| A large body of work, both in experimental animals and more recently in humans, has showed that the amygdala is a critical component of the brain network underpinning emotional processing. However, controversies still exist about specific aspects of its role in emotion, particularly in terms of hemispheric specialization, sex differences and its sensitivity to stimulus valence and/or arousal. The large functional neuroimaging literature developed over the past two decades can help us clarify whether these and other factors influence the magnitude of the amygdala response. To this end, I performed a quantitative meta-analysis of published functional brain imaging studies of emotional perception that have reported an amygdala activation. Critically, this meta-analysis included the magnitude (effect size) and reliability (variance) associated with each of the activations. Our main findings were that the amygdala responds to a larger extent to positive than negative stimuli, and for emotional faces than scenes, pictures, words or films. No evidence for amygdala lateralization as a function of sex or valence was observed. Instead, our findings provide strong support for a functional hemispheric dissociation in terms of temporal dynamics, namely that the right amygdala is more involved in the initial rapid and transient response to the presence of an emotional stimulus, which is followed by a slower and more sustained activation by the left amygdala. We also explored the question of amygdala lateralization in emotional memory, in particular in terms of sex differences. To do so, we re-analyzed data from our previous fMRI study of emotional memory for fearful faces (Sergerie et al. 2006). We observed a sex-specific hemispheric laterality of the amygdala involvement in successful memory for fearful faces only when the sex of the face stimuli was also taken into account. Specifically, the left amygdala was more active in the successfully remembered female fearful faces in women, whereas, for men, the right amygdala was more involved in memory for male fearful faces. These findings provide support and further refine existing models of sex-specific lateralization of the amygdala in emotional memory. I also investigated the emotion-specificity of memory for faces by conducting an fMRI study of recognition memory for sad, happy and neutral faces. The main goal of this study was to assess whether memory for sad faces, a negatively valenced but low arousal emotion, was similar or not to that for fearful faces. Our behavioral results revealed that sad faces were associated with a decrease in recognition accuracy, compared to both happy and neutral expressions, in stark contrast to the previously observed enhanced memory for fearful faces. This effect, however, was due to a significant familiarity bias; that is, subjects were more likely to believe that they had previously seen a sad face, regardless of whether this was true or not. In contrast, happy faces were associated with a novelty response bias. This bias correlated with amygdala and prefrontal cortex activity, whereas the familiarity bias was associated with a superior temporal gyrus activation.;In a follow-up experiment, the same paradigm was applied to individuals with schizophrenia to further investigate their often reported behavioral and neural abnormalities in memory and emotional processing. Despite an overall lower memory performance, patients showed the same influence of emotion on memory as controls, both in terms of accuracy and response bias. For sad faces, this similar behavioral pattern was mirrored by a largely overlapping neural network, mostly involved in familiarity-based judgments (e.g., parahippocampal gyms). In contrast, controls activated a larger network for happy faces, including regions involved in recollection-based memory retrieval (e.g., hippocampus) suggesting that the two groups may utilize different strategies or processes when retrieving emotional information from memory. |
