Preliminary Results for Measurement of Eye Fixation and Speech to Identify Cognitive Decline
Amir Laghai, Bruce Wallace, Kathleen Fraser, Manuela Kunz, Rafik A Goubran, Frank D. Knoefel, Sina Shafiyan, Neil William Douglas Thomas · 2025
Early diagnosis of dementia is essential for timely intervention and care. This study explores a novel approach to dementia assessment by integrating eye-tracking data and speech analysis during the Cookie Theft test. By defining regions of interest (ROIs) within the Cookie Theft picture, we introduce new features and visualizations that provide deeper insights into gaze behavior and verbal identification patterns. Timestamps of fixations within defined ROIs provide insights into how participants visually explore the image. Spoken words were mapped to corresponding ROIs, serving as verbal identifications. A gaze video visualization was developed, highlighting ROI fixations and verbal identifications as they occur. Additionally, a timeline plot was created to illustrate how participants explore the image, verbally identify regions over time, and how these interactions relate to each other. This preliminary study involved two groups of participants: six healthy controls and six individuals with cognitive decline. Although aggregate fixation and verbal metrics did not reach statistical significance, region-specific analyses revealed some significant differences between groups. Results indicate that healthy controls exhibited longer average and maximum time differences between visual fixation on the woman and verbal identification. These results highlight the potential of combined eye-tracking and speech features in characterizing cognitive differences, which may be leveraged for classification. Future work will focus on expanding the participant pool to increase statistical power and on exploring alternative, cost-effective webcam-based eye-tracking methods to improve accessibility and scalability for dementia assessment.