Dynamic Time Warping Analysis of Smooth Pursuit Eye Movements for Dementia Detection
Mamoru Hiroe, Yutaka Kawaguchi, Kentaro Takemura, Y. Zheng, Hisatomo Kowa, Takashi Nagamatsu · Procedia Computer Science · 2025
The increasing prevalence of Alzheimer’s disease (AD) highlights the need for efficient and early detection. However, traditional cognitive assessments, such as the Mini-Mental State Examination (MMSE), which are administered in person, may not capture subtle early stage impairments. Recent studies have suggested that eye movements can serve as a biomarker for AD. In this study, we examined the use of dynamic time warping (DTW) to analyze smooth-pursuit eye movements to detect cognitive decline in 24 elderly participants, of whom 18 had valid eye tracking data for analysis. DTW was chosen because of its ability to capture the temporal alignment between the target and gaze trajectories. DTW demonstrated a stronger correlation with the MMSE scores (R = -0.782 and R 2 = 0.612) than with the conventional Pearson correlation coefficient (R = 0.456 and R 2 = 0.208). These findings highlight the potential of DTW-based analyses to complement existing cognitive screening tools in clinical settings. Future research should expand the participant pool and explore different smooth-pursuit paradigms to improve the predictive accuracy.