Clinical Decision Support System for Automatic Detection of Subjective Pain Intensity from Gaze Data

Poster Number: P62
Presentation Time: 05:00 PM - 06:30 PM

Abstract Keywords: Human-computer Interaction, Clinical Decision Support, Biomarkers, Machine Learning, User-centered Design Methods
Working Group: Clinical Decision Support Working Group
Primary Track: Applications
Programmatic Theme: Clinical Informatics

Pain, a significant health challenge, stands to benefit from sensor-based technologies, with eye movements showing promise as objective measures. However, traditional self-reported pain assessments lack objectivity, emphasizing the need for alternative measures. In this study, we aimed to develop a machine-learning engine to detect subjective pain intensity levels from physiological eye movements, potentially serving as a biomarker for pain. We conducted an eye-tracking experiment with 63 adults, 40 pain-free individuals, and 22 chronic pain individuals. machine learning techniques, including Random Forest and Neural Network models, we analyzed eye movement features such as fixations, visits, and saccade metrics. Our results demonstrated the ability of these features to capture nuances in attention related to pain experience, with the Random Forest model achieving an accuracy of 82% and the Neural Network model achieving 78%. These findings underscore the potential of eye movement metrics as objective measures for assessing subjective pain intensity.

Speaker(s):
Doaa Alrefaei, MS
Worcester Polytechnic Institute

Author(s):
Doaa Alrefaei, MS - Worcester Polytechnic Institute; Read Alharbi, PhD - Saudi Electronic University; Soussan Djamasbi, PhD - Worcester Polytechnic Institute; Diane Strong, PhD - Worcester PolytechnicInstitutes;