Real-Time Stress Assessment Using Sliding Window Based Convolutional Neural Network.

Area of Science:

• Computational Neuroscience
• Artificial Intelligence in Healthcare
• Psychophysiology

Background:

• Mental stress is a significant contributor to various health issues, including depression and cardiovascular diseases.
• Current stress assessment methods are often subjective, time-consuming, and lack precision.
• Existing machine learning (ML) approaches for mental state assessment require offline processing, limiting real-time application.

Purpose of the Study:

• To develop and evaluate a real-time mental stress assessment system.
• To leverage convolutional neural networks (CNNs) for accurate and immediate stress detection.
• To compare the performance of the CNN-based approach against traditional ML techniques.

Main Methods:

• Implementation of a novel approach utilizing convolutional neural networks (CNNs).
• Focus on enabling real-time processing for immediate mental stress evaluation.
• Comparative analysis with existing state-of-the-art machine learning methods.

Main Results:

• The proposed CNN-based system achieved a high accuracy of 96% for real-time mental stress assessment.
• Demonstrated superior performance with 95% sensitivity and 97% specificity.
• Outperformed other ML techniques in accuracy, time efficiency, and feature quality.

Conclusions:

• Convolutional neural networks offer a highly accurate and efficient solution for real-time mental stress assessment.
• The developed CNN approach surpasses conventional methods and existing ML techniques in performance metrics.
• This technology holds significant potential for immediate and objective mental health monitoring.