Functional mapping of dynamic happy and fearful facial expressions in young adults with familial risk for psychosis - Oulu Brain and Mind Study.

Background: Social interaction requires mirroring to other people's mental state. Psychotic disorders have been connected to social interaction and emotion recognition impairment. We compared the brain activity between young adults with familial risk for psychosis (FR) and matched controls during visual exposure to emotional facial expression. We also investigated the role of the amygdala, the key region for social interaction and emotion recognition.

Methods: 51 FR and 52 control subjects were drawn from the Northern Finland 1986 Birth Cohort (Oulu Brain and Mind Study). None of the included participants had developed psychosis. The FR group was defined as having a parent with psychotic disorder according to the Finnish Hospital Discharge Register. Participants underwent functional MRI (fMRI) using visual presentation of dynamic happy and fearful facial expressions. FMRI data were processed to produce maps of activation for happy and fearful facial expression, which were then compared between groups. Two spherical regions of interest (ROIs) in the amygdala were set to extract BOLD responses during happy and fearful facial expression. BOLD responses were then compared with subjects' emotion recognition, which was assessed after fMRI. Psychophysiological interaction (PPI) for the left and right amygdala during happy and fearful facial expression was conducted using the amygdala as seed regions.

Results: FR subjects had increased activity in the left premotor cortex and reduced deactivation of medial prefrontal cortex structures during happy facial expression. There were no between-group differences during fearful facial expression. The FR group also showed a statistically significant linear correlation between mean amygdala BOLD response and facial expression recognition. PPI showed that there was a significant negative interaction between the amygdala and the dorsolateral prefrontal cortex (dlPFC) and superior temporal gyrus in FR subjects.

Conclusions: Increased activations by positive valence in FR were in brain regions crucial to emotion recognition and social interaction. Increased activation of the premotor cortex may serve as a compensatory mechanism as FR subjects may have to exert more effort on processing the stimuli, as has been found earlier in schizophrenia. Failure to deactivate PFC structures may imply error in the default mode network. Abnormal PFC function in FR was also suggested by PPI, as the dlPFC showed decreased functional connectivity with the amygdala in the FR group. This may indicate that in FR subjects the amygdala have to take a greater role in emotion recognition and social functioning. This inference was supported by our discovery of statistically significant correlations between the amygdala BOLD response and emotion recognition in the FR group but not in controls.