The Sawtell laboratory at Columbia University, in collaboration with the Lichtman laboratory and Connectomics team at Google, are releasing the electron microscopy (EM) dataset from the electrosensory lobe (ELL) connectome of Gnathonemus petersii (African pulse-type electric fish). This project combines modeling and electrophysiolgy with targeted (sparsely reconstructed) connectomics, building on decades of knowledge about the behavior and physiology of this fascinating animal to understand how a neural circuit can support multi-layer learning that is rapid, accurate, and robust to noise.

The electron microscopy (EM) dataset, acquired at a resolution of 4 nm × 4 nm × 30 nm, provides ultrastructural details such as synaptic connections and subcellular features. This volume includes ~10,000 cell bodies encompassing all main layers of the passive electrosensory lateral line lobe (ELL). For the first publication of this project [DOI link], ~5000 bodies were reconstructed and ~3000 of those were sortable into the main cell types of ELL: EAF, MG, Output, SG, Gr+, SP-, GC, and MLI (see companion paper for cell type descriptions and connectomic and modelling analysis). Approximately 30,000 synapses were manually annotated to determine the connectivity among these cells.