定量评估
计算机科学
功能(生物学)
上肢
运动功能
人工智能
定量分析(化学)
人机交互
上下界
物理医学与康复
计算机视觉
生物力学
电动机控制
任务(项目管理)
作者
Yingying Hao,Yuanyuan Zhang,Jian Zhang,Jieming Hao,Ying Wang,Wenhao Hu,Xiaoling Chen,Ping Xie
标识
DOI:10.1109/jsen.2026.3742189
摘要
Motor dysfunction after stroke involves complex alterations in the interaction between the brain and peripheral motor system. Objective and quantitative assessment of motor function is important for evaluating rehabilitation outcomes and understanding neural mechanisms of motor recovery. This study proposes a multimodal information interaction framework (MIIF) for quantitative assessment of upper limb motor function after stroke. The proposed MIIF provides a flexible framework for characterizing brain and peripheral motor system interactions through multimodal physiological information and graph theoretical approaches. As a representative implementation of MIIF, corticomuscular coupling derived from electroencephalography (EEG) and electromyography (EMG) signals was used to characterize brain-muscle information interaction, and an XGBoost regression model was applied for motor function prediction. Specifically, multi-channel EEG and EMG signals were collected from stroke patients during multi-joint compound movements. A coherence network was constructed based on corticomuscular coupling, and network features were extracted to quantify alterations in motor control networks. The results showed that the clustering coefficient (C) of muscle nodes on the affected side was significantly higher than that on the unaffected side, whereas the betweenness centrality (BC) of network nodes was significantly reduced. Analysis of the affected side revealed significantly stronger information exchange among cortical regions than among muscle groups. The proposed MIIF achieved a mean absolute error of 5.024 in predicting upper limb motor function, with alpha and gamma band network features contributing more to the assessment. The proposed framework provides a flexible approach for analyzing brain-muscle information interaction and offers a potential tool for objective evaluation of upper limb motor function after stroke.
科研通智能强力驱动
Strongly Powered by AbleSci AI