Real-time Cognitive Load Measurement Using Wearable Photoplethysmogram Sensor

Measurement of the cognitive load in mobile situations using wearable devices plays important roles in smart human-computer interactions, physical health monitoring, mental health monitoring, and so forth in daily life. Among wearable devices, wristbands are similar to traditional watches and suitable for most scenarios, such as classrooms and offices; thus, they are widely accepted and becoming increasingly popular. However, most wristbands are only used for step counting and heart rate monitoring. In this paper, we propose a cognitive load monitoring method based on wristband heart rate sensors. The sensors used to detect heart rate are generally photoplethysmogram sensors, but the current major manufacturers extract only the heart rate parameter of the photoplethysmogram. We make full use of the waveform of the wristband photoplethysmogram signal to measure real-time cognitive load, and establish a real-time cognitive load measurement system, which expands the application scope of the wristband photoplethysmogram sensor and shows great potential for measuring cognitive load in daily life. In experimental validation using an n-back task, our waveform-based CNN model achieved approximately 20% higher accuracy in cognitive load classification versus traditional feature-based methods (e.g., LR, SVM, and GNB), reaching an average of 80.11%. The system successfully classifies the cognitive load and demonstrates the practical viability of the proposed approach.