Cognitive Load Measurement with Pen Orientation and Pressure

Cognitive load measurement during handwriting has broad prospects both for computer-assisted education and for user interface adaptation, but it is difficult to achieve due to the lack of established methods for extracting behavioral information from writing that is indicative of cognitive load variations. This study aimed to measure cognitive load with the direct information available from the writing samples with a digital pen, using the orientation of the pen and the pressure of the pen-tip. This information was modeled with Gaussian Mixture Models (GMM), and cognitive load levels were classified on a handwriting dataset including 600 handwritten sentences. The overall classification accuracy showed that the pen orientation and pressure reflected cognitive load variation well, and the significant improvement in cognitive load classification from 52.8% to 75.4% validated the effectiveness of sample selection using altitude. ACM Classification: H5.2 [Information interfaces and presentation]: Evaluation/methodology