Real-time Assessment of Upper Extremity Motor Function in Stroke Patients Using Machine Learning
Abstract views: 12
/ 
PDF downloads: 12
Keywords:
Stroke Rehab, NIHSS, AI-Augmented, RGB Camera, Machine LearningAbstract
This paper proposes a novel AI-based system that leverages a smartphone camera to assess
upper limb motor functions in stroke survivors, following the National Institute of Health Stroke Scale
(NIHSS) guidelines. Accurate assessment of motor functions is crucial for effective stroke rehabilitation,
yet current methods often require direct involvement from healthcare professionals, who may be in short
supply even in developed countries. Designed for remote and autonomous rehabilitation, our system
enables patients or their attendants to conduct assessment exercises using only a smartphone. The AI
algorithm evaluates motor function based on NIHSS scoring criteria, helping to address the shortage of
healthcare professionals in clinical settings. The system employs Mediapipe Pose to extract 33 skeletal
features from the camera feed, arranging these in a sequential manner using the sliding window
technique, which enhances temporal analysis. These keypoints are then analyzed by machine learning
models (SVM, KNN, RF, and MLP), which were trained on publicly available stroke rehabilitation
datasets, including UI-PRMD and TRSP. The models achieved impressive average accuracies of 95.27%,
94.62%, 96.90%, and 95.94%, respectively. By providing real-time, NIHSS-aligned assessments of motor
functions, this application has the potential to reduce the burden on healthcare professionals, increase
accessibility to rehabilitation services, and ultimately improve patient outcomes during the recovery
process.
Downloads
References
T. N. e. a. Rochmah, "Economic burden of stroke disease: a systematic review," International journal of environmental research and public health, vol. 18, no. 14, p. 7552, 2021.
J. S. e. al., "Multidisciplinary care of patients with facial palsy: Treatment of 1220 patients in a German facial nerve center," Journal of Clinical Medicine, vol. 11, no. 2, p. 427, 2022.
S. J. e. al., "MOG encephalomyelitis after vaccination against severe acute respiratory syndrome coronavirus type 2 (SARS-CoV-2): Case report and comprehensive review of the literature," Journal of neurology, vol. 269, no. 10, p. 5198–5212, Oct 2022.
S. A. e. al., "Rehabilitation of upper limb motor impairment in stroke: A narrative review on the prevalence, risk factors, and economic statistics of stroke and state of the art therapies," MDPI Healthcare, vol. 10, no. 2, p. 190, Feb 2022.
R. T. e. al., "Canadian stroke best practice recommendations: Rehabilitation, recovery, and community participation following stroke. Part one: Rehabilitation and recovery following stroke; update 2019," International Journal of Stroke, vol. 15, no. 07, p. 763–788, 2020.
B. P. H. e. al., "Sensitivity to change and responsiveness of the upper extremity Fugl-Meyer assessment in individuals with moderate to severe acute stroke," Neurorehabilitation and Neural Repair , vol. 37, no. 8, p. 545–553, 2023.
S. R. e. al., "AI-driven stroke rehabilitation systems and assessment: A systematic review," IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 31, p. 192–207, 2022.
L. Y. e. al., "Identifying stroke-related quantified evidence from electronic health records in real-world studies," Artificial Intelligence in Medicine, vol. 140, p. 102552, 2023.
D. E. e. al., "Biosignal-based human–machine interfaces for assistance and rehabilitation: A survey," Sensors, vol. 21, no. 20, p. 6863, 2021.
L. G. S. R. e. al., "Supervised classification of motor-rehabilitation body movements with RGB cameras and pose tracking data," Journal on Interactive Systems, vol. 13, no. 1, pp. 221-231, 2022.
S. S. e. al., "Artificial intelligence for skeleton-based physical rehabilitation action evaluation: A systematic review," Computers in Biology and Medicine, vol. 158, p. 106835, 2023.
L. L. M.-C. S. L. Z. a. S.-C. Y. J. Cui, "An evidence-based intelligent method for upper-limb motor assessment via a VR training system on stroke rehabilitation," IEEE Access, vol. 9, p. 65871–65881, 2021.
Y. H. H. P.-C. H. S.-Y. Y. H.-T. L. Y.-C. C. a. B.-S. L. I.-J. Lee, "AI-based automatic system for assessing upper-limb spasticity of patients with stroke through voluntary movement," IEEE Journal of Biomedical and Health Informatics, vol. 28, no. 2, p. 742–752, 2022.
S. S. A. K. M. N. H. M. M. U. M. F. I. a. S. D. S. Rahman, "AI-driven stroke rehabilitation systems and assessment: A systematic review," IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 31, p. 192–207, 2022.
M.-C. S. S.-H. L. E. H.-K. W. L.-H. T. a. S.-C. Y. C.-R. Chung, "An intelligent motor assessment method utilizing a bi-lateral virtual-reality task for stroke rehabilitation on upper extremity," IEEE Journal of Translational Engineering in Health and Medicine, vol. 10, p. 1–11, 2022.
M. M. I. e. al, "Human activity recognition using tools of convolutional neural networks: A state of the art review, data sets, challenges, and future prospects," Computers in biology and medicine , vol. 149, p. 106060, 2022.
M. G. F. C. a. C. F. F. C. F. G. Aquino, "Explaining one-dimensional convolutional models in human activity recognition and biometric identification tasks," Sensors, vol. 22, no. 15, p. 5644, 2022.
H.-p. J. D. P. a. R. B. A. Vakanski, "A data set of human body movements for physical rehabilitation exercises," MDPI Data, vol. 3, no. 2, pp. 1-15, 2018.
E. D. e. al., "The Toronto Rehab Stroke Pose dataset to detect compensation during stroke rehabilitation therapy," in Proceedings of the 11th EAI international conference on pervasive computing technologies for healthcare, 2017.
W. A. a. U. M. T. Sherazi, "SPVD: A stroke patient motor movements video dataset for AI augmented rehabilitation," in Proceedings of the 20th IBCAST, Bhurban , 2023.
V. B. e. al., Blazepose: On-device real-time body pose tracking, arXiv preprint, 2020.
A. G. O. a. D. B. J. Ortiz Laguna, "A dynamic sliding window approach for activity recognition," in Proc. 19th Int. Conf. User Modeling, Adaptation, and Personalization (UMAP 2011), Girona, Spain, 2011.
P.-W. e. a. Chen, "Measuring activities of daily living in stroke patients with motion machine learning algorithms: A pilot study," International journal of environmental research and public health , vol. 18, no. 4, p. 1634, 2021.
