LIBPhys-EGG: An Open-Access Multichannel Electrogastrography and Accelerometry Dataset for Reproducible Research in Gastrointestinal Electrophysiology

Surface electrogastrography (sEGG) records gastrointestinal (GI) myoelectric activity non-invasively from the abdominal surface, offering a potential glimpse into gastric and intestinal function. Despite nearly a century of investigation, clinical adoption remains limited by low Signal-to-Noise Ratio, motion artifacts, and the lack of standardized, large-scale public datasets particularly those capturing intestinal activity and variations in electrode placement. This work presents a multichannel sEGG dataset acquired using a wearable sensor belt, synchronized with tri-axial accelerometry, designed to support reproducible research in GI signal processing and machine learning. OVERVIEWThis dataset contains multisite surface electrogastrography (sEGG) recordings from 15 healthy adult participants during controlled meal challenge experiments.The LIBPhys-EGG dataset captures gastric and intestinal myoelectrical activity during fasting (pre-prandial), meal ingestion, and prolonged post-prandial states. The recordings include: - 4-channel sEGG signals from anatomically defined abdominal sites - 3-axis accelerometer data for motion artifact identification - Detailed annotations including meal timing, digestive states, and artifacts KEY STATISTICS - Total recording volume: 111.8 hours - Number of recording sessions: 30 - Number of participants: 15 - Median session duration: 4.00 hours - Mean primary session duration: 5.16 ± 1.22 hours - Minimum recorded window: 30 min pre-prandial + 1 hour post-prandial PARTICIPANT INFORMATION - Number of participants: 15 (6 male, 9 female) - Age: 30.6 ± 10.9 years - Male: 34.8 ± 12.7 years - Female: 27.8 ± 9.2 years - BMI: 22.2 ± 3.9 - Male: 23.7 ± 4.9 - Female: 21.0 ± 2.9 - All participants were healthy adults free from: - Known GI disorders (IBD, IBS, gastroparesis) - History of major abdominal surgery - Medications affecting GI motility RECORDING PARAMETERS - Sampling rate: 1000 Hz - ADC resolution: 16-bit - Recording device: 8-channel biosignalsplux hub - sEGG sensor specifications: - Bandwidth: 0.0023–0.75 Hz - Gain: 6114 - Input impedance: >100 GΩ - CMRR: 100 dB - Accelerometer specifications: - Technology: MEMS - Range: ±3.60 g - Bandwidth: 0–50 Hz - Axes: 3 (x, y, z) - Electrodes: Pre-gelled, disposable Ag/AgCl (Cardinal Health, 24mm diameter)表面胃电图（surface electrogastrography, sEGG）可通过腹部体表无创记录胃肠道（gastrointestinal, GI）的肌电活动，为探究胃与肠道功能提供了潜在途径。尽管相关研究已开展近百年，但临床应用仍受限于低信噪比（Signal-to-Noise Ratio）、运动伪影干扰，以及缺乏标准化的大规模公开数据集，尤其是能够捕捉肠道活动与电极放置位置差异的数据集。本研究公开了一套基于可穿戴传感器腰带采集的多通道sEGG数据集，该数据集与三轴加速度计数据同步，旨在为胃肠道信号处理与机器学习领域的可复现性研究提供支撑。 ### 数据集概述 本数据集包含15名健康成年受试者在可控膳食激发实验期间的多部位表面胃电图（sEGG）记录。LIBPhys-EGG数据集采集了受试者在禁食（餐前）、进食及餐后较长时段内的胃与肠道肌电活动。 记录内容包括： - 来自解剖学定位腹部位点的4通道sEGG信号 - 用于运动伪影识别的三轴加速度计数据 - 包含进食时间、消化状态与伪影的详细标注 ### 关键统计信息 - 总记录时长：111.8小时 - 记录会话数量：30次 - 受试者人数：15名 - 会话中位时长：4.00小时 - 单会话平均时长：5.16 ± 1.22小时 - 最小记录窗口：餐前30分钟 + 餐后1小时 ### 受试者信息 - 受试者总数：15名（男性6名，女性9名） - 年龄：30.6 ± 10.9岁 - 男性受试者：34.8 ± 12.7岁 - 女性受试者：27.8 ± 9.2岁 - BMI指数：22.2 ± 3.9 - 男性受试者：23.7 ± 4.9 - 女性受试者：21.0 ± 2.9 - 所有受试者均为健康成年人，无以下情况： - 已知胃肠道疾病（炎症性肠病、肠易激综合征、胃轻瘫） - 腹部大手术史 - 影响胃肠道动力的药物使用史 ### 记录参数 - 采样率：1000 Hz - 模数转换器（Analog-to-Digital Converter, ADC）分辨率：16位 - 记录设备：8通道biosignalsplux集线器 - sEGG传感器参数： - 带宽：0.0023–0.75 Hz - 增益：6114 - 输入阻抗：>100 GΩ - 共模抑制比（Common Mode Rejection Ratio, CMRR）：100 dB - 加速度计参数： - 技术：微机电系统（Micro-Electro-Mechanical Systems, MEMS） - 量程：±3.60 g - 带宽：0–50 Hz - 轴数：3轴（x、y、z轴） - 电极：预凝胶一次性银/氯化银（Ag/AgCl）电极（Cardinal Health，24mm直径）