Early detection of depression through facial expression recognition and electroencephalogram-based artificial intelligence-assisted graphical user interface

• Published in: Neural computing & applications Vol. 36; no. 12; pp. 6937 - 6954
• Main Authors: Kumar, Gajendra, Das, Tanaya, Singh, Kuldeep
• Format: Journal Article
• Language: English
• Published: London Springer London 01.04.2024; Springer Nature B.V
• Subjects: Accuracy; Artificial Intelligence; Biomarkers; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Science; Data Mining and Knowledge Discovery; Diagnostic software; Electroencephalography; Emotion recognition; Emotions; Face recognition; Graphical user interface; Image Processing and Computer Vision; Mental depression; Mental disorders; Original Article; Probability and Statistics in Computer Science; Sensitivity; Signs and symptoms; User interface; Emotion recognition; Depression; Artificial intelligence; EEG; Graphical user interface
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-024-09437-z

Psychological disorders have increased globally at an alarming rate. Among these disorders, depression stands out as one of the leading and most prevalent conditions that have affected more than 280 million people. However, it remains widely undiagnosed and untreated due to lack of sensitive and reliable diagnostic tools. This underscores the imperative for the development of a sensitive and accurate diagnostic tool facilitating the early diagnosis of depression symptoms to mitigate the impending mental illness epidemic. To address this need, we developed an artificial intelligence (AI)-assisted tool utilizing facial expression-based emotion recognition and electroencephalogram (EEG) analysis for the detection of depression symptoms along with their severity level assessment. Our approach yielded successful detection of depression symptoms with an accuracy of 93.58%, a sensitivity of 92.70%, a specificity of 93.40%, and an f1-score of 93.68% through facial emotion recognition task. Additionally, severity level detection employing EEG biomarkers achieved an accuracy of 99.75%, a sensitivity of 99.75%, a specificity of 99.92%, and an f1-score of 99.75%. Consequently, a graphical user interface (GUI) tool was developed that seamlessly integrated the AI with facial image and EEG data inputs, enabling efficient recognition of depression from both real-time and pre-recorded data. The resulting AI assistant demonstrates high sensitivity, precision, and accuracy in the early detection of depression, establishing its potential as a reliable diagnostic tool. The application of our tool may be extended to clinicians, therapists, and hospitals for the identification of depression at its early stage.