HS Stroke dataset: an upper-limb motor imagery EEG dataset of chronic stroke patients
收藏DataCite Commons2026-04-24 更新2026-04-25 收录
下载链接:
https://figshare.com/articles/dataset/HS_Stroke_dataset_an_upper-limb_motor_imagery_EEG_dataset_of_chronic_stroke_patients/30747806
下载链接
链接失效反馈官方服务:
资源简介:
The HS Stroke dataset contains motor imagery (MI) EEG recordings from 14 chronic stroke patients performing left- and right-hand MI tasks. Each participant completed 20 sessions over approximately 20 days, with 200 trials per session, yielding a total of 56,000 MI EEG trials. The dataset also includes pre- and post-training clinical and neurophysiological measures, including FMA, sEMG, and MEPs, providing complementary assessments of motor function and neural activity. All data are organized according to the EEG-BIDS standard, including experimental stimuli, EEG data and derivatives, patient information, and analysis code.With this dataset, we first examined the neurophysiological differences between left- and right-hand MI in stroke patients. Subsequently, we performed a binary classification task using baseline data and state-of-the-art methods, achieving an average classification accuracy of 82.65%. This result affirms the quality of the collected EEG signals and demonstrates their successful discriminability based on the execution of MI tasks. Exploratory analyses further showed that MI-related EEG features are significantly associated with FMA scores, indicating their relevance to motor recovery. We expect that this dataset will support future MI-BCI research and inform post-stroke rehabilitation strategies.
HS Stroke数据集收录了14名慢性脑卒中患者执行左右手运动想象(motor imagery, MI)任务时的脑电图(electroencephalogram, EEG)记录。每位受试者在约20天内完成20轮实验,每轮含200次试次,总计获取56000次运动想象脑电图试次。该数据集还包含训练前后的临床与神经生理学测量指标,包括FMA(Fugl-Meyer运动功能评分)、表面肌电图(surface electromyography, sEMG)及运动诱发电位(motor evoked potentials, MEPs),可为运动功能与神经活动提供互补评估。所有数据均按照脑电图脑成像数据结构(EEG-Brain Imaging Data Structure, EEG-BIDS)标准进行组织,涵盖实验刺激、脑电图数据及其衍生数据、受试者信息与分析代码。依托该数据集,我们首先探究了脑卒中患者左右手运动想象的神经生理学差异。随后,我们采用基线数据与最先进方法开展二分类任务,平均分类准确率达82.65%。这一结果证实了所采集脑电图信号的质量,并证明其可基于运动想象任务的执行实现有效区分。探索性分析进一步显示,与运动想象相关的脑电图特征与FMA评分存在显著关联,表明其与卒中后运动恢复具有相关性。我们期望该数据集可助力未来运动想象-脑机接口(motor imagery-Brain-Computer Interface, MI-BCI)研究,并为卒中后康复策略提供参考依据。
提供机构:
figshare
创建时间:
2025-12-03



