Autism spectrum disorder (ASD) traits modulate context-dependent use of prior information in vocoded speech perception

Predictive processing accounts propose that perception relies on integrating top-down prior knowledge with bottom-up input. In degraded speech, successful perception relies on learning from previous listening experiences and generalising this knowledge to novel contexts. Autistic spectrum disorder (ASD)-related traits have been linked to reduced generalisation, yet it is unclear how this affects generalisation of prior events. We assessed learning via lexical (sentence-level) and perceptual-acoustic (speaker-level) generalisation for vocoded speech, and examined how these are moderated by ASD traits. 42 participants with/without ASD diagnosis completed a 2 (Sentence Exposure: Old, New) x 2 (Speaker Exposure: Old, New) within-subjects vocoded speech task, where they listened to and transcribed previously presented vs. novel degraded sentences. ASD traits were expected to be associated with reduced generalisation and thus lower speech perception accuracy for both novel sentences and speakers. Participants were divided into Low (N=21) and High (N=21) ASD-trait groups using k-means clustering of Social Responsiveness Scale-2 (SRS-2) subscales, measuring ASD-related traits. Mixed-effects modelling showed both groups performed comparably overall. Group differences in accuracy depended on both prior sentence and speaker exposure (Sentence*Speaker*Group interaction): High-ASD group showed a sentence-exposure benefit only when the speaker was previously encountered, whereas Low-ASD group benefitted from sentence priors regardless of speaker identity. Contrary to predictions, higher ASD traits were not associated with reduced performance in all novel conditions but instead only reduced generalisation of sentence priors across speaker changes. This may reflect greater sensitivity to lower-level acoustic variation between speakers, resulting in more context-specific/constrained use of higher-level lexical priors.

Authors

Institutions

Publication Details

Published
2026-11-14
DOI
https://doi.org/10.33774/coe-2026-zm4x1
Primary Topic
Autism Spectrum Disorder Research
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Autism spectrum disorder (ASD) traits modulate context-dependent use of prior information in vocoded speech perception

Anastasia Klimovich‐Gray, Agnieszka Konopka, Irina Guliaeva, Philip Alex et al.
Autism Spectrum Disorder Research
preprint

Autism spectrum disorder (ASD) traits modulate context-dependent use of prior information in vocoded speech perception

Anastasia Klimovich‐Gray, Agnieszka Konopka, Irina Guliaeva, Philip Alex, Marta Brzeska
preprint en

Abstract

Predictive processing accounts propose that perception relies on integrating top-down prior knowledge with bottom-up input. In degraded speech, successful perception relies on learning from previous listening experiences and generalising this knowledge to novel contexts. Autistic spectrum disorder (ASD)-related traits have been linked to reduced generalisation, yet it is unclear how this affects generalisation of prior events. We assessed learning via lexical (sentence-level) and perceptual-acoustic (speaker-level) generalisation for vocoded speech, and examined how these are moderated by ASD traits. 42 participants with/without ASD diagnosis completed a 2 (Sentence Exposure: Old, New) x 2 (Speaker Exposure: Old, New) within-subjects vocoded speech task, where they listened to and transcribed previously presented vs. novel degraded sentences. ASD traits were expected to be associated with reduced generalisation and thus lower speech perception accuracy for both novel sentences and speakers. Participants were divided into Low (N=21) and High (N=21) ASD-trait groups using k-means clustering of Social Responsiveness Scale-2 (SRS-2) subscales, measuring ASD-related traits. Mixed-effects modelling showed both groups performed comparably overall. Group differences in accuracy depended on both prior sentence and speaker exposure (Sentence*Speaker*Group interaction): High-ASD group showed a sentence-exposure benefit only when the speaker was previously encountered, whereas Low-ASD group benefitted from sentence priors regardless of speaker identity. Contrary to predictions, higher ASD traits were not associated with reduced performance in all novel conditions but instead only reduced generalisation of sentence priors across speaker changes. This may reflect greater sensitivity to lower-level acoustic variation between speakers, resulting in more context-specific/constrained use of higher-level lexical priors.

University of Aberdeen (GB)
Autism Spectrum Disorder Research
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon