EMOTIONAL MEMORY AND THE AMYGDALA LECTURE

[figure of brain]

-- The amygdala is located in the anterior part of the temporal lobe, next to the hippocampus.

It has dense interconnections with the rest of the brain.

-- early interest in amygdala  came from kluver and bucyís findings with monkeys with temporal lobe lesions, who developed profound changes in emotional and social behavior, known as the Kluver-Bucy syndrome

[figure of Kluver-Bucy]

1. Taming effect: show no fear of previously feared stimuli such as snakes or humans.
2. ìPsychic blindnessî, or inability to appreciate the psychological significance of visually presented objects
3. Hyperorality
4. Indiscriminate hypersexuality

Since these early findings  there have been many additional studies linking the amygdala to emotion.

For example, electrical stimulation of the amygdala in humans has been shown to produce a variety of emotional reactions, usually negative responses, most frequently fear.

One last note before I go on to facial emotion. I should note that  ìthe amygdalaî, is in fact a complex of about 15 smaller cell groups or nuclei

[figure here of the nuclei]

each of which has its own specific connections, primary neurotransmitters, and functions.

In the work Iíll be describing though, Iíll be ignoring this complexity because the data rarely allow distinctions to be made at this level.
 

--  an emotional stimulus is detected and reacted to, itís adaptive to have the relevant aspects of this event encoded into memory for future reference

-- the amygdala appears to play 2 major roles in this regard

 -- first, it is involved in learning the association between a previously neutral stimulus and an emotional response.

 -- for example, in classical fear conditioning, the pairing of a tone with a shock will eventually cause the tone to elicit the emotional response originally evoked by the shock.

-- the second role of the amygdala is in modulating declarative or consciously accessible memory such as recall and recognition

-- the neural basis of the first type of memory has been well worked out in both animal models and more recently, in humans, so I will focus on the second, modulatory role, which is much less well understood

------
-- a priori, it seems reasonable that the amygdala, which is involved in detecting and reacting to emotional stimuli, should play a role in strengthening memories for emotional stimuli

-- weíre all probably familiar with the extreme example of emotion influencing memory, the so-called flashbulb memory phenomenon, but emotion has a major effect on memory for less arousing stimuli as well

-- and, although most of the focus has been on unpleasant or negative emotion and memory, positive emotion also enhances memory. the degree of arousal is the primary factor.

-- now, if the amygdala really plays a role in emotional memory, one would expect emotional memory deficits to follow amygdala damage.
 

-------

-- Again, here the data are sparse due to the rarity of selective bilateral lesions to the amygdala, but one can still make out a general pattern

-- All of the data in the literature are with one paradigm, an emotional story task.

-- in this task, an emotional middle section is sandwiched between two nonemotional sections. memory increases for the middle section

-- subjects see a slide show and hear an accompanying story. the story describes how a boy and his mother go to visit the father at work
[figure of boy and mother]
how the boy is hit by a car and rushed to the hospital, where his legs are reattached
-- memory peaks for this slide
[2 figures, one of accident, the other of his legs re-attached]
finally, the mother returns home
[slide of mother going home]

-- the initial link to the amygdala came through a beta-blocker study Larry Cahill conducted with humans
-- subjects who received a beta-blocker drug before the task did not show memory enhancement for the emotional material, although they rated it as emotionally arousing. nonemotional memory was not affected.
-- based on animal data, the site of action of this effect was linked to the amygdala

[figure of site of action of drugs in amygdala]

-- the logical next step was to investigate whether patients with bilateral damage to the amygdala would also lack this enhancement

-- here are the data from 2 Urbach-Wiethe patients.  The emotional ìlegsî slide is shown with the arrow, the emotional middle phase begins with slide 5.

-- neither patient shows normal memory enhancement for the legs slide or for the middle phase. Although BP is not normal for the last phase, SM is.

_____

-- before we can conclude that this deficit indicates the loss of a normal emotional memory mechanism, we have to consider the possibility that the story simply wasnít emotional for the patients.

-- As I indicated before, amygdala damage does not appear to disrupt normal emotional reactions to stimuli, and the patients rated the story as being highly arousing

-- but perhaps they rated it highly because of their semantic knowledge of what should be arousing, although they were not physiologically aroused

-- We followed this up with patient DBB and recorded physiological responses during the story.

[figure of DBB study]

-- on the top are the memory data. Like the other patients, DBB did not show enhanced memory for either the peak ìlegsî slide or the middle phase.

-- However, notice that for the legs slide DBB has a large physiological response. He also rated the story as 9 out of 10 in arousal and made comments afterward about how painful it was to see a child get hurt  -- by the way, this response (note first upturn) is an artifact caused by a loud noise.

[back to previous slide]

-- So, amygdala damage does not affect the reactions to this story, consistent with previous findings.

-- But somehow amgydala damage interferes with the translation of emotional reactions into stronger memory.

----------

-- well, these data are suggestive, but they are with only one paradigm.

-- weíve followed up these studies with patient DBB in a variety of emotional memory tasks designed to get at two basic issues.

-- first, is this deficit general to both positive and negative stimuli?

-- second, is memory for verbal as well as nonverbal memory affected?

[figure of DBB data]

-- here are data from an experiment that used the positive, negative, and neutral stimuli from the PET study I described earlier that had been shown to activate the amygdala.

-- notice that whereas control subjects show enhancement for both the positive and negative stimuli, DBB fails to show enhancement

-- similar results were found with a study that used socially unacceptable or taboo words. Controls showed enhancement for the arousing taboo words, but DBB did not show this enhancement

[back to previous slide]

-- so these data suggest that the emotional memory deficit is general to both positive and negative emotion as well as verbal and nonverbal stimuli,
that is, a general lack of enhancement of declarative memory by emotion.

----------

-- now, how does the amygdala actually affect declarative memory?

-- it could act like the hippocampus, directly participating itself in the processing of emotional memory

-- or, it could modulate other memory-related structures such as the hippocampus, but not participate directly in the formation or retrieval of memory

-- actually, it does both.

-- for fear conditioning and other conditioning phenomena, it processes and stores the engram within itself

-- but for declarative memory, it appears to have only a modulatory role

-- thereís a lot of animal data supporting this.

-- but also there is some human data.  for example, if the amygdala itself processed emotional memory, we would expect to see a selective sparing of emotional memory in amnesic patients with hippocampal damage.  For example, fear conditioning is normal in such patients because the amygdala processes that type of memory rather than the damaged hippocampus.

-- but here with the emotional story task again, amnesic patients with hippocampal damage but normal amygdalas showed a normal enhancement of memory by emotion.

-- this suggests that the intact amygdala is modulating whatever hippocampal function is left in these patients in a normal way.

______

-- so, these neuropsychological results are remarkably consistent
 
 

PET study of emotional memory encoding

 Predictions based on prior animal and neuropsychological findings:

1  The amygdala should be activated by emotionally     arousing stimuli, regardless of whether they are     pleasant or aversive

2 Amygdala activity should be correlated with enhanced    conscious memory (recall, recognition) for emotional    stimuli (both pleasant and aversive) but not for     highly memorable but nonemotional stimuli

3 Hippocampal activity should be correlated with     enhanced conscious memory for emotional stimuli,    and should also be present for highly memorable,     nonemotional stimuli

Subtraction analysis -- conclusions

ï Both pleasant and aversive stimuli elicited significant    amygdala activation, as predicted

ï  First neuroimaging evidence of amygdala activation    elicited by pleasant stimuli

ï Laterality of activation is consistent with previous     studies that have found left-sided amygdala activation    in tasks that elicited aversive emotion

ï However, unexpectedly, the interesting condition was   also associated with amygdala activation, to a lesser    extent, and subtraction using the interesting condition    as a control eliminated the significant amygdala     activations for pleasant and unpleasant stimuli

ï  Suggests that the amygdala is active during      processing of some kinds of nonemotional stimuli,     specifically those that attract high interest and attention

ï Strong visual area activations in the pleasant and     aversive conditions are consistent with the idea that    the amgydala can enhance processing in visual areas    via back-projections

Correlational results

ï  Analysis was conducted for all three memory tests, but    only the recognition test yielded correlations in the    regions of interest

ï  A significant correlation was found between individual    subjectsí recognition enhancement and bilateral     amygdala activity during the pleasant picture scans

ï Recognition enhancement was also correlated for the    aversive picture scans, though at a slightly more     inferior level and more right-lateralized

ï This relation was not found for interesting pictures,    indicating that the amygdalaís relation to enhanced    recognition was specific to emotional stimuli

ï Bilateral hippocampal/parahippocampal activity was    correlated with enhanced recognition for pleasant and    aversive pictures. Right hippocampal activity was     correlated with enhanced memory for interesting     pictures
 
ï For pleasant pictures, the nucleus accumbens, an area  that has been linked to pleasure and reward      mechanisms, also showed memory-correlated activity
 

PET study conclusions
 

1 Results supported each of the three predictions

2 Results confirm previous Cahill FDG-PET study of    aversive emotional memory and extend the results to    pleasant memory

3 Findings are consistent with a modulatory view of the    amygdalaís role in influencing emotional memory

4 Overall, the results suggest that the amygdala in     humans plays an important adaptive role in enhancing    conscious memory for events according to their     emotional importance (arousal), regardless of whether    the nature of the emotion is pleasant or aversive

Summary

The findings from animal, neuropsychological, and neuroimaging studies all converge on the following conclusions:
 

ï The amygdala plays a crucial, adaptive role in     enhancing  memory for emotionally arousing stimuli

ï  Memory for non-emotional, neutral stimuli does not    engage the amygdala

ï The amygdalaís role in enhancing memory includes    both aversive and pleasant stimuli

ï Normal cognitive and physiological emotional     reactions can take place without the amygdala

ï For conscious memory, the amygdala appears to     modulate the operation of other, memory-related brain    systems such as the hippocampal system rather than    participating directly in memory encoding and storage