Schizophrenia ranks among the most debilitating neuropsychiatric disorders, affecting roughly 1% of people worldwide. It manifests through a range of symptoms, from hallucinations and delusions-collectively known as psychosis-to cognitive impairments, diminished motivation, and depression. Typically diagnosed in individuals’ around late adolescence, schizophrenia leads to significant life disruptions, adversely affecting educational achievements, employment prospects, and social interactions. Over time, individuals with schizophrenia place considerable mental and financial strain on their families from an early stage and become progressively isolated, eventually becoming dependent on state resources, thereby imposing a societal financial burden.
Research underscores the benefits of early diagnosis in schizophrenia, including better treatment outcomes, reduced duration of illness and functional decline, and a lower risk of progressing from clinical at-risk mental states to full-blown psychosis. Despite ongoing efforts to pinpoint a biomarker for schizophrenia, there has been a burgeoning interest in the early identification and treatment of individuals at heightened risk of developing psychosis. To date, biomarker research in psychiatry and clinical neuroscience has largely concentrated on those experiencing their first psychotic episode or those at high risk of psychosis. Yet, comparisons limited to the extremes of the disorder may cloud our understanding of how psychosis develops in a small fraction of the general population.
This project aims to leverage machine learning to distinguish between at-risk and not-at-risk psychosis categories in healthy individuals based on brain alterations and cognitive scores. We intend to explore how these patterns evolve over time and their correlation with symptom changes. Utilizing the extensive dataset from the UK Biobank, which includes data from approximately half a million individuals, will enhance the robustness and predictive power of our machine learning model.
Additionally, we plan to apply this model to an adolescent cohort to assess its effectiveness in identifying young individuals at an increased risk of developing psychosis later in life.
The project is projected to span three years, during which we aim to make significant contributions to the biomarker research of schizophrenia and offer valuable neuropsychological insights for early psychosis risk prediction.