Longitudinal ALS EEG Dataset for Motor Imagery Studies
收藏资源简介:
This dataset comprises EEG recordings from eight ALS patients aged between 45.5 and 74 years. Patients exhibited revised ALS Functional Rating Scale (ALSFRS-R) scores ranging from 0 to 46, with time since symptom onset (TSSO) varying between 12 and 113 months. Notably, no disease progression was reported during the study period, ensuring stability in clinical conditions. The participants were recruited from the Penn State Hershey Medical Center ALS Clinic and had confirmed ALS diagnoses without significant dementia. This rigorous selection criterion ensured the validity and reliability of the dataset for motor imagery analysis in an ALS population.The EEG data were collected using 19 electrodes placed according to the international 10-20 system (FP1, FP2, F7, F3, FZ, F4, F8, T7, C3, CZ, C4, T8, P7, P3, PZ, P4, P8, O1, O2), with signals referenced to linked earlobes and a ground electrode at FPz. Additionally, three electrooculogram (EOG) electrodes were employed to facilitate artifact removal, maintaining impedance levels below 10 kΩ throughout data acquisition. The data were amplified using two g.USBamp systems (g.tec GmbH) and recorded via the BCI2000 software suite, with supplementary preprocessing in MATLAB. All experimental procedures adhered strictly to Penn State University’s IRB protocol PRAMSO40647EP, ensuring ethical compliance.Each participant underwent four brain-computer interface (BCI) sessions conducted over a period of 1 to 2 months. Each session consisted of four runs, with 10 trials per class (left hand, right hand, and rest) for a total of 40 trials per session. The sessions began with a calibration run to initialize the system, followed by feedback runs during which participants controlled a cursor's movement through motor imagery, specifically imagined grasping movements. The study design, focused on motor imagery (MI), generated a total of 160 trials per participant over two months.This dataset holds significance in studying the longitudinal dynamics of motor imagery decoding in ALS patients. To ensure reproducibility of our findings and to promote advancements in the field, we have received explicit permission from Prof. Geronimo of Penn State University to distribute this dataset in the processed format for research purposes. The original publication of this collection can be found below.<b>How to use this dataset: </b>This dataset is structured in MATLAB as a collection of subject-specific structs, where each subject is represented as a single struct. Each struct contains three fields:<code><strong>L</strong></code>: Trials corresponding to <i>Left</i> Motor Imagery.<code><strong>R</strong></code>: Trials corresponding to <i>Right</i> Motor Imagery.<code><strong>Re</strong></code>: Trials corresponding to <i>Rest</i> state.Each field contains an array of trials, where each trial is represented as a matrix with, <b>Rows </b>as Timestamps, and <b>Columns </b>as channels.<br><br>Primary Collection: Geronimo A, Simmons Z, Schiff SJ. Performance predictors of brain-computer interfaces in patients with amyotrophic lateral sclerosis. Journal of neural engineering 2016 13. 10.1088/1741-2560/13/2/026002.<br><br>All code for any publications with this data has been made publicly available at the following link:<br>https://github.com/rishannp/Auto-Adaptive-FBCSP<br>https://github.com/rishannp/Motor-Imagery---Graph-Attention-Network<br>
本数据集包含8名肌萎缩侧索硬化症(ALS)患者的脑电(EEG)记录,患者年龄介于45.5至74岁之间。患者的修订版ALS功能评定量表(ALSFRS-R)得分范围为0至46,症状发作后时间(TSSO)跨度为12至113个月。值得注意的是,研究期间未报告疾病进展,确保了临床状态的稳定性。 受试者均招募自宾夕法尼亚州立大学赫尔希医疗中心ALS门诊,且经确诊为ALS,无显著痴呆症状。这一严格的筛选标准保障了本数据集用于ALS人群运动想象分析的有效性与可靠性。 脑电数据采用符合国际10-20系统的19个电极采集(FP1、FP2、F7、F3、FZ、F4、F8、T7、C3、CZ、C4、T8、P7、P3、PZ、P4、P8、O1、O2),信号以双侧乳突作为参考电极,接地电极置于FPz。此外,还使用了3个眼电(EOG)电极以辅助伪迹去除,在整个数据采集过程中保持阻抗低于10千欧(kΩ)。 数据通过两台g.USBamp系统(g.tec GmbH公司)进行放大,并借助BCI2000软件套件进行记录,后续在MATLAB中完成补充预处理。所有实验流程严格遵循宾夕法尼亚州立大学伦理审查委员会(IRB)协议PRAMSO40647EP,确保了研究的伦理合规性。 每名受试者需在1至2个月内完成4次脑机接口(BCI)实验场次。每次实验场次包含4个运行区块,每个类别(左手运动想象、右手运动想象、静息态)各设10次试次,因此单次实验场次共计40次试次。实验流程以校准运行作为起始,用于初始化系统,随后进入反馈运行阶段,受试者通过运动想象(MI)——具体为想象抓握动作——控制光标移动。本研究聚焦于运动想象(MI),在两个月内为每名受试者共计收集了160次试次。 本数据集对于研究ALS患者运动想象解码的纵向动态变化具有重要学术价值。为保障研究结果可复现并推动该领域的技术进步,我们已获得宾夕法尼亚州立大学Geronimo教授的明确许可,可将处理后的数据集用于科研用途分发。本数据集的原始发表文献如下:<b>数据集使用方法:</b>本数据集以MATLAB格式存储,为按受试者分类的结构体集合,每名受试者对应一个单独的结构体。每个结构体包含三个字段:<code><strong>L</strong></code>:对应左手运动想象的试次数据;<code><strong>R</strong></code>:对应右手运动想象的试次数据;<code><strong>Re</strong></code>:对应静息态的试次数据。每个字段均包含试次数组,每个试次以矩阵形式表示,<b>行</b>为时间戳,<b>列</b>为电极通道。<br><br>原始文献:Geronimo A, Simmons Z, Schiff SJ. 肌萎缩侧索硬化症患者脑机接口的性能预测因子[J]. 神经工程杂志, 2016, 13. DOI:10.1088/1741-2560/13/2/026002.<br><br>使用本数据集发表相关成果的代码已公开至以下链接:<br>https://github.com/rishannp/Auto-Adaptive-FBCSP<br>https://github.com/rishannp/Motor-Imagery---Graph-Attention-Network<br>




