Application of Deep Learning in Mental Disorder: Challenges and Opportunities

Abstract

In this chapter, application of deep learning techniques in mental disorder is proposed to enable medical practitioners to detect the disease accurately and help them in efficient decision making subsequently. Mental illness is a type of health condition that changes a person’s mind, emotions or behaviour (or all three), which witnessed its impact on an individual’s physical health. The idea behind using deep learning lies in its potential in classifying medical data, detecting objects, recognizing speech, translating languages and knowledge discovery processes. An extensive review is conducted based on state-of-the-art machine learning approaches considering the trends, gap and challenges in mental health disorder scenarios and then proposed several deep learning methods including convolution neural network (CNN) and autoencoder; for building an effective and efficient model to address the potential pitfalls. Experimental studies are conducted using publicly available mental disorder dataset with various performance metrics such as accuracy, complexity and error rate. The obtained results and discussions will pave the way for the researchers and practitioners to understand the suitability of deep learning applications to address the mental health issues in terms of developing plans for effective mental health treatment, identifying behavioural biomarkers and predicting crises with utmost caution to name a few.