Sensory Characteristics and Autistic Traits Influence Neural Responsivity to Predictable Versus Unpredictable Visual Information

Background:  Hypo- and hyper-sensory response and preference for predictability are common features of autism spectrum disorder (ASD). Prior research investigating sensory sensitivities and using event-related potentials (ERPs) has demonstrated atypical early visual processing in individuals with ASD. However, most ERP paradigms are repetitive and temporally predictable. Because early visual brain response is influenced by expectancy, the temporal regularity of previous ERP experiments represents a potential confound, raising the possibility that differences attributed to abnormalities in low level visual processing might reflect atypical response to predictable stimuli rather than disruptions in the functional integrity of the visual pathway.

Objectives: This study investigates the (a) relationship between sensory processing, as indexed by visual evoked potentials (VEPs), and the predictability of visual stimulation and (b) the degree to which this relationship is modulated by autistic traits and sensory characteristics.

Methods: ERPs were recorded from 10 typically developing adults (data collection is ongoing) using 128-channel sensor nets. Eye movements were recorded concurrently with an SR eye-tracking system. In two experimental paradigms, black and white checkerboards appeared on screen and reversed phase. In the predictable condition, the phase reversal occurred every 500ms. In the unpredictable condition, reversal occurred randomly between 300 and 1000ms. ERPs were time-locked to each phase reversal; occipital P1 and N1 amplitude and latency were extracted for analyses. Self-report questionnaires captured sensory features (Glasgow Sensory Questionnaire; GSQ) and autism characteristics (Broad Autism Phenotype Questionnaire; BAP-Q; Social Responsiveness Scale; SRS).

Results:  Preliminary data analyses indicated that predictability of visual change had a statistically significant effect on P1 amplitude, F(1,9)=6.323, p=.03, such that P1 amplitudes were larger for the unpredictable condition. Larger N1 peak amplitudes in the unpredictable condition correlated with greater subthreshold autistic symptomatology (SRS Restricted Interest and Repetitive Behavior subscale, r=.707, p<.05). Difference scores for P1 amplitude between the unpredictable and predictable conditions correlated with the Visual Modality of the GSQ, r=.764, p<.05, such that increased amplitudes in the unpredictable condition explained higher levels of visual sensitivity. Additionally, difference scores for P1 latency demonstrate that earlier P1 responses in the unpredictable compared to the predictable condition were associated with an increased level of autistic traits (BAP-Q Pragmatic Language Subscale, r=-.773, p=.009).

Conclusions: Distinct neural responses were elicited by visual stimuli presented at predictable versus unpredictable rates. Moreover, variability associated with expectancy explained variance in the autism phenotype. Individuals who displayed enhanced or more rapid response to unpredictable stimuli reported increased visual sensory sensitivities, higher level of social-communication difficulties, and more repetitive interests and behaviors. Ongoing analyses will explore oscillatory harmonics of the VEP and the relationship between pupil dilation, brain activity, and symptomatology. These data show that symptom variability is associated with both early visual processing and top down expectancies and that these relationships are dissociable. By exploring the intersection of top-down and bottom-up sensory driven brain activity we are better poised to determine how these factors influence sensory and social symptomology and uncover sources of heterogeneity in ASD.