Objectives: We sought to characterize the neural substrates of abnormal gaze fixation in autism utilizing functional magnetic resonance imaging (fMRI). Specifically, we asked how brain regions involved in social cognition would differ in subjects with autism versus IQ-matched typically developing comparison subjects in response to dynamic social stimuli of somewhat greater ecological validity than commonly used still pictures of faces, by presenting a common and relatively simple social scenario.
Methods: Six subjects with autism (mean age 25.7 years, 5 males) and 4 typical controls (mean age 33.7 years, 4 males) underwent high-resolution fMRI in a 3-T scanner while viewing 6-sec computer-generated virtual-reality movies of a man approaching down a hallway (Run 1). In a second condition, the same videos were shown with a fixation cross-hair moving over the eyes of the figure; subjects were instructed to maintain fixation on the cross-hairs (Run 2). Subjects viewed 20 trials of each condition. Functional regions of interest (ROIs) were drawn in the FFG, STS, and AMY based on average activation of all subjects across both conditions. The ROIs were then applied to each group and condition separately, and hemodynamic responses were extracted for each region.
Results: The task resulted in significant activation across subjects in the right STS, bilateral FFG, and bilateral AMY. Examination of the hemodynamic response revealed an effect of group on latency in the right STS and bilateral FFG, with the onset of the BOLD response premature in autism as compared to controls for both conditions. Additionally, activation in the right FFG was greater for subjects with autism than for controls.
Conclusions: The FFG is known to be selectively activated by faces in typical individuals. The latency effect suggests that, rather than a selective response to the face that is stronger towards the end of the trial, when the virtual actor is closer to the subject and occupying more of the visual field, subjects with autism display a broadly responsive activation that peaks earlier, indicative of a lack of selectivity to the face stimuli. The earlier onset of activation in turn drives an overall greater magnitude at peak in subjects with autism as compared to neurotypical subjects. Similarly, the STS may be less specifically responsive to socially-relevant motion in autism, responding more broadly to motion in general. These findings implicate a relative lack of specialization of brain regions involved in social cognition in individuals with autism.