Objective: Demonstrate the concurrent validity, diagnostic accuracy, and early detection ability of an iPad app-based eye tracking tool for the early detection of Parkinson’s disease (PD).
Background: Despite advances in understanding the pathogenesis and treatment of PD, early detection remains a major challenge due to the lack of definitive early detection tools1. Currently, proposed methods such as neuroimaging, skin biopsies, and spinal fluid assays are costly, invasive, and impractical for large-scale screening. A promising alternative is the quantitative assessment of eye movements, as the neural circuits controlling them are linked to nearly every part of the brain and are highly sensitive to early, localized, degenerative changes2-5. Eye movement tracking using a widely available platform, the iPad, could offer a practical and easily scalable solution6.
Method: Twenty-five participants (10 with PD and 15 healthy controls) completed a randomized sequence of one reflexive (prosaccade) and two volitional (antisaccade and memory-guided saccade) tasks7. Forty trials per task were obtained concurrently on the iPad and EyeLink 1000 Plus ®, the gold-standard for video-based eye tracking. The iPad displayed visual stimuli on its screen and recorded participants via its front-facing camera. The EyeLink simultaneously tracked the same eye movements. Pulses sent every 16ms from the iPad synchronized the two systems. Offline, a deep learning model estimated gaze from the iPad recordings. We then fit a curve to the estimated gaze to increase data resolution and reliably determine saccade metrics of amplitude, accuracy, and gain. Standard processing methods were used to extract the same data from the EyeLink8-10.
Results: We found excellent concurrent validity (Cronbach alpha >0.95) and diagnostic accuracy of the iPad (>95% specificity and sensitivity). In addition, in the HC group we were able to identify trials that were closer to PD than HC performance, potentially classifying them as ‘high-risk’ trials.
Conclusion: By demonstrating validity, diagnostic accuracy, and possible early detection capability of our tool, we provide an easily scalable solution to track eye movement with a potential application for early PD detection.
Results
Subject Demographics and MDS-UPDRS Scores
References: 1. Tolosa E, Garrido A, Scholz SW, Poewe W. Challenges in the diagnosis of Parkinson’s disease. Lancet Neurol 2021;20(5):385-397.
2. Antoniades CA, Spering M. Eye movements in Parkinson’s disease: from neurophysiological mechanisms to diagnostic tools. Trends in Neurosciences 2024;47(1):71-83.
3. Przybyszewski AW, Sledzianowski A, Chudzik A, Szlufik S, Koziorowski D. Machine Learning and Eye Movements Give Insights into Neurodegenerative Disease Mechanisms. Sensors 2023;23(4):21.
4. Terao Y, Fukuda H, Hikosaka O. What do eye movements tell us about patients with neurological disorders? – An introduction to saccade recording in the clinical setting. Proc Jpn Acad Ser B Phys Biol Sci 2017;93(10):772-801.
5. Vidailhet M. Eyes as a window to brain pathology in parkinson’s disease: a narrative review. J Neural Transm (Vienna) 2024.
6. Lai HY, Saavedra-Pena G, Sodini C, Heldt T, Sze V. App-Based Saccade Latency and Directional Error Determination Across the Adult Age Spectrum. Ieee Transactions on Biomedical Engineering 2022;69(2):1029-1039.
7. Terao Y. Making saccades, fast and slow: The voluntary versus reflexive saccade systems in Parkinson’s disease. Clinical neurophysiology : official journal of the International Federation of Clinical Neurophysiology 2022;143:143-144.
8. Munoz MJ, Reilly JL, Pal GD, et al. Benefits of subthalamic nucleus deep brain stimulation on visually-guided saccades depend on stimulation side and classic paradigm in Parkinson’s disease. Clinical Neurophysiology 2024;162:41-52.
9. Munoz MJ, Goelz LC, Pal GD, et al. Increased subthalamic nucleus deep brain stimulation amplitude impairs inhibitory control of eye movements in Parkinson’s disease. Neuromodulation 2022;25(6):866-876.
10. Munoz MJ, Reilly JL, Pal GD, et al. Medication adversely impacts visually-guided eye movements in Parkinson’s disease. Clin Neurophysiol 2022;143:145-153.
To cite this abstract in AMA style:
F. David, J. Koerner, E. Zou, J. Karl, C. Poon, C. Sodini, V. Sze, L. Verhagen-Metman, T. Heldt. iPad App-based Eye Tracking for the Early Detection of Parkinson’s Disease [abstract]. Mov Disord. 2025; 40 (suppl 1). https://www.mdsabstracts.org/abstract/ipad-app-based-eye-tracking-for-the-early-detection-of-parkinsons-disease/. Accessed October 5, 2025.« Back to 2025 International Congress
MDS Abstracts - https://www.mdsabstracts.org/abstract/ipad-app-based-eye-tracking-for-the-early-detection-of-parkinsons-disease/