Deciphering ankle dynamics: EEG BCI-robotic system to predict continuous ankle joint movements in passive, active, and imagined training

• 主要作者: Tong, Grace Min
• 其他作者: Guan Cuntai
• 格式: Final Year Project
• 語言: English
• 出版: Nanyang Technological University 2024
• 主題: Computer and Information Science; Engineering; Medicine, Health and Life Sciences; Brain computer interface; Rehabilitation device; Robotic rehabilitation system; BCI-robotic rehabilitation system; Lower limb extremities rehabilitation; Ankle joint rehabilitation; Dorsiflexion and plantarflexion
• 在線閱讀: https://hdl.handle.net/10356/175097

Background: The rehabilitation of ankle-joint movements, specifically dorsiflexion and plantarflexion, is crucial for individuals suffering from motor disabilities and neural limitations due to ageing, strokes, and locked-in syndrome (LIS). These conditions often result in foot drop, significantly impairing individuals’ daily activities and quality of life. The integration of Electroencephalogram (EEG) with robotic devices presents a novel approach to enhancing rehabilitation training, leveraging the advancements in BCI and robotics. Objective: This study aims to design and develop an Integrated EEG-Robotic Lower Limb Rehabilitation System for ankle-joint dorsiflexion and plantarflexion movement. Thereafter, our primary goal is decoding and interpreting EEG signals to differentiate between active, passive, and imaginative rehabilitation modes, predicting continuous joint angle movements leading towards a close-looped multimodal BCI-Robotic system. Methods: I designed a multi-session multimodal sensory experiment and collected data from fifteen healthy subjects with each session comprising four phases: Proprioceptive Testing, Passive Movement, Active Movement, and Imaginative Movement for continuous ankle dorsiflexion and plantarflexion. Participants were also evaluated on their Motor Imagery skills to determine how such personal and learnable skills affected the performance of EEG decoding of continuous movements in lower limb rehabilitation exercises. Results: Preliminary results suggest a significant correlation between participants' motor imagery skills and their performance in the imaginative movement phase, with higher skills associated with less fatigue. The “Active Mode” emerged as the most preferred rehabilitation mode, indicating higher cognitive engagement. Our findings highlight the necessity of active rehabilitation modes and demonstrate how EEG neurorehabilitation training enhances neuroplasticity. Conclusion: The integration of BCI and robotic technologies improves ankle-joint rehabilitation training, not only making it more engaging for users but also leading to better training outcomes for faster motor recovery. I believe this study demonstrates how integrated motor imagery detection for EEG-controlled movement of the ankle robot and real-time feedback mechanisms is capable of improving overall rehabilitation training efficiency and effectiveness, laying the groundwork for future efforts in these areas.