Viewpoint on Multisensory Integration in Psychiatric Disorders

22/08/2026

AI Summary

Shift from descriptive accounts to normative computational models (maximum likelihood estimation, causal inference), with goal of linking computations to neural circuits and clinical heterogeneity.
Inconsistent clinical findings, limited statistical power, and challenges defining and measuring multisensory function highlight need for larger, better characterised, well phenotyped cohorts.
Emerging directions include naturalistic paradigms, adaptive coding, open team science, precision psychiatry, and AI analysis of multimodal datasets; promise for perception, cognition, belief updating.

AI Summary

The discussion highlights the shift from descriptive accounts of multisensory phenomena toward normative models, including maximum likelihood estimation, causal inference, and correlation detection, as well as the need to link these computations to neural circuits and clinical heterogeneity.

The authors consider ongoing challenges in the field, including inconsistent findings across clinical studies, limited statistical power, difficulties in defining and measuring multisensory function, and the need for larger, better-characterized cohorts.

Basic summary

Multisens Res. 2026 Aug 18:1-10. doi: 10.1163/22134808-bja10207. Online ahead of print.

ABSTRACT

This article presents a compiled interview with three researchers in the field of multisensory integration, with an emphasis on how multisensory integration may go awry in psychiatric and neurodevelopmental conditions. Through structured dialogue, the authors reflect on the origins of the field, the development of key empirical and computational frameworks, and the growing relevance of multisensory science for understanding autism, schizophrenia, dyslexia, and related conditions. The discussion highlights the shift from descriptive accounts of multisensory phenomena toward normative models, including maximum likelihood estimation, causal inference, and correlation detection, as well as the need to link these computations to neural circuits and clinical heterogeneity. The authors consider ongoing challenges in the field, including inconsistent findings across clinical studies, limited statistical power, difficulties in defining and measuring multisensory function, and the need for larger, better-characterized cohorts. They also discuss emerging directions, including naturalistic paradigms, adaptive coding, open and team science, precision psychiatry, and the application of artificial intelligence to complex multimodal datasets. The interview concludes by emphasizing the promise of multisensory integration as a framework for understanding perception, cognition, and belief updating in health and disease.

PMID:42628959 | DOI:10.1163/22134808-bja10207

Document this CPD