CASCADE: Calibrated inference of spiking from calcium ΔF/F data using deep networks

CASCADE translates calcium imaging ΔF/F traces into spiking probabilities or discrete spikes. CASCADE is based on deep networks and runs on Python (Linux/Windows/Mac); it can be used either as a web application (Colaboratory notebook), or locally on a CPU or GPU. CASCADE therefore does not necessarily require any local installations, and it does not require any parameter tuning to different datasets. The algorithm is trained on a large ground truth database of simultaneous calcium imaging and cell-attached electrophysiological recordings (almost 300 neurons). The database is freely accessible and part of the Github repository.

Project Author(s)

Peter Rupprecht; Adrian Hoffmann; Rainer Friedrich; Fritjof Helmchen

This post was automatically generated by Peter Rupprecht

Edit this page