Artificial Intelligence and Wearable Technologies for Upper Limb Neurorehabilitation
Summary
Wearable neural interfaces (NIs) show promise for upper limb neurorehabilitation by decoding motor intentions with AI. This review of 51 studies highlights challenges and discusses future AI advancements for personalized, interpretable systems.
Area of Science:
- Biomedical Engineering
- Neuroscience
- Rehabilitation Technology
Background:
- Non-invasive neural interfaces (NIs) utilize biosignals like electroencephalography (EEG) and electromyography (EMG) for upper limb neurorehabilitation.
- Traditional NIs are complex, limiting clinical use; wearable devices offer comfort and long-term monitoring potential.
- Existing AI methods need adaptation for wearable constraints, with a lack of clear state-of-the-art summarization.
Purpose of the Study:
- To systematically review the state-of-the-art of AI-driven non-invasive neural interfaces for upper limb neurorehabilitation using wearable devices.
- To analyze current literature based on biosignals, wearable technology, AI methods, rehabilitation focus, and clinical applications.
- To identify challenges and discuss the potential of advanced AI, including explainable AI (XAI) and generative AI (GenAI).
Main Methods:
- A systematic literature review encompassing 51 studies.
- Analysis focused on key concepts: biosignals (EEG, EMG), wearable sensors, AI algorithms, upper limb rehabilitation, and clinical applications.
- Evaluation of methodological heterogeneity, sensor configurations, and performance metrics.
Main Results:
- Significant methodological heterogeneity and diverse wearable sensor configurations were identified across studies.
- Key challenges include achieving accuracy, robustness, and thorough clinical validation for wearable NI systems.
- The review identified a need for clearer summarization of current AI approaches for wearable neurorehabilitation.
Conclusions:
- Wearable NIs present a promising avenue for upper limb neurorehabilitation, but require further development to overcome current limitations.
- Explainable AI (XAI) and Generative AI (GenAI) hold potential for enhancing interpretability and personalization in future neurorehabilitation systems.
- Addressing accuracy, robustness, and clinical validation is crucial for the widespread adoption of these technologies.
