Work

Rethinking the effects of working memory training on executive functions in schizophrenia: A machine learning approach

Year: 2025

Type: article

Abstract: Background: Executive dysfunction in schizophrenia profoundly impairs functional outcomes and remains insufficiently addressed by standard pharmacological treatments. Although computerized cognitive training offers a promising intervention, traditional evaluation methods often fail to capture nuanced improvements along the psychosis-health continuum. This study aims to quantify changes in executive function (EF) profiles following cognitive training and identify candidate causal predictors of treatment response.Methods: Patients with schizophrenia were randomized into adaptive N-back training (n = 32), non-adaptive 1-back control (n = 33), or treatment-as-usual (n = 29) groups. EF was evaluated across working memory, cognitive flexibility, and inhibitory control domains at baseline and post-intervention. A support vector machine classifier, trained on an independent sample (195 schizophrenia patients, 169 controls) and calibrated via Platt scaling, quantified EF profile changes. An integrative causal machine learning framework identified baseline predictors of treatment response.Results: Adaptive working memory training resulted in significant near-transfer effects on untrained working memory updating tasks and reduced general psychopathology symptoms (pfdr< 0.05). The probability of being classified as having a neurotypical EF profile increased substantially in the adaptive training group (from 13.21% at baseline to 38.79% at follow-up, p < 0.001), with these changes correlating with symptom reduction. Working memory maintenance and response inhibition emerged as robust causal predictors of treatment response.Conclusions: Working memory training induces meaningful shifts in EF profiles in schizophrenia, promoting movement along the psychosis-health continuum toward neurotypical functioning. The classifier-based approach provides a more refined assessment of cognitive gains compared to traditional binary measures, while the causal analysis identifies specific EF domains that predict treatment response.

Source: International Journal of Clinical and Health Psychology

DOI: 10.1016/j.ijchp.2025.100628

Authors: Tongyi Zhang, Meifang Su, Xiaoning Huo, Xin Zhao

Institution: Northwest Normal University

Language: English

Topic: Schizophrenia research and treatment

Subfield: Psychiatry and Mental health

Field: Medicine

Domain: Health Sciences

SDG: Good health and well-being

Open Access status: gold

Estimated APC paid: $2,970