EARLY IDENTIFICATION OF PSYCHOSIS USING BEHAVIORAL, COGNITIVE, AND NEUROIMAGING MARKERS

Abstract

Prompt intervention through early psychosis identification is needed in order to change the trajectory of the disease and improve the long-term outcomes. In this paper, the effectiveness of behavioural, cognitive and neuroimaging markers, both individually and together, in early identification of psychosis was reviewed. The multimodal framework was used comprehensively for healthy controls, clinical high risk and early psychosis patients. The behavioral analyses showed that the responses were more variable and there was less adaptive performance in the high-risk and the early psychosis group. Cognitive assessments showed that there was a significant deficit in executive functioning, working memory and speed of processing, with the deficits becoming increasingly worse as clinical severity increased. These alterations in the structure of the brain, the rupture of connections between significant neural networks, which are involved in cognition, manipulation of significant information and emotional control, were observed by neuroimaging. Using these markers in conjunction with multimodal feature fusion with machine learning based classification models, these markers outperformed the single-modality. Longitudinal data revealed that integrated risk scores were sensitive to the symptom evolution, which highlights their possible prognostic value. Overall, the results suggest that abnormalities associated with the psychosis will be apparent in behavioural, cognitive and neurobiological domains before the onset of the disease itself. The study will validate the clinical effectiveness of multimodal assessment strategies in identifying psychosis and risk stratification at an early stage, and provide a precision mental health care intervention in a scalable and data-driven way.