Neurophysiological alterations during sensory processing in autism - a meta-analysis

While sensory processing alterations in autism are well-documented, the neurophysiological correlates remain unclear. This meta-analysis examined differences in early event-related potentials (ERP) and event-related fields (ERF) between autistic and non-autistic individuals using electroencephalography and magnetoencephalography to identify neurophysiological alterations that may underlie variations in sensory perception, communication, and social interaction in autism. Following PRISMA guidelines, a database search was conducted for peer-reviewed studies published from January 1980 onwards. Random-effects meta-analyses were performed using the metafor package in R. Standardised mean-group differences in early ERP/ERF latencies and amplitudes were analysed with moderator analyses exploring demographic and methodological factors, including neurophysiological technique, sensory modality, age group, sex, and language impairment. 145 studies (3778 autistic, 3484 non-autistic participants) were included. Autistic individuals exhibited significantly longer latencies in P/M50 (SMD = 0.44; SE = 0.21; 95% CI 0.03-0.86; p = 0.04), P/M100 (SMD = 0.18; SE = 0.08; 95% CI 0.01-0.36; p = 0.03), N170 (SMD = 0.33; SE = 0.12; 95% CI 0.10-0.56; p = 0.01), and P/M200 (SMD = 0.30; SE = 0.09; 95% CI 0.12-0.48; p = 0.00) components. P/M50 showed the greatest latency alteration, with an effect-size nearing medium, especially in individuals with language impairment (Q_B(2) = 7.70, p = 0.02), followed by N170 most notable in autistic adolescents and adults (Q_B (3) = 12.30, p = 0.01). No significant amplitude alterations were found, and substantial heterogeneity was observed. Neurophysiological characteristics of sensory processing in autism implicate multiple mechanisms and stages given prolonged P/M50- and N170-latency (associated with sensory filtering challenges and social perception alterations, respectively). These component timings show potential as biomarkers, though heterogeneity and modest effect-sizes limit clinical application, highlighting the need for further research.