Intra-Individual Variability in Behavioural and fMRI Sensory-Evoked Responses in Autism

Background:  Autism is a neurodevelopmental disorder that affects neural circuitry across the brain.  One indicator of altered cortical processing in autism is greater variability in sensory-evoked responses (found using fMRI and EEG) from one trial to the next, suggesting unreliable sensory-evoked responses (Milne, 2011; Dinstein et al., 2012).

Objectives:  To confirm the greater variability in fMRI sensory-evoked responses across sensory modalities in autism and to investigate the functional significance of the greater cortical variability in ASD by assessing variability in roughness perception.  A subgroup of participants with ASD participated in both experiments to allow for direct comparison between cortical and behavioural responses.

Methods:  Two experiments were conducted.  For experiment 1, we used the same method as reported by Dinstein et al. (2012), to measure the visual (moving white dots), auditory (pure tone beeps) and somatosensory responses (air puffs to the back of the hand) in primary sensory cortices in a new group of individuals with autism.  Each stimulus was repeated 24 times over two scanning sessions.  The data from the original study by Dinstein et al. (2012) were used to compare the responses from the new autism group (N=12) to the original autism group (N=14).  For experiment 2, 17 individuals with autism and 17 age- and gender-matched control were asked to rate 12 surfaces on roughness using a Likert scale of 1 (not very rough) to 10 (very rough).  Each plate was presented 14 times to gauge a measure of variability.  A sub-set of individuals with autism took part in both experiments (N=13).  

Results:  For the roughness experiment, we found that the individuals with autism perceived the surfaces as rougher and were more variable in their ratings (from trial to trial) compared to matched controls.  For the fMRI study, we found similar results in the new autism group compared to the original autism group (from Dinstein et al., 2012): greater trial-to-trial variability in sensory-evoked responses and weaker SNR compared to typical individuals.  For the sub-group of individuals with autism who took part in both roughness perception and fMRI experiments, those with greater fMRI response variability rated the plates as feeling rougher, suggesting a functional correlate of the variable sensory responses, and their contribution to the sensory sensitivities that are commonly associated with autism.  In addition, greater variability in the roughness ratings was associated with greater fMRI response variability in the visual cortex.  This was despite participants wearing headphones playing white noise and an eye mask.  This is consistent with the hypothesis that visual imagery is used to help interpret tactile stimuli, and consequently, that greater response variability in visual cortex may impact tactile perception. 

Conclusions:  Individuals with autism produce more cortical and behavioural intra-individual variability than controls, and variability in fMRI responses correlates with behavioural responses.  This suggests that there is a functional consequence of the greater intra-individual variability, and that the variability is an endophenotype in autism.  Unreliable sensory responses could contribute to the unusual sensory sensitivities that are a part of the DSM-V criteria for autism.