ENHANCING EARLY DETECTION AND INTERVENTION IN MENTAL HEALTH DISORDERS THROUGH MULTI-MODAL AI TECHNIQUES

This study presents a novel multimodal AI framework for early detection of mental health disorders using speech and text analysis. The system employs a transformer-based text encoder, and a hybrid convolutional-recurrent speech encoder integrated via an attention-based fusion mechanism. This dynamic fusion approach in diagnostic accuracy outperforms traditional uni-modal and static fusion methods. Experiments on benchmark datasets yielded an accuracy of 89.5%, precision of 88.0%, recall of 90.2%, F1-score of 89.1%, and AUC of 0.93. Additionally, the framework supports proactive intervention by providing real-time clinical recommendations and enhancing treatment outcomes. The proposed method offers significant potential for improving early mental health diagnostics and intervention, contributing to the shift towards AI-enhanced mental health care.