Enhanced Early Visual Responses to Emotional Faces in ASD

Background:

Sensory atypicalities in all the modalities in ASD have recently gained a crucial role in the description of the autistic symptomatology. In the visual domain, atypical neural responses have been reported in several low- and high- level tasks. Facial expressions, are essential for successful social interactions and represent a core problem for those with ASD. To better understand implicit and rapid processing of emotions in ASD, the visual cortex and emotional network should be considered in neuroimaging studies from its earliest stages.

Objectives:

The present study investigated the time course of brain activity in a large number of regions implicated in various stages of the processing of emotional faces, utilizing both the high spatial and temporal resolution of magnetoencephalography (MEG).

Methods:

Nineteen ASD adults and 19 typically developed adults (TD) took part in an implicit emotional task while MEG was recorded. Stimuli consisted of a face displaying a happy, an angry or a neutral expression and a scrambled pattern presented simultaneously just right and left of central fixation. The participants responded whether the pattern occurred on the left or right side of the screen. Emotions were thus implicit in the task. MEG responses in 23 bilateral regions of interest (ROI) were selected and seven time-windows of 30ms were visually identified to measure mean power. Statistical analyses were performed at each ROI when a deflection was present to investigate group emotion, hemisphere and interaction effects.

Results:  Angry faces elicited a stronger response in both groups (110-140ms) in the fusiform gyri (FG), but only the TD presented similar emotional enhancement in the inferior temporal gyri. Hypo-activation in the postcentral gyrus was found in the ASD group to all facial expressions. In the occipital region and the middle temporal poles (125-155ms) the ASD group presented an enhanced activity and an emotion-specific response, while a hypo-activation in the right FG, the bilateral postcentral gyri and in the parietal cortex was found around 155-185ms. Happy faces elicited a greater parietal response in TD (220-250ms) while neutral faces elicited a stronger response than happy faces in the occipital gyri in the ASD (245-275ms). An emotion-specific responses were found (315-345ms) in the FG and in the parietal lobules for both groups. Differences between groups in hemisphere dominance were observed at several ROI and time-windows.

Conclusions:

These results demonstrate a distinct spatio-temporal organization during the processing of implicit emotional faces in adults with ASD. Consistent with perceptual and sensory theories of autism, an early enhancement in the visual cortices was found in the ASD group. Thus, the atypical implicit facial expression processing is not only due to hypo-activation of the FG, but to a broader atypical emotional face network, including atypical visual and emotion-specific responses. According to reports of enhanced visual functioning and to difficulties in configural face processing, a visual bias might be partly responsible for the impairment in the emotional face response in ASD. Thus, our data indicate that atypical early visual responses should be considered while investigating social cognition in autism.