Graduate School Life


Amit Etkin and Kristen Klemenhagen

Developing a
Window into the
Unconscious Mind

BY JOHN KOESTER, PH.D.,
PROFESSOR OF CLINICAL NEUROBIOLOGY AND BEHAVIOR AND
ACTING DIRECTOR, CENTER FOR NEUROBIOLOGY & BEHAVIOR

AMIT ETKIN AND KRISTEN KLEMENHAGEN, STUDENTS IN THE Neurobiology and Behavior Doctoral Program, both chose dissertation projects that use genetically modified mice to study the neural control of behavior. In the lab of Dr. Eric Kandel, Mr. Etkin, an M.D./Ph.D. student, studied the molecular mechanisms that underlie long-term memory. Ms. Klemenhagen, a student in Dr. René Hen's lab, is investigating the role of serotonin receptors in anxiety states.
Both were inspired to broaden the scope of their experimental approaches by taking a graduate course in "Functional Neuroimaging of the Human Brain," taught by Joy Hirsch, Ph.D., professor of functional neuroradiology in radiology. When the course ended, Ms. Klemenhagen, Mr. Etkin, and their colleagues teamed up to do a special brainimaging project to investigate the differences between conscious and unconscious processing of information.
The goal of this "side project," which eventually turned into a study of major significance, was to investigate the neural basis of responses to threatening stimuli in healthy human subjects. Earlier experimental and theoretical studies had suggested that both conscious and unconscious processing of threatening stimuli occur, that the unconscious processing may be related to an individual's baseline level of anxiety, and that these two processes may be carried out in different regions of the nervous system, including parts of the amygdala. In testing these ideas, the authors targeted two regions of the amygdala for particular attention: the basolateral amygdala and the dorsal amygdala. The activity levels of these and other regions of the brain to various stimuli were measured by functional magnetic resonance imaging (fMRI).
A key to interpreting the imaging results was the authors' determination of each subject's characteristic tendency toward anxiety. As with any randomly selected sample of normal individuals, subjects varied in their inherent anxiety characteristics along a scale that can be determined by Spielberger's Trait-Anxiety Inventory. This psychological test provides an indication of how strongly each subject is likely to respond to a potentially anxiety-provoking situation. The anxiety-triggering stimuli used in the brain-imaging part of the study were a set of randomly presented faces exhibiting fearful expressions (see illustration). The difference between the fMRI responses to fearful and neutral faces was used to assess the response of a specific brain region to the threatening aspect of the fearful face. To measure the subconscious response, each fearful face was presented for only 1/8 of the normal stimulus presentation time, followed immediately by a longer presentation of a neutral face. Under these "backward masking" conditions, the fearful face is perceived unconsciously but its fearful nature is not consciously perceived by the subject. Each face was overlaid with an arbitrarily chosen color, and subjects were instructed to respond by identifying the color of each stimulus presentation. The reaction time of this response was used as a measure of the subject's level of emotional vigilance, which is thought to be related to a subject's trait anxiety score.


A: Examples of the fearful (left) and nonfearful (right) faces used as stimuli. Each face was overlaid with an arbitrary color used in the reaction-time test.
B: Plots of fMRI responses vs. trait anxiety score for the individual subjects in response to fearful faces. In the basolateral amygdala, the response to unconsciously perceived faces varied directly with trait anxiety (B1), but no significant response was observed for consciously perceived faces (B2). In the dorsal amygdala, responses to consciously perceived faces were significantly above baseline but unrelated to the subject's anxiety trait score (B4), while the study showed no significant response to unconsciously perceived faces (B3).

The two areas of the amygdala differed in the way they processed the two types of stimuli. The fMRI response of the dorsal amygdala was greater for nonmasked (consciously perceived) fearful faces than for neutral faces and was independent of each individual's anxiety trait score. However it generated no response above control to masked (unconsciously perceived) fearful faces. In contrast, the basolateral amygdala showed no consistent response to the nonmasked fearful faces, but its responses to masked fearful faces did correlate with subjects' trait anxiety scores: The greater the anxiety rating, the greater the response to these unconsciously perceived fearful faces. These results are consistent with the notion that the dorsal amygdala is engaged in analyzing the information content in consciously processed threatening stimuli, while the basolateral amygdala is involved in the unconscious processing of potentially threatening stimuli.
Such a threat-evaluating mechanism, which had been postulated to be linked to an individual's level of anxiety, may have the adaptive effect of increasing awareness of potential danger. In support of this hypothesis, the students and their colleagues found that in response to masked fearful faces, reaction time and response strength of the basolateral amygdala co-varied with the anxiety trait score. Consciously perceived (nonmasked fearful faces) showed no such correlation.
The authors further extended their analysis by looking at fMRI responses at higher levels of the nervous system under the two stimulus presentation conditions. Their results are consistent with previously developed cognitive theories of anxiety according to which a person's innate anxiety level may bias his or her neural processing to one of two potential pathways. For an obviously threatening signal, the neural network that is involved is part of the mechanism for conscious perception. For stimuli that are subthreshold or ambiguous in their threat level, the later steps in sensory processing in an anxious individual may serve to direct increased attentional resources toward the stimulus, thereby increasing the possibility that the individual can respond quickly to a developing threat.
These results are of major interest for two reasons. First, they provide fundamental new insights into the functional architecture of the amygdala, including its role in the differential response of the nervous system to threatening stimuli that are either above or below the threshold for conscious perception. Second, the paradigm developed for this study may provide insights into the mechanisms by which post-traumatic stress and other anxiety syndromes respond to psychotherapy. One goal of therapy for such conditions is to prevent the interpretation of neutral stimuli as being threatening. Using this imaging paradigm, it may be possible to determine whether such therapy works by changing unconscious processing, or, alternatively, by working at the level of the conscious control of thought processes.
Their study was published in the Dec. 16, 2004, issue of the journal Neuron.


| TOP |