Ultra-Short Window Length and Feature Importance Analysis for Cognitive Load Detection from Wearable Sensors

Human cognitive capabilities are under constant pressure in the modern information society. Cognitive load detection would be beneficial several applications of human–computer interaction, including attention management and user interface adaptation. However, current research into accurate real-time biosignal-based lacks understanding optimal minimal window length data segmentation which allow for more timely, continuous state detection. This study presents a comparative analysis ultra-short (30 s or less) lengths with wearable device. Heart rate, heart rate variability, galvanic skin response, temperature features extracted at six different used to train an Extreme Gradient Boosting classifier detect between rest. A 25 showed highest accury (67.6%), is similar earlier studies using same dataset. Overall, model accuracy tended decrease as decreased, lowest performance (60.0%) was observed 5 window. The contribution physiological classification most useful that react short windows also discussed. provides promising basis future sensors.