The impact of sex and menstrual phase on emotional reactivity and emotion regulation

Epidemiological research has reliably determined that women are significantly more likely to develop anxiety disorders than men, with women typically developing these disorders at twice the rate of men. While sex differences in the development and prevalence of such psychiatric conditions are established, the mechanisms underlying these differences are currently unknown. One proposed mechanism is that women exhibit greater emotional reactivity to negative emotions than men, leading to enhanced sensitivity for processing unpleasant or threatening stimuli. This pattern of responding reflects a negativity bias. An alternative explanation of the vulnerability of women in developing anxiety disorders is that women, when compared to men, have greater difficulty regulating their response to unpleasant stimuli and subsequent negative emotions. Behavioural, physiological, and neuroimaging data indicate that women are overall more responsive to emotional stimuli, particularly unpleasant stimuli, relative to men. However, previously reported sex-related differences in electrophysiological cortical activity during emotion processing, and particularly in emotion regulation, have been understudied and existing data is varied and inconsistent. Two key theories have been developed to explain the processing of emotional information; the motivational model and the negativity bias hypothesis. The motivational model asserts that emotional (pleasant and unpleasant) stimuli require greater processing relative to neutral stimuli. In contrast, the negativity bias hypothesis proposes prioritised processing of unpleasant compared with pleasant and neutral stimuli. The current thesis was designed to investigate the competing theoretical perspectives and possible mechanisms which may explain sex differences in psychopathologies such as anxiety disorders. This research project used high temporal resolution event-related potentials (ERPs) to investigate sex differences in the cortical processing of emotion. In Study 1, ERPs were recorded from healthy women and men during a dual oddball task containing pleasant, unpleasant, and neutral stimuli to test the competing motivational and negativity bias models. N2 amplitude for women was significantly greater, reflecting more emotional orienting processes, to neutral and unpleasant relative to pleasant stimuli while N2 activation for men was increased to neutral compared to both pleasant and unpleasant stimuli, with unpleasant stimuli eliciting higher N2 amplitude than pleasant stimuli. Irrespective of sex, P3 activation was greater, indexing increased conscious appraisal and subsequent allocation of attention, to pleasant and unpleasant relative to neutral stimuli. During the dual-task condition, both women and men exhibited increased LPP amplitudes, signifying an enhanced regulatory response, to pleasant and unpleasant compared to neutral stimuli, with women displaying significantly greater LPP amplitude than men to all valences. While women rated the unpleasant stimuli as more arousing than men, no ERP evidence was found for the negativity bias. Some support, during late (P3, LPP) processing, was shown for the motivational model, however, no sex differences to emotional stimuli were demonstrated. Taken together, there was little evidence for a female negativity bias and while some support for the motivational model was demonstrated, few sex differences in emotional reactivity were shown. The findings of Study 1 were somewhat divergent from previous literature and may be explained by methodological factors, a key one of which was failure to control for menstrual phase. As recent neuroimaging data has emerged suggesting significant differences in cortical emotional processing according to menstrual phase, Study 2 aimed to extend Study 1 by examining the impact of menstrual phase on emotional processing in women compared to men. Accordingly, ERPs were recorded from healthy women in their early follicular menstrual phase (day 1-7 (low estradiol/low progesterone)), healthy women in their midluteal menstrual phase (day 18-24 (high estradiol/high progesterone)), and from healthy men while they viewed neutral, and low- and high- arousing pleasant and unpleasant stimuli in a passive viewing task. Modulation by menstrual phase was demonstrated during early visual processing, as midluteal women exhibited significantly larger P1 amplitude at the occipital region to all visual images, relative to men, suggesting that mifluteal women have superior generally enhanced visual processing. Early follicular and midluteal women both showed greater N1 amplitudes, reflecting increased automatic preconscious processing, compared to men (although this only reached significance for the midluteal women) to the visual stimuli. No menstrual phase or sex differences were revealed during later (N2, P3, LPP) processing. In addition to statistical significance, reporting of effect sizes is important as it promotes a more scientific approach to the accumulation of knowledge by indicating the practical and clinical significance of research findings. As such, Cohen's d (Cohen, 1988) effect sizes were reported in Studies 2 and 3 to aid interpretation of the findings involving menstrual phase effects. Cohen's rule of thumb for interpreting effect sizes is that small effects (d=.2) represent findings of weak practical and clinical significance, medium effects (d=.5) represent findings of moderate significance, whereas large effects (d=.8) reflect findings with strong practical and clinical significance. For Study 2, the finding of enhanced P1 amplitude for midluteal women relative to men reflected a large effect size at both O1 site (d=.9) and O2 (d=.83) site whereas the greater N1 amplitude elicited by midluteal women compared to men represented a moderate to large effect size (d=.74). The results of Study 2 demonstrate that, as compared to men, women have greater early automatic visual processing, with this effect particularly strong in midluteal women at the earliest stage of visual attention processing. However, this was found to all emotional and neutral stimuli, which does not confirm predictions of the motivational model or the negativity bias hypothesis but suggests there is a generalised enhancement of visual processing in women when sex hormones are elevated. Recent theoretical models propose that early enhanced emotional processing or reactivity to stimuli may impair later emotion regulation processes. Study 3 was designed to investigate sex differences in emotional reactivity and emotion regulation controlling for menstrual phase. To this end, ERPs were recorded from healthy early follicular women, midluteal women, and men while they completed an emotion regulation task. Midluteal women reported greater effort and distress when attempting to suppress emotional responses to unpleasant images than did men. Further, larger N2 amplitude, reflecting greater early conscious emotional orienting processes, was demonstrated during suppression for midluteal women compared to early follicular women and men. P1 and N1 amplitudes were shown to be greater in midluteal women compared to men regardless of instructional set, indicating enhanced early unconscious attentional processing. No menstrual phase or sex differences were demonstrated during late (P3, LPP) processing. Evidence from Study 3 suggest that midluteal women have difficulty down-regulating their behavioural and mid-latency (but not later) cortical responses to unpleasant stimuli during suppression, which suggests early reactivity in midluteal women may be related to difficulties with suppressing emotional responses. For Study 3, the finding of increased distress (d=.67) and effort (d=.64) reported by midluteal women relative to men during suppression reflected moderate to large effect sizes. Similarly, the larger N1 amplitude in midluteal women during suppression (d=.64) and reappraisal (d=.74) represented moderate to large effect sizes, as did the enhanced N2 component during suppression in midluteal women (d=.69). During reappraisal, the increased P1 amplitude represented a moderate to large effect size at O2 site but a large effect size at O1 site. When considered together, the evidence from the three studies in this thesis does not provide definitive support for either the motivational model or the negativity bias hypothesis (see Appendix A). In Study 1, no evidence of the negativity bias was revealed, however, some evidence for women having greater late processing in line with the motivational model was found. Given the lack of clarity in the obtained results, we examined menstrual phase as a potential powerful and often uncontrolled influence on emotional processing in previous studies. In contrast to Study 1, when controlling for menstrual phase no support for the motivational model was found in Study 2, but some evidence for the negativity bias during late processing was found across both women and men. However, rather than find evidence of a negativity bias during early reactivity, Study 2 revealed a generalised enhancement of early visual processing for midluteal women (when sex hormone levels are high). Subsequently in Study 3 we tested whether this early visual reactivity impacted on emotion regulation, and found initial evidence of greater visual reactivity alongside reported difficulty with suppression and greater mid-latency cortical processing during suppression in midluteal women. Overall, Study 3 extended existing emotion processing literature by examining aspects of emotion regulation in conjunction with menstrual phase. This thesis presents novel ERP evidence in Studies 2 and 3 for enhanced early visual processing in midluteal women and of deficits in suppression (with enhanced mid-latency cortical processing) in midluteal women. The finding that this suppression effect is particularly pronounced during the midluteal phase suggests that women may...