Virtual reality (VR) is one of the versatile technologies which extends reality into the virtual domain. It offers an enormous virtual environment that is partially or completely independent from the real world. The objective of VR is to create an environment that will simulate or build the digital forms of the real world. For this feature of simulation, VR technology is being rapidly adopted in the healthcare industry for different purposes, i.e., training, treatment, disease awareness, etc., where many of its applications are in the testing phase. Depression is one of the psychological health issues that has become a very common problem worldwide. Different methods like antidepressants and psychotherapy can be used to treat this problem. Some VR technologies have already been adopted for the treatment of depression. In this chapter, the authors are going to discuss the significant applications and approaches of VR towards depression.
TopRazavi, R. et. al. in their work have taken mobile phones to screen the depressed people. They have reported 412 participants from their mobile usage statistics. The data was collected from the Amazon Mechanical Turk portal. Researchers used Beck Depression Inventory-2nd ed (BDI-II) to measure the depression severity. Also, different ML algorithms i.e., K-Nearest Neighbors, Logistic Regression, Neural Networks, Random Forests (RF), Support Vector Machines-Linear Kernel and Radial Basis Function Kernel were trained to detect the patients with depression symptoms. RF gave the best accuracy of 0.768, which was increased to 0.811 based on the gender and age of the patients.
Ansari, L. et. al. in their research took social media based mental health assessment to classify depression. The main objective of this study was to detect depression in its early stage to prevent more harm. Natural Language Processing (NLP), ML and Deep Learning (DL) models were taken into consideration for this paper. Long short-term memory (LSTM) gave the highest accuracy for this study.
Rosa, R. L. et. al. presented a Knowledge-Based Recommendation System (KBRS), which included an emotional health monitoring system. This system mainly detected the users with potential psychological disturbances i.e., stress and depression. The detection of sentences with stressful and depressive contents were performed through Convolutional Neural Network (CNN), Recurrent Neural Network (RNN) and Bi-Directional Long-Short Term Memory (BLSTM). RNN gave the highest accuracy of 94% for this study.
Zhou, X. et. al. worked on facial depression recognition. For this study, label distribution learning and deep metric learning techniques were proposed. Two depression datasets were used for this paper. A metric learning method (DJ-LDML) was proposed to address different depression mental issues. They selected the depression score with level (14-19) mild, (20-28) moderate, and (29-63) severe for this study.
Deshpande, M. et. al. detected the depression using Feeling of worthlessness or hopelessness, Worsened ability to think and concentrate, Depressed Mood, Loss on interest in activities, Suicidal thoughts. Naive-Bayes and Support vector machine classifiers were used. NB gave the best accuracy with 83%.