response of rats to the cham- ski in my laboratory have determined but the connections were fairly sparse. ber, or context, in which they had been that in the absence of the auditory cor- Working with monkeys, David Amaral conditioned. We found that lesions of tex, rats can learn to respond fearfully and Asla Pitkanen of the Salk Institute the amygdala interfered with the ani- to a single tone. If, however, projections for Biological Studies in San Diego dem- mals’ response to both the tone and the from the thalamus to the amygdala are onstrated that the lateral nucleus ex- chamber. But lesions of the hippocam- removed, projections from the thalamus tends directly to an adjacent site, called pus—a region of the brain involved in to the cortex and then to the amygdala

72 Mysteries of the Mind Emotion, Memory and the Brain Copyright 1997 Scientific American, Inc. are sufficient. Romanski went on to es- and Romanski found that essentially ev- tablish that the lateral nucleus can re- ery cell responding to the auditory stim- PRESYNAPTIC ceive input from both the thalamus and uli also responded to the shock. The ba- NEURON GLUTAMATE the cortex. Her work in the rat comple- sic ingredient of conditioning is thus ments earlier research in primates. present in the lateral nucleus. Once we had a clear understanding Bordi was able to divide the acousti- of the mechanism through which fear cally stimulated cells into two classes: conditioning is learned, we attempted to habituating and consistently responsive. find out how emotional memories are Habituating cells eventually stopped re- established and stored on a molecular sponding to the repeated sound, suggest- level. Farb and I showed that the excita- ing that they might serve to detect any NMDA POSTSYNAPTIC RECEPTOR tory amino acid transmitter glutamate sound that was unusual or different. NEURON ORPOLE

is present in the thalamic cells that reach They could permit the amygdala to ig- IAN W the lateral nucleus. Together with Chiye nore a stimulus once it became familiar. J. Aoki, we showed that it is also pres- Sound and shock pairing at these cells ) ent at synapses in the lateral nucleus. might reduce , thereby al- ORPOLE Because glutamate transmission is impli- lowing the cells to respond to, rather IAN W cated in memory formation, we seemed than ignore, significant stimuli. BEFORE LTP

to be on the right track. The consistently responsive cells had MILLIVOLTS high-intensity thresholds: only loud Long-Term Potentiation sounds could activate them. That finding is interesting because of the role loud- lutamate has been observed in a ness plays in judging distance. Nearby process called long-term potentia- sources of sound are presumably more G AFTER LTP

tion, or LTP, that has emerged as a mod- dangerous than those that are far away. NEURAL IMPULSE ( el for the creation of memories. This pro- Sound coupled with shock might act on cess, which is most frequently studied these cells to lower their threshold, in- 0 10 20 in the hippocampus, involves a change creasing the cells’ sensitivity to the same TIME (MILLISECONDS) in the efficiency of synaptic transmis- stimulus. Consistently responsive cells sion along a neural pathway—in other were also broadly tuned. The joining of words, signals travel more readily along a sound and a shock could make the this pathway once LTP has taken place. cells responsive to a narrower range of The mechanism seems to involve gluta- frequencies, or it could shift the tuning mate transmission and a class of post- toward the frequency of the stimulus. synaptic excitatory amino acid recep- In fact, Weinberger has recently shown tors known as NMDA receptors. that cells in the auditory system do alter Various studies have found LTP in their tuning to approximate the condi- the fear-conditioning pathway. Marie- tioned stimulus. Bordi and I have de- Christine Clugnet and I noted that LTP tected this effect in lateral nucleus cells could be induced in the thalamo-amyg- as well. dala pathway. Thomas H. Brown and The apparent permanence of these Paul Chapman and their colleagues at memories raises an important clinical Yale discovered LTP in a cortical pro- question: Can emotional learning be jection to the amygdala. Other research- eliminated, and, if not, how can it be ers, including Davis and Fanselow, have toned down? As noted earlier, it is actu- been able to block fear conditioning by ally quite difficult to get rid of emotion- blocking NMDA receptors in the amyg- al memories, and at best we can hope dala. And Michael T. Rogan in my lab- oratory found that the processing of sounds by the thalamo-amygdala path- MEMORY FORMATION has been way is amplified after LTP has been in- linked to the establishment of long-term duced. The fact that LTP can be dem- potentiation, or LTP. In this model of onstrated in a conditioning pathway of- memory the neurotransmitter glutamate fers new hope for understanding how and its receptors, called NMDA receptors LTP might relate to emotional memory. (top), bring about strengthened neural In addition, recent studies by Fabio transmission. Once LTP is established, the same neural signals produce larger re- DOUX

Bordi, also in my laboratory, have sug- E

sponses (top, middle). Emotional mem- . L

gested hypotheses about what could be H E ories may also involve LTP in the amyg- going on in the neurons of the lateral nu- dala. Glutamate (red circle in top photo- cleus during learning. Bordi monitored graph) and NMDA receptors (red circle HS BY JOSEP the electrical state of individual neurons in bottom photograph) have been found AP OGR in this area when a rat was listening to in the region of the amygdala where fear T HO the sound and receiving the shock. He conditioning takes place. P

Emotion, Memory and the Brain Mysteries of the Mind 73 Copyright 1997 Scientific American, Inc. only to keep them under wraps. Studies between this region of the cortex and sory cortex and hippocampus—delivers by Maria A. Morgan in my laboratory the amygdala may make it more difficult unique information to the organ. Path- have begun to illuminate how the brain for some people to change their emotion- ways originating in the sensory thala- regulates emotional expressions. Mor- al behavior. Davis and his colleagues mus provide only a crude perception of gan has shown that when part of the have found that blocking NMDA recep- the external world, but because they in- is damaged, emotional tors in the amygdala interferes with ex- volve only one neural link, they are quite memory is very hard to extinguish. This tinction. Those results hint that extinc- fast. In contrast, pathways from the cor- discovery indicates that the prefrontal tion is an active learning process. At the tex offer detailed and accurate represen- areas—possibly by way of the amyg- same time, such learning could be situ- tations, allowing us to recognize an ob- dala—normally control expression of ated in connections between the pre- ject by sight or sound. But these path- emotional memory and prevent emotion- frontal cortex and the amygdala. More ways, which run from the thalamus to al responses once they are no longer use- experiments should disclose the answer. the sensory cortex to the amygdala, in- ful. A similar conclusion was proposed Placing a basic emotional memory volve several neural links. And each link by Edmund T. Rolls and his colleagues process in the amygdalic pathway yields in the chain adds time. at the University of Oxford during stud- obvious benefits. The amygdala is a ies of primates. The researchers studied critical site of learning because of its the electrical activity of neurons in the central location between input and out- frontal cortex of the animals. put stations. Each route that leads to Functional variation in the pathway the amygdala—sensory thalamus, sen-

VISUAL CORTEX VISUAL THALAMUS

AMYGDALA

HEART RATE O OSTI T OBER BLOOD PRESSURE MUSCLE R

CORTICAL AND SUBCORTICAL PATHWAYS in the brain— tex also receives information from the thalamus and, with more generalized from our knowledge of the auditory system—may perceptual sophistication and more time, determines that there bring about a fearful response to a on a hiker’s path. Visu- is a snake on the path (blue). This information is relayed to the al stimuli are first processed by the thalamus, which passes amygdala, causing heart rate and blood pressure to increase and rough, almost archetypal, information directly to the amygdala muscles to contract. If, however, the cortex had determined that (red). This quick transmission allows the brain to start to re- the object was not a snake, the message to the amygdala would spond to the possible danger (green). Meanwhile the visual cor- quell the fear response.

74 Mysteries of the Mind Emotion, Memory and the Brain Copyright 1997 Scientific American, Inc. Emotional and Conserving time may be the reason is not declarative learn- And for this reason, the there are two routes—one cortical and ing. Rather it is mediat- declarative trauma may men- one subcortical—for emotional learning. ed by a different system, memories are tal and behavioral func- Animals, and humans, need a quick-and- which in all likelihood tions in later life, albeit dirty reaction mechanism. The thalamus operates independently joined seamlessly through processes that activates the amygdala at about the same of our conscious aware- in our conscious remain inaccessible to time as it activates the cortex. The ar- ness. Emotional infor- . rangement may enable emotional re- mation may be stored experience. That Because pairing a tone sponses to begin in the amygdala before within declarative mem- does not mean and a shock can bring we completely recognize what it is we ory, but it is kept there we have direct about conditioned re- are reacting to or what we are . as a cold declarative sponses in animals The thalamic pathway may be partic- fact. For example, if a conscious access throughout the phyla, it ularly useful in situations requiring a person is injured in an to our emotional is clear that fear condi- rapid response. Failing to respond to automobile accident in tioning cannot be de- danger is more costly than responding which the horn gets memory. pendent on conscious- inappropriately to a benign stimulus. stuck in the on position, ness. Fruit flies and For instance, the sound of rustling leaves he or she may later have snails, for example, are is enough to alert us when we are walk- a reaction when hearing the blare of car not creatures known for their conscious ing in the woods without our having horns. The person may remember the mental processes. My way of interpret- first to identify what is causing the details of the accident, such as where ing this phenomenon is to consider fear sound. Similarly, the sight of a slender and when it occurred, who else was in- a subjective state of awareness brought curved shape lying flat on the path ahead volved and how awful it was. These are about when brain systems react to dan- of us is sufficient to elicit defensive fear declarative memories that are depen- ger. Only if the organism possesses a responses. We do not need to go through dent on the hippocampus. The individ- sufficiently advanced neural mechanism a detailed analysis of whether or not ual may also become tense, anxious and does conscious fear accompany bodily what we are seeing is a snake. Nor do depressed, as the emotional memory is response. This is not to say that only we need to think about the fact that reactivated through the amygdalic sys- humans experience fear but, rather, that are reptiles and that their skins tem. The declarative system has stored consciousness is a prerequisite to sub- can be used to make belts and boots. All the emotional content of the experi- jective emotional states. these details are irrelevant and, in fact, ence, but it has done so as a fact. Thus, or are con- detrimental to an efficient, speedy and Emotional and declarative memories scious products of unconscious process- potentially lifesaving reaction. The brain are stored and retrieved in parallel, and es. It is crucial to remember that the simply needs to be able to store primi- their activities are joined seamlessly in subjective experiences we call feelings tive cues and detect them. Later, coordi- our conscious experience. That does not are not the primary business of the sys- nation of this basic information with the mean that we have direct conscious ac- tem that generates them. Emotional ex- cortex permits verification (yes, this is a cess to our emotional memory; it means periences are the result of triggering snake) or brings the response (scream- instead that we have access to the con- systems of behavioral that ing, sprinting) to a stop. sequences—such as the way we behave, have been preserved by . Sub- the way our bodies feel. These conse- jective experience of any variety is chal- Storing Emotional Memory quences combine with current declara- lenging turf for scientists. We have, tive memory to form a new declarative however, gone a long way toward un- lthough the amygdala stores primi- memory. Emotion is not just uncon- derstanding the neural system that un- A tive information, we should not scious memory: it exerts a powerful in- derlies fear responses, and this same consider it the only learning center. The fluence on declarative memory and oth- system may in fact give rise to subjec- establishment of memories is a function er thought processes. As James L. Mc- tive feelings of fear. If so, studies of the of the entire network, not just of one Gaugh and his colleagues at the neural control of emotional responses component. The amygdala is certainly University of California at Irvine have may hold the key to understanding sub- crucial, but we must not lose sight of convincingly shown, the amygdala plays jective emotion as well. SA the fact that its functions exist only by an essential part in modulating the stor- virtue of the system to which it belongs. age and strength of memories. Memory is generally thought to be The distinction between declarative Further Reading the process by which we bring back to memory and emotional memory is an mind some earlier conscious experience. important one. W. J. Jacobs of the Uni- The Amygdala: Neurobiological As- The original learning and the remem- versity of British Columbia and Lynn pects of Emotion, Memory and Men- tal Dysfunction. Edited by John P. Ag- bering, in this case, are both conscious Nadel of the University of Arizona have gleton. Wiley-Liss, 1992. events. Workers have determined that argued that we are unable to remember Brain Mechanisms of Emotion and declarative memory is mediated by the traumatic events that take place early in Emotional Learning. J. E. LeDoux in hippocampus and the cortex. But re- life because the hippocampus has not Current Opinion in Neurobiology. Vol. 2, moval of the hippocampus has little ef- yet matured to the point of forming No. 2, pages 191–197; April 1992. fect on fear conditioning—except con- consciously accessible memories. The The Role of the Amygdala in Fear and ditioning to context. emotional memory system, which may Anxiety. M. Davis in Annual Review of Neuroscience, Vol. 15, pages 353–375; In contrast, emotional learning that develop earlier, clearly forms and stores 1992. comes about through fear conditioning its unconscious memories of these events.

Emotion, Memory and the Brain Mysteries of the Mind 75 Copyright 1997 Scientific American, Inc.