Multineuron whole-cell patch-clamp dataset of layer 5 pyramidal cell synaptic connectivity in juvenile rat somatosensory cortex

Reconstructing the wiring rules of the neocortical microcircuit ideally requires simultaneous electrophysiological access to many neighboring neurons together with a precise record of the stimuli that probe their connections. This dataset extends the whole-cell patch-clamp recordings from Perin, Berger, and Markram (2011, Proceedings of the National Academy of Sciences of the United States of America [PNAS]), in which synaptic connectivity was mapped within local groups of thick-tufted layer 5 pyramidal neurons in the juvenile rat somatosensory cortex. It comprises 47 recording sessions, stored as Neurodata Without Borders (NWB) HDF5 files, in which up to 12 neurons were patched and recorded simultaneously in current-clamp mode using a purpose-built multineuron (“12-patch”) setup. Across these sessions, the dataset includes 150,121 membrane-voltage acquisition sweeps (CurrentClampSeries; each a 5,000–20,000-sample waveform sampled at 5–10 kHz), 68,968 injected-current stimulus series (CurrentClampStimulusSeries—recorded on the secondary channel where available, otherwise reconstructed from the Igor wave-note protocol definitions), delivered under 160 distinct stimulus protocols, and 482 patched-cell electrode configurations. To make the collection queryable and artificial intelligence (AI)-ready without opening every HDF5 file, four analytical index tables are provided in Apache Parquet format—session_metadata (47 rows), acquisition_series (150,121 rows), electrode_metadata (482 rows) and stimulus_series (68,968 rows)—joined on session_id and the source NWB file path. The package is distributed as a FAIR² data product with machine-readable metadata in MLCommons Croissant format, World Wide Web Consortium (W3C) PROV Ontology (PROV-O) provenance, a full data dictionary and reproducible extraction scripts, under an Open Data Commons (ODC)-By version 1.0 attribution license. It offers a large, standardized, connectivity-oriented electrophysiology resource for microcircuit modeling, synaptic physiology methods development, and machine-learning applications on raw intracellular signals.