Authors
Samantha Mombelli, Samaneh Roghani Zanjani, Kian Godhwani, Alban Voppel, Aryamman Aiyaar, Tihare Zamorano, Timothy Friesen, Ashley Choucroun, Valentin Guigon, Caroline J Charpentier, Franziska Knolle, Ryan Smith, Jerome Waldispuhl, Nace Mikus, Christoph Mathys, Andreea O Diaconescu, James Gilleen, Lena Palaniyappan, Jai Shah, Albert Powers, David Benrimoh
Published in
Journal of psychiatry & neuroscience : JPN. Volume 51. Pages 1-13. Jan 01, 2026.
Abstract
Computational psychiatry has advanced through two parallel traditions: theory-driven models that aim to explain the mechanisms underlying psychiatric symptoms, and data-driven approaches designed to improve prediction and clinical decision-making. Although both approaches have generated important insights, their integration could support clinical translation. Using psychosis onset prediction as a case study, we propose a mechanism-first, iterative framework in which computational models, empirical data, and predictive tools inform and refine one another. Within this framework, theory guides model and task design, while data-driven approaches support stratification, validation, and comparison across computational hypotheses. This integration may improve generalizability across clinical cohorts, support individualized risk estimation, and facilitate the identification of mechanistically informed subgroups. We also discuss the potential role of computational parameters as biomarker-like features within predictive models, alongside challenges related to reliability, scalability, and ethical implementation. Integrating theory- and data-driven approaches may help computational psychiatry develop clinically useful and mechanistically interpretable models.
PMID:
43018144
Bibliographic data and abstract were imported from PubMed on 12 Nov 2026.
Read full publication at:
Please sign in
to see all details.
Advertisement
Stats
- Recommendations n/a n/a positive of 0 vote(s)
- Views 10
- Comments 0