Neurophysiological Measures of Sensory Processing and Integration Are Predictive of Autistic Symptom Severity, but Not of Sensory Sensitivities, in a Group of High-Functioning Children and Adolescents with ASD

Background:  Difficulties processing and integrating sensory information have long been noted in individuals with autism spectrum disorders (ASD), and it has been suggested that this contributes significantly to some of the major symptoms of autism, including aberrant language development, unusual and repetitive behaviors, and deficits in social perception.  In light of these potentially far-reaching consequences, surprisingly few studies have explored the neurophysiological basis of sensory processing atypicalities, or how these might relate to the clinical presentation of ASD. 

Objectives:  To assess whether neurophysiological measures of sensory processing and multisensory integration are predictive of autistic symptom severity and of sensory sensitivities in high-functioning children and adolescents with ASD.    

Methods:  The sample consisted of 44 individuals between the ages of 6-17 years who met criteria for a diagnosis of ASD.  Event related potentials (ERP) data were collected while participants performed a simple speeded detection task in response to auditory and visual stimuli presented together and alone, in random order.  The ERP components of interest were the major sensory deflections recorded in the first 200 ms, decided a priori and including the auditory P1, N1a, N1b, and N1c, the visual P1 and N1, and 3 multisensory effects between 100-210 (based on prior research).  A linear regression analysis was conducted to investigate whether ERP responses to basic auditory, visual, and audiovisual stimuli predict the severity of current ASD symptoms, as indexed by the Calibrated Severity Scores (CSS) from the Autism Diagnostic Observation Schedule (ADOS).  An additional hierarchical linear regression analysis examined whether these same ERP measures predict the severity of visual and auditory sensitivities (VAS), as reported by parents on the Short Sensory Profile.  Both analyses controlled for participant characteristics (age, sex, VIQ, PIQ, race, and maternal education) that were shown to correlate with our outcome measures.

Results:  The linear combination of ERP measures was significantly related to CSS, with about half of the variance of autistic symptom severity in the sample accounted for by measures of basic sensory responses and multisensory integration.  The best ERP predictors of CSS were the auditory N1a and N1b, and multisensory integration between 100-130 ms over frontal sites.   In contrast, the linear combination of ERP measures did not account for a significant proportion of the variance in VAS after controlling for the effects of verbal IQ (which was highly correlated with VAS).

Conclusions:  Brain responses to basic sensory stimuli predict severity of autistic symptoms in a group of high-functioning individuals with ASD.  This provides preliminary evidence that early sensory processing differences  may well  contribute to the core symptoms of ASD.  Perhaps even more importantly, this study supports the use of ERPs as a potential tool for developing endophenotypic markers for ASD, an approach that may be crucial in the future for identifying, characterizing and subtyping  this complex group of neurodevelopmental disorders.