A Multidimensional Visible Evaluation Model for Stroke Rehabilitation: A Pilot Study

Ping Xie, Ying Wang, Xiaoling Chen, Yingying Hao, Haoxiang Yang, Yinan Yang, Meng Xu

Abstract

Efficient rehabilitation state evaluation is important to the design of rehabilitation strategies after stroke. However, most traditional evaluations have depended on subjective clinical scales, which do not entail quantitative evaluation of the motor function. Functional corticomuscular coupling (FCMC) can be used to quantitatively describe the rehabilitation state. However, how to apply FCMC to clinical evaluation still needs to be studied. In this study, we propose a visible evaluation model which can combine the FCMC indicators with a Ueda score to comprehensively evaluate the motor function. In this model, we first calculated the FCMC indicators based on our previous study, including transfer spectral entropy (TSE), wavelet package transfer entropy (WPTE) and multiscale transfer entropy (MSTE). We then apply Pearson correlation analysis to determine which FCMC indicators are significantly correlated with the Ueda score. Then, we simultaneously introduced a radar map to present the selected FCMC indicators and the Ueda score, and described the relation between them. Finally, we calculated the comprehensive evaluation function (CEF) of the radar map and applied it as a comprehensive score of the rehabilitation state. To verify the model’s effectiveness, we synchronously collected the electroencephalogram (EEG) and electrocardiogram (EMG) data from stroke patients under the steady-state force task and evaluated the state by the model. This model visualized the evaluation results by constructing a radar map and presented the physiological electrical signal features and the clinical scales at the same time. The CEF indicator calculated from this model was significantly correlated with the Ueda score (P= 0.001<0.01 ). This research provides a new approach to evaluation and rehabilitation training after stroke, and explicates possible pathomechanisms.

摘要

高效的康复状态评估对中风后康复策略的设计非常重要。然而，大多数传统评估都依赖于主观临床量表，无法对运动功能进行定量评估。功能性皮质肌肉耦合（FCMC）可用于定量描述康复状态。然而，如何将 FCMC 应用于临床评估仍有待研究。在本研究中，我们提出了一种可视化评估模型，该模型可将 FCMC 指标与上田评分相结合，对运动功能进行综合评估。在该模型中，我们首先根据之前的研究计算了 FCMC 指标，包括传递谱熵 (TSE)、小波包传递熵 (WPTE) 和多尺度传递熵 (MSTE)。然后，我们采用皮尔逊相关分析法确定哪些 FCMC 指标与上田得分有显著相关性。然后，我们同时引入雷达图来呈现所选的 FCMC 指标和上田评分，并描述它们之间的关系。最后，我们计算了雷达图的综合评价函数（CEF），并将其作为康复状态的综合评分。为了验证模型的有效性，我们同步采集了脑电图（EEG）和心电图（EMG）数据，并通过模型对脑卒中患者在稳态用力任务下的状态进行了评估。该模型通过构建雷达图将评估结果可视化，并同时呈现生理电信号特征和临床量表。该模型计算出的 CEF 指标与上田评分有显著相关性（P= 0.001<0.01 ）。这项研究为脑卒中后的评估和康复训练提供了一种新方法，并阐述了可能的病理机制。