DISCUSSION
The present results uniquely extend recent observations in
rodents that stress-induced interactions between AEA and
CRH signaling modulate amygdala function associated with
anxiety and fear extinction (9). In rodents, stress-induced CRH
signaling via CRHR1 in the basolateral amygdala results in
increased activity of FAAH. The increased activity of this
catabolic enzyme subsequently results in decreased AEA
and a loss of inhibitory tone necessary for reducing anxiety
and maintaining fear extinction. Here, we demonstrate parallel
effects in humans using two functional genetic polymorphisms
to model variability in AEA inhibitory tone and CRH signaling.
Specifically, we found the least temporal habituation of the
basolateral amygdala, a neuroimaging correlate of fear extinc-
tion, in individuals who have relatively high AEA inhibitory tone
(i.e., FAAH 385A allele carriers) and relatively high CRH
signaling (i.e., CRHR1 A allele homozygotes). Moreover, the
blunted habituation of the left amygdala in these individuals
mediated a significantly increased risk for anxiety disorders.
Although speculative, the laterality of this mediated risk may
reflect the preferential contributions of the left amygdala to
sustained evaluation of threat (42), which is a distinguishing
feature of anxiety disorders (43).
Our study is not without limitations. Modeling variability in
signaling pathways using functional genetic polymorphisms
does not provide direct evidence for these interactions in
humans; establishing functional correlates of our target poly-
morphisms through assays of circulating cortisol or AEA
concentrations would underscore the accuracy of our model
(44). The amygdala habituation phenotype we examined may
be associated with fear extinction; for instance, during early
extinction trials the amygdala is activated by stimuli previously
conditioned to aversive outcomes, while in late extinction trials
it is not (20,21). However, our task did not condition individuals
to specific stimuli. Instead, it relied upon stimuli that have
presumably been conditioned in everyday experience (e.g.,
facial expressions of fear). As such, amygdala habituation as
presently measured is not a direct analog of traditional fear
extinction (e.g., to a stimulus previously conditioned within the
laboratory) [e.g., (25)]. Further, we did not collect behavioral
measures of fear extinction (e.g., skin conductance response)
(1,13), which may, speculatively mediate associations between
amygdala habituation and anxiety disorder risk. The concur-
rent examination of neural and behavioral phenotypes in the
context of imaging genetics research may provide clues to
these relationships in future research.
The cross-sectional nature of our study is a further limi-
tation regarding the contribution of the observed pathways to
risk for anxiety disorders. Our moderated mediational model
makes specific directional predictions regarding the link
between amygdala habituation and psychopathology. While
there is a robust but primarily nonhuman animal literature that
is consistent with these directional assumptions, it is possible
that alterations in habituation follow rather than precede the
development of an anxiety disorder. Notably, however, blunted