New Insights from Old Data: Multimodal Classification of Schizophrenia using Automated Deep Learning Configurations

Abstract

AbstractSchizophrenia is a heterogeneous cognitive disorder where clinical classification is challenging because of the lack of well-established, non-invasive diagnoses biomarkers. There is, hence, a need for objective systems that can classify Schizophrenia despite challenges such as overlapping symptomatic factors, diverse internal clinical manifestations, and complex diagnostic process leading to delayed treatment. Thus, experimentation with automated machine learning architectural frameworks (AutoML) is presented in order to handle multimodal Functional Network Connectivity(FNC) and Source Based Morphometry(SBM) features based on functional magnetic resonance imaging(fMRI) and structural magnetic resonance imaging(sMRI) components respectively. On evaluating the resultant AutoML models with respect to approximately 280 machine learning architectures on the Overall AUC metric, the former outperforms the latter despite remarkable limitations including complex high dimensional feature space with very little data.