A Methodological Review of Artificial Intelligence Technology Application in Mental Health Disorders

Abstract

Artificial Intelligence (AI) has increasingly gained attention for its potential in diagnosing, treating, and managing mental health disorders, addressing critical gaps in traditional mental healthcare. This study explores the application of AI technologies, models, and datasets to mental health disorders, including autism spectrum disorder (ASD), schizophrenia, depression, anxiety, bipolar disorder, and PTSD. A systematic review of 40 peer-reviewed studies published between 2014 and 2024 was conducted, sourced from PubMed and Google Scholar. The review examines AI techniques such as machine learning (ML), deep learning (DL), and natural language processing (NLP), used in building AI models with diverse data types, including neuroimaging and body signals. Popular models observed include Support Vector Machine (SVM), Convolutional Neural Network (CNN), and Ensembling models. These models have various applications in developing technologies and phenomena such as wearable sensors and digital phenotyping, intending to improve mental healthcare. One of the potentials of AI is in overcoming traditional mental health challenges, including limited access and high costs. Although our findings indicate significant advancements in AI's diagnostic accuracy and personalised intervention potential. However, limitations persist, raising questions about AI’s reliability. This review identifies challenges and limitations with present-day AI technologies. Some of these are model performance, and barriers to widespread implementation, such as privacy concerns, algorithmic bias, and the need for diverse datasets. Recommendations for future research are provided, emphasising the need for larger, more inclusive datasets, research diversity, and the integration of AI into real-world clinical settings to optimise patient outcomes.