Reduced Frontal P1 Amplitude Differentiation As a Neural Signature of Speech Sound Disorder in ASD

Background:  Auditory statistical learning supports acquisition of language and social skills and is therefore of particular interest in autism spectrum disorder (ASD). The extant literature is inconclusive regarding the role of statistical learning in language deficits associated with ASD. Prior research on this topic has been limited by inclusion of only high functioning individuals and reliance on explicit measurement of this implicit process (e.g. Mayo & Egsti, 2012). Moreover, the heterogeneity of the ASD phenotype has potential to mask patterns of neurodevelopmental differences among clinical subgroups. The P1 event related potential (ERP) is an index of auditory cortical development (Sharma et al., 2005) that can be measured during passive statistical learning tasks, and thus has potential as a marker of disrupted, bottom-up processing of auditory information in ASD.

Objectives:  The current study aims to characterize the P1 ERP component in youth with ASD in the context of auditory statistical learning. We hypothesize that the ASD group will have more difficulty encoding and recognizing auditory stimuli, as evidenced by reduced frontal P1 differentiation between learned and unlearned sequences of speech-like sounds. Additionally, we expect that within the ASD group, weaker frontal P1 differentiation will be associated with a profile of lower verbal IQ, verbal memory deficits, and impaired speech production.

Methods: 102 youth with ASD (n=58), comorbid ASD and speech sound disorder (ASD+SSD; n=12), or no neurodevelopmental diagnosis (n= 32) ages 7 to 17 years (mean = 12.74) participated in electroencephalogram (EEG) testing as part of two larger studies of ASD. ASD and comorbid diagnoses were established by trained clinical professionals following a clinical evaluation that included thorough clinical interview, cognitive and behavioral assessment. ERPs were measured during an eight minute, passive, auditory word segmentation task. During an initial exposure period, participants listened to streams of nonword syllabic phonemes. This was followed by a test phase wherein both novel and previously heard tri-syllabic combinations were randomly presented across 48 trials.

Results: Results of the two-level, repeated measures analysis are presented in Table 1. Auditory statistical learning in the test phase was evidenced across the entire sample by a significantly stronger mean frontal P1 amplitude response to the learned versus unlearned stimuli. Age attenuated this effect, with older youth demonstrating smaller differences between stimuli and weaker mean amplitudes overall. Contrary to expectations, ASD diagnosis, verbal IQ and verbal memory did not moderate the effect. However, ASD+SSD youth did show significantly weaker differentiation between learned and unlearned stimuli, indicating impaired auditory statistical learning within this clinical subgroup.

Conclusions: Increasingly, research on ASD is emphasizing the heterogeneity of the disorder, including variability in psychiatric comorbidity. In the current study, abnormal frontal P1 amplitude is identified as a neural signature of comorbid ASD + SSD. The results are consistent with a deficit in bottom-up processing of auditory input that affects encoding and reproduction of speech sounds, but may also be consistent with a bi-directional effect, wherein poor speech production impairs refinement of specialized neural circuits that support phonemic awareness.