In this study, we used fMRI to investigate sex differences in BOLD-signal changes of healthy subjects exposed to a fear conditioning and fear extinction paradigm. No statistically significant sex differences in SCRs were observed during any experimental phase. However, we note that men exhibited a trend towards generally elevated SCRs during acquisition of conditioned fear responses and the extinction recall test. Regarding the BOLD responses, we observed significant sex differences in several brain regions during fear conditioning and extinction recall. Specifically, during fear conditioning, women showed greater activation in dACC, rACC and amygdala relative to men. During extinction recall, women showed greater activation in the insula cortex relative to men, while men showed greater activation in the rACC region of the mPFC relative to women. Moreover, we found that regions where sex differences were previously identified in response to stress  also exhibited sex differences during fear conditioning and extinction, including the anterior hypothalamus.
Sex differences at the behavioral level have been reported in humans and rodents across a number of paradigms such as fear conditioning, active avoidance, conditioned taste aversion and eye blink conditioning . Some studies report increased conditioned responding in males rats during fear learning , which is consistent with our findings of sex differences in SCR, whereas others found no significant sex differences in fear acquisition . We previously reported that men show elevated conditioned fear responses relative to women [17, 23]. While not statistically significant in this study, men exhibited a trend towards elevated conditioned fear responses. The lack of a statistically significant different sex difference may be due to variability of endocrine status or use of contraceptives among the women in our sample. That is, we recently demonstrated that estradiol significantly enhances fear extinction recall and its neural correlates during fear extinction , and others have reported that contraceptives can impact learning and memory . Moreover, it has been shown that estradiol facilitated contextual fear extinction via estradiol’s effect on hippocampal long-term potentiation in rats . One important caveat is that in our data, men also showed a trend towards increased skin conductance responses to the CS- compared with the women. This would suggest enhanced general skin conductance reactivity in men relative to women that may not be specific to fear learning or extinction per se given that they are higher across conditions. Future studies will need to further investigate the role of sex-steroid hormones in understanding variability due to gender during conditioning and the neural responses of the fear extinction network.
Using an anatomical ROI analysis approach, we observed sex differences in fear circuitry activation. Women exhibited greater amygdala, mPFC and dACC activation during fear acquisition and greater insula activation during extinction recall. In contrast, men showed significantly greater activation within the rACC, which is in close proximity to the locus we previously reported showing increased activation during fear extinction [35, 46, 47]. The mPFC and dACC have been implicated in pain processing, conflict monitoring and error processing, fear expression, and appraisal of emotionally salient stimuli [41, 48–50]. The amygdala is also well known for signaling novelty and mediating emotional learning such as fear conditioning . The vmPFC has been implicated in emotion regulation and fear extinction recall [2, 52]. The increased mPFC activation in men during extinction recall predicted facilitated fear extinction recall in men and fear responses in women, and the increased amygdala and dACC activation in women during fear conditioning again predicted facilitated fear responses in women during this phase. The behavioral data showed lack of sex differences in fear responses. Thus brain activity differences in neural responses may be contributing to producing similar behavioral responses suggesting that men and women use different neural strategies to produce homeostasis in the brain in response to fear. This was similar to Goldstein’s previous findings of sex differences in neural responses to stress to maintain homeostasis in the brain in response to stress, which was dependent on menstrual cycle phase in the women . Further studies are needed to explore whether men and women use different neural networks to acquire and control fear to a similar degree.
In fact, the functional ROI analyses revealed overlap between sex differences in stress-arousal circuitry activation and sex differences in fear-arousal circuitry reported here. For example, Goldstein and colleagues  reported that men, compared with women in the late follicular menstrual cycle phase, exhibited significantly greater BOLD-signal changes in response to negative versus neutral stimuli in ACC, OFC, mPFC, anterior hypothalamus, hippocampus and periaqueductal gray. Although the regions of activation overlap, data from the present study indicated that women exhibited significantly greater BOLD-signal changes, compared to men, during fear conditioning in a number of these brain regions. However, no women in the current study were scanned during the mid-cycle menstrual phase, as distinct from all women in the Goldstein study who contributed to the sex difference effect were scanned during this phase. Men in both studies showed hyperactivation in vmPFC, other orbitofrontal regions, and right hypothalamus. In our study, hyperactivation was observed during extinction recall even without controlling for menstrual cycle phase. The differences between our findings and those reported by Goldstein and colleagues may be due to differences in levels of sex hormones, particularly estradiol and/or progesterone , but this hypothesis needs further investigation. Nevertheless, the approach underscores the importance of analyzing sex differences in shared brain circuitry across behavioral domains for understanding psychiatric disorders.