Hosting institute
ICube Laboratory (The Engineering science, computer science and imaging laboratory) at the University of Strasbourg is a leading research center in Computer Science, with more than 300 permanent researchers, with the recently opened AI graduate school supported by the French government. This PhD offer is provided by the ENACT AI Cluster and ITI HealthTech. Find all ENACT PhD offers and actions on https://cluster-ia-enact.ai/.
Work place
The thesis work will take place in the MLMS (Machine Learning, Modeling & Simulation) research team of the ICube laboratory (The Engineering science, computer science and imaging laboratory) of the University of Strasbourg, a leading research center with more than 300 permanent researchers. The workplace is located on the hospital site of the laboratory, a 10-minute walk from the heart of downtown Strasbourg, listed as a UNESCO World Heritage Site.
In the course of the thesis, approximately 200 patients will be recorded during their clinical visits at Hautepierre University Hospital and at the Expert Centre for Parkinson’s Disease, where Prof. Anheim conducts his consultations.
Supervisors
– co-supervisors: Dr. Hyewon Seo (ICube, Univ. Strasbourg), Prof. Mathieu Anhem (Reference Centre for Rare Neurogenetic Diseases and Inter‑Regional Expert Centre for Parkinson’s Disease)
Staring date
October 2026
Context
Neurodegenerative diseases such as Alzheimer’s disease (AD), Dementia with Lewy Bodies (DLB) and Parkinson’s disease (PD), are frequent progressive disorders of the nervous system, impacting a significant proportion of the elderly population, with prevalence increasing sharply after 65 years (30–40% of individuals over 85 years). Given that neurodegenerative disorders fundamentally disrupt cognition, movement, and behavior, the analysis of motor function has become a vital and objective means of monitoring clinical progression [1-3]. However, traditional clinical assessments remain subjective and episodic, hampering fine-grained longitudinal monitoring and limiting sensitivity to subtle motor changes outside controlled clinical settings. Recently, computational and reproducible methods have emerged for motor function analysis. While most motor assessment methods depend on wearable sensors or motion capture devices, video-based approaches are increasingly being explored, offering the potential for scalable and unobtrusive monitoring, even outside clinical settings. Our recent work on deep gait analysis in AD and DLB from videos [4-6] has shown encouraging results, suggesting the potential of computer vision as a scalable tool for clinical assessment and beyond
Objective
In this project, we will develop learning-based approaches to motion analysis from videos in the context of neurodegenerative diseases, addressing current limitations and challenges:
̶ Data scarcity: Despite the availability of some recent datasets on Parkinson’s disease [7], most data are limited to 3D motion recordings due to privacy concerns, and the research community remains highly data hungry. Throughout the project, we will collect patient data to help address current limitations in assessing and monitoring disease progression.
̶ Neuromotor analysis beyond gait: While gait is the most commonly analyzed neuromotor activity, other movements—including facial expressions, hand gestures, and upper-body motions—provide complementary insights. We aim to integrate these motion types into our analyses to achieve a more comprehensive and reliable assessment of neuromotor impairments.
̶ Explainability and interpretability in motion prediction: While current methods achieve reasonable performance, understanding the reasoning behind model predictions remains a challenge. This is crucial for detecting distinguishing closely related diseases such as PD and DLB.
̶ Longitudinal analysis: Neurodegenerative diseases develop over decades, so single-timepoint used by current state-of-the-art methods often fail to capture early indicators or accurately reflect disease stage. We will perform longitudinal analyses from repeated video recordings, enabling continuous monitoring, and precise tracking of disease progression.
Candidate profile
− Master degree in Computer Science, Electronic & Electrical Engineering, or Applied Mathematics
− Solid programming skills
− Proficiency in Deep Learning techniques
− Good communication skills
Application
Send your CV and your academic transcripts (Bachelor and Master) to seo@unistra.fr (Application deadline: 23rd April).
Bibliography
[1] W. G. Meissner, P. Remy, C. Giordana, D. Maltête, P. Derkinderen, J. L. Houéto, M. Anheim, et al., “Trial of Lixisenatide in early Parkinson’s disease,” N. Engl. J. Med., vol. 390, no. 13, pp. 1176–1185, Apr. 2024.
[2] M. Béreau et al., “Motivational and cognitive predictors of apathy after subthalamic nucleus stimulation in Parkinson’s disease,” Brain, vol. 147, no. 2, pp. 472–485, Feb. 2024.
[3] T. Wirth, J. Faber, C. Depienne, E. Roze, J. Honnorat, W. G. Meissner, P. Giunti, C. Tranchant, T. Klockgether, and M. Anheim, “Progress and challenges in sporadic late-onset cerebellar ataxias,” Nat. Rev. Neurol., vol. 21, no. 12, pp. 687–705, Dec. 2025.
[4] D. Wang, K. Yuan, C. Muller, F. Blanc, N. Padoy, and H. Seo, « Enhancing gait video analysis in neurodegenerative diseases by knowledge augmentation in vision language model, » in Proc. Med. Image Comput. Comput.-Assist. Intervent. (MICCAI), 2024, vol. 15005, pp. 251–261.
[5] D. Wang, K. Yuan, and H. Seo, « GaVA‑CLIP: Refining multimodal representations with clinical knowledge and numerical parameters for gait video analysis in neurodegenerative diseases, » IEEE J. Biomed. Health Inform., 2025.
[6] D. Wang, C. Zouaoui, J. Jang, H. Drira, and H. Seo, « Video‑based gait analysis for assessing Alzheimer’s disease and dementia with Lewy bodies« , in Proc. MICCAI Workshop Appl. Med. AI, 2023, vol. 14313.
[7] V. Adeli et al., « CARE‑PD: A multi‑site anonymized clinical dataset for Parkinson’s disease gait assessment, » in Proc. Adv. Neural Inf. Process. Syst., 2025. [8] S. T. Wasim, M. Naseer, S. H. Khan, F. Shahbaz Khan, and M. Shah, “Vita-CLIP: Video and text adaptive CLIP via multimodal prompting,” in Proc. IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR), 2023.
