RT Journal Article SR Electronic T1 Granger causality analysis for calcium transients in neuronal networks: challenges and improvements JF bioRxiv FD Cold Spring Harbor Laboratory SP 2022.06.27.497721 DO 10.1101/2022.06.27.497721 A1 Xiaowen Chen A1 Faustine Ginoux A1 Thierry Mora A1 Aleksandra M. Walczak A1 Claire Wyart YR 2022 UL http://biorxiv.org/content/early/2022/06/29/2022.06.27.497721.abstract AB A major challenge in neuroscience is to understand how information flows between neurons, triggering specific behaviour. At large scales, Granger causality (GC) has been proposed as a simple and effective measure for identifying dynamical interactions. At single-cell resolution however, GC analysis is rarely used compared to directionless correlation analysis. We discuss here the applicability of Granger Causality analysis for population calcium imaging data. We used recordings from motoneurons in the zebrafish embryo and the entire brainstem region of zebrafish larvae during active visuomotor behavior, and synthetic data simulating intracellular calcium fluctuations of spiking neurons in a chosen neuronal network. We first show that despite underlying linearity assumptions, GC analysis can successfully retrieve non-linear interactions in a synthetic network. We then discuss the potential pitfalls and challenges when applying GC analysis on population calcium imaging data, including the effects of calcium signal preprocessing and motion artefacts. We show how to optimize the choice of GC analysis parameters, such as the relevant time delay and the GC value significance threshold to account for the properties of the data. Applied to motoneuron and hindbrain datasets from larval zebrafish, we show how the improved GG better reflects information flow between neurons.Competing Interest StatementThe authors have declared no competing interest.