Homogenized $\textit{C. elegans}$ Neural Activity and Connectivity Data

Quilee Simeon,Anshul Kashyap,Konrad P Kording,Edward S Boyden

2024-11-19

Abstract:There is renewed interest in modeling and understanding the nervous system of the nematode $\textit{Caenorhabditis elegans}$ ($\textit{C. elegans}$). This is particularly interesting as this model system provides a path to bridge the gap between structure and function, from nervous system connectivity to physiology. However, the many existing physiology datasets, both recording and stimulation, as well as connectome datasets, are in distinct formats, requiring extra processing steps before modeling or other analysis can commence. Here we present a homogenized dataset of neural activity, including during stimulation, compiled from 11 neuroimaging experiments and from 10 connectome reconstructions. The physiology datasets, collected under varying experimental protocols, all measure neural activity via calcium fluorescence in labeled subsets of the worm's 300 neurons. Our preprocessing pipeline standardizes these datasets by consistently ordering labeled neurons and resampling traces to a common sampling frequency. The resulting dataset includes neural recordings from approximately 900 worms and 250 uniquely labeled neurons. The connectome datasets, collected from electron microscopy (EM) reconstructions, all contain the entire nervous system of the worm, preprocessed into a graph of connections across the neurons. Using our collection of datasets is facilitated through easy data sharing on HuggingFace. We believe that our joint dataset of physiology and connectivity will facilitate modeling, for example in terms of recurrent neural network or transformer architectures, making it easier to check how well models generalize across animals and labs.

Neurons and Cognition