Title:

PACE: Pupillometric Analysis of Cognitive Effort

The PACE (Pupillometric Analysis of Cognitive Effort) project investigates whether pupil dilation can serve as a reliable indicator of cognitive load during standardized neuropsychological testing. Current cognitive-load measurement systems rely on expensive, specialized eye-tracking hardware, limiting accessibility, and long-term deployment. During Semester One, our team developed a proof-of-concept pipeline using Tobii Pro Glasses 3 and PEBL cognitive assessments (N-Back and BCST) to collect, process, and analyze pupil diameter data. We developed a reproducible end-to-end Python pipeline to import, clean, and analyze Tobii pupil data, addressing issues such as blinks and invalid samples while generating interpretable visualizations. Controlled testing sessions followed a standardized procedure, with results showing consistent increases in IPA and pupil variability during higher cognitive-load tasks. A preliminary Random Forest model achieved a test R² of 0.64, supporting the feasibility of estimating cognitive load from engineered pupil-based features. Semester Two focused on developing a method for estimating cognitive-load using webcam-based capture scripts and open-source computer vision tools, enabling real-time pupil measurement with low-cost hardware. We expanded our dataset, refining feature engineering, and improved predictive modeling performance. These efforts resulted in a more accessible, scalable approach to cognitive-load monitoring, demonstrating the feasibility of replacing specialized eye-tracking hardware with webcam technology.

Name

Comment*

Comments are viewable only by submitter

Captcha*