Integrating Eye Tracking and Inertial Sensing for Enhanced Freezing of Gait Detection in Parkinson's Disease

Freezing of gait (FOG), a disabling symptom in Parkinson's disease, presents a major challenge for wearable classification algorithms that struggle to distinguish freezes from voluntary stops. To address this ambiguity, we evaluated whether incorporating eye-gaze kinematics could improve classification accuracy compared to using ankle-mounted inertial measurement units (IMUs) alone. We analyzed data from 10 participants performing standardized walking tasks and compared two deep learning classifiers differing only in their inputs: an IMU-only model (bilateral ankle accelerometer and gyroscope) and an IMU + Gaze model that improved macro-averaged F1 from 0.657 (95% bootstrap CI: 0.461-0.756) to 0.757 (0.591-0.832; Δ = 0.099, bootstrap p = 0.016). Class level improvements were largest for standing (F1: 0.600 vs. 0.356; Δ = 0.244, p = 0.019, Holm-corrected p = 0.056), driven by recall increasing from 36.4% to 81.8%, and standing windows misclassified as freezing reduced from 59.1% (13/22) to 13.6% (3/22). These findings show that gaze kinematics complement ankle kinematics for disambiguating voluntary stopping from FOG and potentially strengthen automated monitoring, clinician-facing assessment, and patient-facing assistive technologies.