Infants at high and low likelihood for autism show different EEG developmental trajectories in speech tracking and statistical learning.

The study mapped early speech-processing trajectories and found that weaker syllable tracking related to later verbal scores. These EEG measures are research candidates, not diagnostic tests for autism.

Godel M, Fló A, Benjamin L, Dehaene-Lambertz G, Schaer M.
eLife · Publication date: 2026-09-21
Added to site: 2026-09-24
DOI: 10.7554/elife.109901

Publication date verified in Europe PMC and the full-text article XML; first indexed 2026-09-22. This source date is separate from the site's addition date.

Modalities

EEG, ERP

Clinical focus

Language development, Autism likelihood, Longitudinal biomarkers

Population

Infants aged 2.5 to 22.6 months, with verbal outcomes assessed around 20 months

Sample

83 high-density EEG recordings from 44 infants: 25 at high likelihood and 19 at low likelihood for autism

Intervention / method

Longitudinal neural entrainment to syllable and word frequencies during multi-speaker artificial speech, followed by an ERP word-recognition task

Comparator

High- versus low-likelihood groups across development

Clinician summary

High-likelihood infants had reduced syllable-level tracking, while statistical learning itself appeared preserved. They also showed reduced late novelty-related ERP responses. Analyses excluding the final visit retained associations with later verbal outcomes.

How to interpret the findings

Small cohort, uneven age distribution, limited repeated measures, 19.4% recording exclusion, sleep-state effects, unquantified additional language exposure and no established autistic specificity. High likelihood does not mean every infant received or will receive an autism diagnosis.

Equipment and disclosures

High-density EEG; specific acquisition hardware was not summarized in the indexed record.

Supported by six Swiss National Science Foundation grants. The authors declared no competing interests, and the full text states that funders had no role in study design, data collection, interpretation or publication decisions.

Related evidence

Compare findings from the same clinical area, including mixed results.