Trait anxiety and sex have been shown to separately account for some of the observed individual differences in amygdala responses to emotional stimuli, but the combined effect of both factors remains unknown. In this fMRI study, participants varying in trait anxiety scores viewed a series of superimposed face/scene composite images (containing fearful or neutral faces) and were instructed to direct attention to either the face or the scene content. We observed an interaction between sex and trait anxiety in amygdala responses to fearful faces as a function of attention. In females, higher trait anxiety was associated with a stronger amygdala response to unattended fearful faces, whereas no such relationship was present in males. This observed interaction between sex and individual differences in trait anxiety at the level of the brain may have clinical implications for a better understanding of the higher incidence of anxiety disorders in women than men.
Functional neuroimaging studies have consistently reported amygdala responses to various types of threat-related stimuli (Davis and Whalen, 2001), supporting the notion of a general role of this structure in the detection of biologically meaningful events across species (LeDoux, 2000). None the less, a growing number of studies point towards the existence of individual differences in the magnitude, location and/or laterality of amygdala responses to emotional material (Hamann and Canli, 2004). For example, sex differences in amygdala responses to emotional stimuli, particularly facial expressions, have been reported (Hamann, 2005 and Cahill, 2006). In general, women appear to have larger and more sustained bilateral amygdala activation than men (Hall et al., 2004, McClure et al., 2004 and Williams et al., 2005). This is consistent with the observation that women classify emotional expressions, especially those depicting fear, more accurately than men (Hall, 1978, Thayer and Johnsen, 2000 and Wild et al., 2001) and may reflect a greater vigilance in females for threat-related stimuli (Williams et al., 2005).
Individual differences in anxiety have also been shown to influence neural responses to threat stimuli, particularly in situations of reduced awareness or attention. For instance, Etkin et al. (2004) found that trait anxiety scores positively correlated with basolateral amygdala responses to masked fearful faces, and Bishop et al. (2004) reported that right amygdala responses to fearful faces correlated with state anxiety. In the latter study, an expression-by-attention interaction was also observed: participants with higher state anxiety showed a higher response to fearful faces that were unattended compared to attended ones, while the low state anxiety group exhibited the opposite pattern. Thus, these results suggest a relationship between anxiety levels and automatic amygdala responses to threat. Interestingly, enhanced amygdala responses to threat-related stimuli presented below the threshold of awareness have also been reported in patients with anxiety disorders (Rauch et al., 2003).
While previous studies have highlighted the individual effects of anxiety levels and sex on neural function, it remains unknown whether these variables exert their influence independently, in an additive fashion, or whether a more subtle interaction exists between them. Here, we directly tested, using functional magnetic resonance imaging (fMRI), the effects of trait anxiety and its interaction with sex on the neural responses to fearful faces under different attention conditions. We hypothesized that trait anxiety would influence amygdala responses to unattended faces, as shown in previous studies (Bishop et al., 2004, Bishop et al., 2007 and Most et al., 2006) and that, critically, this effect would be stronger in female participants, consistent with the proposed greater vigilance for threat stimuli in women.
