Neural Signature Drift: Can AI Detect Early Personality Disorder Onset?

Signature Drift in Neuronal Networks involves subtle but gradual changes in brain network activity prior to the emergence of symptoms for personality disorder diagnosis. The project evaluates the potential of sophisticated AI algorithms to identify preclinical neural anomalies based on neuroimaging and electrophysiology recordings longitudinally. Based on neurobiology, the project aims to explore neuronal anomalies in larger brain networks, especially in the fronto-limbic circuits, default mode network functions, and salience processing system. The particular focus will be on Borderline Personality Disorder owing to its link with affective dysregulation and self-disruption. By leveraging deep learning frameworks and temporal pattern recognition techniques, the research uses multiple data modalities such as fMRI and EEG data to model latent neural dynamics over time. Such techniques attempt to detect departures from expected developmental trajectories of cognitive and affective processes in order to extract neurobiological biomarkers for psychopathological outcomes. Through a combination of behavioral and neural measurements, the research utilizes the methodology of computational psychiatry to make the link between biological basis and psychopathology explicit. Among the anticipated results are the development of biomarkers that could accurately identify individuals who are vulnerable to personality disorders before they develop any symptoms. These have very important implications in the context of prevention, and will allow better identification of the risk factors associated with such disorders. From a scientific perspective, this project will contribute to the progress in the field of precision psychiatry through a shift away from post hoc diagnosis and toward predictive modeling.