Social Responsiveness Scale Predicts Activity in Limbic Regions for An Emotion Recognition Task

Background: The Social Responsiveness Scale (SRS) is a questionnaire that correlates strongly with diagnoses of autism and is thought to reflect stable behavioural traits that are normally distributed within typical human populations.  It therefore seems likely that neurobiological differences that contribute to the variability of SRS score will inform us about the neurobiology of autism spectrum disorder. This could stem from sensitivity to emotional stimuli or capacity to direct attention towards emotional stimuli.

Objectives: To investigate its neurobiological significance and relationship to emotional stimuli, SRS was used as a correlate in the analysis of an fMRI task of emotion recognition for a typically developing cohort of children and adolescents.

Methods: 33 subjects (23 female, 10 male), right-handed, aged 9.8-16.4 (mean age = 13.8, std. dev. 2.1) were considered in the analysis. Imaging (MRI) was performed using a 3T scanner. The functional protocol consisted of faces displaying an emotional expression that was either happy or fearful, subtle or obvious, and directed at or away from the observer. Conditions were presented in blocks for the tasks of identifying either the emotional state, or the sex of the presented face. Each face was positioned in the centre of the display, accompanied by a visual display of the words 'Happy' and 'Fearful' or 'Male' 'Female', in the bottom left and right corners. The participant was asked to choose the relevant word and respond via a push-button system activated by the left or right index finger.  MRI data were processed and analysed using SPM8 (http://www.fil.ion.ucl.ac.uk/spm/). A whole brain voxel-based single group analysis was performed to produce activation contrasts for task dimensions of attention, emotional valence, emotional arousal and face direction. SRS data was applied to the model as a linear regressor to search for correlations between the metric and contrasts of interest. Statistical significance was considered at thresholded p < 0.001, FWE-corrected p < 0.05 at cluster level.

Results: The amount of activity in limbic and paralimbic areas in response to fearful expression relative to happy expression correlated robustly with score on the Social Responsiveness Scale (clusters included right medial temporal lobe: 701 voxels, z=4.80; right lateral orbitofrontal cortex: 1202 voxels, z=4.77 and anterior cingulate 2220 voxels, z=4.73). SRS score did not correlate with any differences relating to attention or gaze direction.

Conclusions: Higher SRS score was not determined by absolute sensitivity to emotional expression but a higher reactivity to fearful relative to happy emotional expression. Our findings suggest that variability on the autism construct among typical adolescents could be determined by degree of reactivity to expression of negative emotions relative to reactivity to expression of positive emotions.