The data were acquired from publicly available extracellular recordings of human brain organoids. Recordings were collected using high-density CMOS microelectrode arrays for organoid slices and Neuropixels probes for whole-organoid measurements. The dataset captures spontaneous neural activity across developmental time points and under pharmacological perturbation (benzodiazepine treatment). All data were downloaded programmatically from a public repository using an authenticated API workflow.
Each recording was processed through a Slurm-based analysis pipeline. For every dataset, channel-wise voltage root-mean-square (RMS) values were computed directly from the raw extracellular signals using chunked HDF5 reads to ensure memory-efficient processing. When spike timing information was available, per-channel firing rates were calculated from spike counts normalized by recording duration. Results were aggregated across experimental groups to compare: (1) drug concentration versus control conditions, and (2) developmental time points of organoids. Summary statistics (mean ± SEM) and full distributions were computed to examine changes in signal amplitude and spiking activity.
Figure 1: Summary of voltage RMS and firing rate metrics across drug and developmental conditions. Top row shows mean ± SEM channel-wise RMS amplitude for diazepam concentrations compared to control (left) and across developmental ages (right). Middle row displays distributions of voltage RMS values per channel. Bottom row shows firing rate distributions. Drug conditions are color-coded by increasing concentration, while developmental recordings are ordered chronologically. These plots highlight potential changes in network activity amplitude and spiking statistics across experimental manipulations; however, further statistical analyses are required to determine significance.