Multimodal Dataset of Freezing of Gait in Parkinson's Disease

Freezing of gaits (FOG), a debilitating transitory inability to pursue moving, is one of the severest symptoms of Parkinson's disease (PD). An accurate and reliable detection or prediction of FOG is of great significance for PD patients' assessment and rehabilitation. It is difficult to detect FOG with sufficient low-latency and high precision based on single sensor information. In order to improve the detection accuracy and facilitate further research, we gathered and presented a new multimodal dataset by combining rich physical and physiological sensor information. The multimodal data, including electroencephalogram (EEG), electromyogram (EMG), electrocardiogram (ECG), skin conductance (SC), and acceleration (ACC) in walking tasks, were collected using a high-quality hardware system integrated commercial and self-designed sensors. A standard experimental procedure was carefully designed to induce FOG in hospital surroundings. A total number of 12 PD patients completed the experiments and produced a total length of 3 hours and 42 minutes valid data. The FOG episodes in the multimodal data were labeled by two qualified physicians. The multimodal data can be used to efficiently discriminant FOG from normal locomotion, and indicated that changes in the multimodal motional and electrophysiological signals during FOG episodes could be used to guide PD patients' treatment and recovery.

步态冻结（Freezing of gaits, FOG）是一种使人衰弱的暂时性运动不能，属于帕金森病（Parkinson's disease, PD）最严重的症状之一。对FOG进行准确可靠的检测与预测，对帕金森病患者的病情评估与康复治疗具有重要意义。仅依靠单一传感器信息，难以实现低延迟、高精度的FOG检测。为提升FOG检测精度并推动相关研究，我们整合了丰富的物理与生理传感器数据，构建并发布了一个新型多模态数据集。该多模态数据集采集了行走任务中的多类信号，包括脑电图（electroencephalogram, EEG）、肌电图（electromyogram, EMG）、心电图（electrocardiogram, ECG）、皮肤电导率（skin conductance, SC）以及加速度（acceleration, ACC），数据采集依托一套集成了商用传感器与自研传感器的高精度硬件系统完成。研究人员精心设计了标准实验流程，在医院环境下诱发受试者出现FOG症状。共有12名帕金森病患者参与了实验，最终获得总时长3小时42分钟的有效数据。数据集中的FOG发作片段均由两名资质合格的医师进行标注。该多模态数据集可用于高效区分FOG发作与正常行走运动，同时研究表明，FOG发作期间多模态运动与电生理信号的变化特征，可为帕金森病患者的治疗与康复提供指导依据。