Complex Oscillatory Waves Emerging from Cortical Organoids Model Early Human Brain Network Development

Longitudinal electrophysiological dataset from cortical organoids, recorded weekly using planar multi-electrode arrays over 10 months. Please cite DOI of this repository when using the data. Thanks! <strong>Data associated with the following publication:</strong> Cleber A. Trujillo, Richard Gao, Priscilla D. Negraes, Jing Gu, Justin Buchanan, Sebastian Preissl, Allen Wang, Wei Wu, Gabriel G. Haddad, Isaac A. Chaim, Alain Domissy, Matthieu Vandenberghe, Anna Devor, Gene W. Yeo, Bradley Voytek, Alysson R. Muotri, Complex Oscillatory Waves Emerging from Cortical Organoids Model Early Human Brain Network Development, Cell Stem Cell,<br> Volume 25, Issue 4,<br> 2019,<br> Pages 558-569.e7,<br> ISSN 1934-5909,<br> https://doi.org/10.1016/j.stem.2019.08.002.<br> (https://www.sciencedirect.com/science/article/pii/S1934590919303376) --- Instructions: /CTC_YYMMDD/LFP_Sp_YYMMDD.mat files contain preprocessed data for recording on a given date YY-MM-DD. Each .mat file contains the following variables: (explanation given in python and with 0-indexing) - LFP: broadband recording downsampled to 1000Hz; 12-element cell array, one for each well of the MEA. In each cell is a 2D matrix, [time x 64 channels]. - t_ds: downsampled timestamps at 1000Hz (just a continuously increasing vector, 1ms apart) - t_s: original timestamps at 12.5kHz (same idea as t_ds) - spikes: [12 wells x 64 channel] cell array; each cell contains a 1D array denoting multi-unit spike times in sample number (at 12.5kHz), use t_s to find the corresponding time in seconds. e.g., spikes[5, 32] contains the spiketrain from well 5 (of 12), channel 32 (of 64). Briefly, spikes were detected after bandpass filtering (300-3000Hz) and as threshold crossings adaptively set to 5.5 standard deviations. Note that the spikes were not sorted, and thus represent multi-unit activity. Refer to manuscript for details. - spike_cnt: [12 wells x 64 channel] total spike count at each well and channel. Just a count over the channels of the ‘spikes’ variable. - spike_shape: cell array, [12 wells x 64 channel] each cell contains a 2D array [101 samples x N spikes], containing 101-samples of the spike waveform for each of the N detected spike, centered at time 0 of spike, i.e., detected spike peak/trough. Empty arrays denote no spikes detected. - spike_avg: [12 wells x 64 channel x 101 timepoints] 3D array average spike waveform detected at that well and channel, 101-samples (at 12.5kHz), i.e., averaged over all N spikes in each element of the ‘spike_shape” variable. ———————— Note that wells 1-4 are empty, all data analyzed in the paper are from wells 5-12. To load the data: in MATLAB: just open the .mat file and do as you please. in python: from scipy import io data = io.loadmat(‘CTC_YYMMDD/LFP_Sp_YYMMDD.mat’, squeeze_me = True) LFP = data[‘LFP’]本数据集为皮层类器官的纵向电生理数据集，采用平面多电极阵列（planar multi-electrode arrays）每周采集，时长跨越10个月。使用本数据时，请引用本仓库的DOI。感谢！<strong>本数据关联以下发表论文：</strong> Cleber A. Trujillo、Richard Gao、Priscilla D. Negraes、Jing Gu、Justin Buchanan、Sebastian Preissl、Allen Wang、Wei Wu、Gabriel G. Haddad、Isaac A. Chaim、Alain Domissy、Matthieu Vandenberghe、Anna Devor、Gene W. Yeo、Bradley Voytek、Alysson R. Muotri，论文题为《Complex Oscillatory Waves Emerging from Cortical Organoids Model Early Human Brain Network Development》，发表于*Cell Stem Cell*（《细胞干细胞》），2019年，第25卷第4期，页码范围为558-569.e7，ISSN 1934-5909，DOI链接：https://doi.org/10.1016/j.stem.2019.08.002（论文原文链接：https://www.sciencedirect.com/science/article/pii/S1934590919303376） --- 使用说明：/CTC_YYMMDD/LFP_Sp_YYMMDD.mat 文件包含指定日期YY-MM-DD对应的预处理后电生理数据。每个.mat文件包含以下变量（以下说明采用Python语法规范，且以0索引为准）： - LFP：局部场电位（Local Field Potential），宽带原始记录信号经下采样至1000Hz；为包含12个元素的元胞数组，分别对应平面多电极阵列的12个孔板。每个元胞内存储一个二维矩阵，维度为[时间步数 × 64个通道]。 - t_ds：1000Hz采样率下的下采样时间戳，为连续递增的一维数组，相邻采样点间隔1ms。 - t_s：12.5kHz采样率下的原始时间戳，格式与t_ds一致。 - spikes：维度为[12个孔板 × 64个通道]的元胞数组；每个元胞内存储一个一维数组，以采样点数形式记录多单元锋电位（spike）的时间点（采样率12.5kHz），可通过t_s数组转换为对应的秒级时间。例如，spikes[5, 32] 对应12个孔板中的第5个孔、64个通道中的第32个通道的锋电位序列。锋电位检测流程简述如下：先对信号进行300-3000Hz的带通滤波，再以自适应阈值（设为5.5倍标准差）进行跨阈检测。需注意，本数据未对锋电位进行分类，因此记录的是多单元活动信号。详细细节请参阅发表论文手稿。 - spike_cnt：维度为[12个孔板 × 64个通道]的数组，记录每个孔板与通道的总锋电位数量，为spikes变量中各通道内锋电位的总数。 - spike_shape：维度为[12个孔板 × 64个通道]的元胞数组；每个元胞内存储一个二维数组[101个采样点 × N个锋电位]，存储每个检测到的锋电位的101个采样点的波形数据，波形以锋电位的时间零点（即检测到的锋电位峰或谷点）为中心。空数组代表该孔板与通道未检测到锋电位。 - spike_avg：维度为[12个孔板 × 64个通道 × 101个时间点]的三维数组，存储对应孔板与通道的平均锋电位波形，包含101个采样点（采样率12.5kHz），即对spike_shape变量中每个元素内的所有N个锋电位波形取平均所得。 ———————— 重要提示：孔板1至4无有效数据，论文中所有分析所用数据均来自孔板5至12。 数据加载方式： - MATLAB环境：直接打开.mat文件即可进行后续操作。 - Python环境：可通过以下代码加载：`from scipy import io; data = io.loadmat('CTC_YYMMDD/LFP_Sp_YYMMDD.mat', squeeze_me=True); LFP = data['LFP']`