Artificial intelligence-enabled predictive modelling in psychiatry: overview of machine learning applications in mental health research

Abstract

Machine learning, an artificial intelligence (AI) approach, provides scope for developing predictive modelling in mental health. The ability of machine learning algorithms to analyse vast amounts of data and make predictions about the onset or course of mental health problems makes this approach a valuable tool in mental health research of the future. The right use of this approach could improve personalisation and precision of medical and non-medical treatment approaches. However, ensuring the availability of large, good-quality data-sets that represent the diversity of the population, along with the need for openness and transparency of the AI approaches, are some of the challenges that need to be overcome. This article provides an overview of current machine learning applications in mental health research, synthesising literature identified through targeted searches of key databases and expert knowledge to examine research developments and emerging applications of AI-enabled predictive modelling in psychiatry. The article appraises both the potential applications and current challenges of AI-based predictive modelling in psychiatric practice and research.