Fast Nonconvex Deconvolution of Calcium Imaging Data.

Calcium imaging data promises to transform the field of neuroscience bymaking it possible to record from large populations of neurons simultaneously.However, determining the exact moment in time at which a neuron spikes, from acalcium imaging data set, amounts to a non-trivial deconvolution problem whichis of critical importance for downstream analyses. While a number offormulations have been proposed for this task in the recent literature, in thispaper we focus on a formulation recently proposed in Jewell and Witten (2017)which has shown initial promising results. However, this proposal is slow torun on fluorescence traces of hundreds of thousands of timesteps.

Here we develop a much faster online algorithm for solving the optimizationproblem of Jewell and Witten (2017) that can be used to deconvolve afluorescence trace of 100,000 timesteps in less than a second. Furthermore,this algorithm overcomes a technical challenge of Jewell and Witten (2017) byavoiding the occurrence of so-called "negative" spikes. We demonstrate thatthis algorithm has superior performance relative to existing methods for spikedeconvolution on calcium imaging datasets that were recently released as partof the spikefinder challenge (this http URL).

Our C++ implementation, along with R and python wrappers, is publiclyavailable on Github at https://github.com/jewellsean/FastLZeroSpikeInference.