Automated Detection of Mental Stress Using Multimodal Characterization of PPG Signal for AI Based Healthcare Applications

Purpose-The increasing complexity of our society has made mental stress a part of every human life. The early evaluation of long-term mental stress conditions is necessary since they may trigger a number of chronic disorders. Existing electroencephalogram (EEG) based methods for estimating mental stress are often complex, multi-channel, and expert-dependent. On the other hand, the respiratory signal offers intriguing information about stress, but its capture is challenging and requires multimodal support. Method-To overcome this challenge, the respiratory signal can be extracted from the Photoplethysmogram (PPG) signal. The easy-to-acquire PPG signal is multimodally characterised in this proposed approach to determine the stressed status. Exclusively, the created algorithm leverages a major PPG characteristic and, using streamlined approaches, additionally extracts the respiratory rate from the same PPG signal. The method is assessed using PPG records gathered from the publicly available DEAP dataset. Result-A simple threshold-based along with standard classification techniques are used to assess the effectiveness of the proposed algorithm, which classify the stressed and relaxed states with an average accuracy of 98.43%. The suggested approach outperforms the existing methods in terms of performance, and its simple methodology and low acquisition load support its use in real-time standalone, personal healthcare applications. Conclusion-The use of simple features and classification strategy ensures the applicability of the proposed method in artificial intelligence (AI) based healthcare applications. This will upgrade the existing system for monitoring and proper diagnosis of patients suffering from mental stress conditions.