Mapping the Disordered Mind: A Computational Framework for Integrating Neuroimaging and Symptom Data

Abstract

The diagnostic categorization in psychiatry, largely based on clinically observed symptom clusters, has proven insufficient for capturing the vast biological and phenomenological heterogeneity within and across psychiatric disorders. This heterogeneity is a primary obstacle to developing targeted, effective treatments. We propose a novel computational framework designed to map the complex, non-linear relationships between distributed neural circuit dysfunctions and the multidimensional space of psychopathology. This framework, which we term the Integrative Neuroclinical Mapping Framework (INMF), moves beyond case-control comparisons to model psychopathology as a system of continuous, overlapping dimensions. The INMF processes high-dimensional data from functional and structural magnetic resonance imaging (fMRI, sMRI), integrating them with fine-grained, dynamic symptom data acquired through digital phenotyping (e.g., ecological momentary assessment, smartphone sensors). Core to the INMF is a multi-stage analytical pipeline featuring (1) data harmonization and feature extraction using automated preprocessing and source-based morphometry/independent component analysis, (2) manifold learning to uncover low-dimensional latent neural-symptom structures, and (3) graph-based network analysis to model the dynamic interplay between brain networks and symptom domains. We present a proof-of-concept application using a simulated dataset of individuals with schizophrenia, major depressive disorder, and healthy controls, demonstrating the framework's ability to identify transdiagnostic biotypes that are more homogenous than traditional diagnoses. Our results illustrate how the INMF can delineate distinct neural pathways leading to similar symptomatic expressions (equifinality) and common neural risk factors manifesting as different disorders (multifinality). This framework offers a powerful, data-driven approach for parsing the nosological chaos of psychiatry, with the potential to revolutionize diagnosis, prognostication, and the development of personalized neurotherapeutics.