想象多力度单侧手运动的脑电信号分类研究

Research on Single Hand Imagination of Different Strength EEG Classification

目的传统脑机接口实验范式多为左右手运动想象,无力度分级,命令单一,为增加脑机接口命令数,使中风患者在康复期间设计获得更好的治疗方案,设计了想象三种力度下的单侧手运动实验并对其进行分类。方法 9名受试者被要求想分别以三种力度(50%、30%、10%最大自主收缩力)握紧单侧手,同时记录脑电及肌电信号,对脑电信号预处理后进行空间滤波和特征提取,再对处理后的数据进行带通滤波并提取特征,利用线性判别分析作为分类器。结果\结论采用两级特征提取分类方法,平均分类正确率达到72.4%,证明通过分析想象不同力度单侧手运动的脑电信号能够扩展脑机接口命令数。

Aim The traditional brain-computer interface experimental paradigm is mostly for left and right hand movement imagination, no strength distinction, Therefore the order is single, In order to increase the number of brain-computer interface commands, so that stroke patients can get better treatment plans during rehabilitation. a single-handed hand movement experiment with three strengths was designed and classified. Method Nine subjects were asked to imagine clenching their right hands with three different force loads(50% maximum voluntary contraction(MVC), 30%MVC and 10% MVC) and recorded their EEG and EMG signals. After preprocessing the EEG signal, spatial filtering and feature extraction are performed, then the processed data is bandpass filtered and features are extracted, and LDA is used as a classifier. Results\Conclusion Using the two-level feature extraction classification method, the average classification accuracy rate reached 72.4%, which proves that the number of brain-computer interface commands can be extended by analyzing the EEG signals of unilateral hand movements with different strengths.