Computational multi-omics integration in neurocognitive disorders from machine learning to graph-based approaches

Abstract

Neurocognitive disorders (NCDs) pose a major global health challenge. This review summarizes computational methods for multi-omics data integration in NCD research, including conventional machine learning, deep learning, and graph-based approaches. We highlight their strengths, limitations, and suitability for key objectives such as biomarker discovery and patient classification. Overall, this study highlights considerations that can support researchers in selecting suitable methods for multi-omics integration in NCD studies.