Examination of the Role of Social and Sensory Factors in Atypical Speech Processing in Autism Spectrum Disorder

Background: Language and social development are intimately linked. Individuals with Autism Spectrum Disorder (ASD) display differences in language processing, including atypicalities in speech sound discrimination and lateralization of speech processing. It is unclear, however, whether these differences are related to the inherent social nature of speech or its unique sensory features. In light of evidence for both broad sensory and specific auditory processing abnormalities in ASD, further research is necessary to determine how atypicalities in language processing are related to social and sensory aspects of speech.

Objectives: To investigate brain responses to speech and non-speech stimuli in children with ASD using event-related potentials, in order to elucidate how abnormalities in speech processing relate to social and sensory factors associated with speech.

Methods: Participants are children with ASD aged 3-5 years and chronological age-matched typically developing (TD) controls (see Table 1 for sample characteristics). ASD participants were diagnosed by research-reliable doctoral level clinicians using best clinical judgment based on a variety of measures (ADOS-2, Mullen Scales of Early Learning, Vineland Adaptive Behavior Scales). Speech stimuli in this on-going study consist of short nonsense phrases spoken with angry or neutral prosody, selected based upon affect ratings (interrater reliability ≥.875). Non-social sensory-matched controls for the speech sounds were created by synthesizing a complex tone with the same pitch, amplitude, and overtones as the original speech. Speech and non-speech stimuli are presented in separate blocks, each contrasting prosody (angry, neutral). This experimental design allows for analysis of broad (speech vs. non-speech) and specific (angry vs. neutral) social factors while controlling for sensory features. Analyses focus on the P2 component and a late positive component (P450) over frontal cortex, in order to examine both perception and cognitive evaluation of the stimuli. ANOVAs with Stimulus Type (Human, Non-Human), Emotion (Angry, Neutral), and Hemisphere (Left, Right) as within-subjects factors and Group (ASD, TD) as a between-subjects factor were conducted on the amplitudes of these components.

Results: Perceptual P2 component amplitudes exhibited a significant Group by Hemisphere interaction whereby P2 amplitudes were larger in the left hemisphere for ASD participants but larger in the right hemisphere for TD participants (F(1, 13) = 6.265, p<.05, η_p²=.325). Amplitudes of a frontal P450 cognitive component also displayed an interaction effect between Stimulus Type and Group in which TD participants exhibited larger discrimination of speech and non-speech stimuli, (F(1, 13) = 6.838, p<.05,  η_p²=.345).

Conclusions: Preliminary analyses found that ASD participants exhibit atypical left-lateralization of auditory perceptual processing regardless of the stimulus, suggesting the possibility that abnormal hemispheric specialization is independent of both social and prosodic factors. However, these preliminary results suggest that ASD participants also exhibited reduced discrimination of speech and non-speech sounds relative to control participants at a later cognitive stage of processing. Because the speech and non-speech stimuli were carefully matched on physical properties, this finding suggests that impairments in later cognitive processing of speech versus non-speech in ASD may be driven by social factors. Data collection is on-going and full sample ASD and TD data will be reported.