Unsupervised spike sorting for large scale, high density multielectrode arrays

Abstract

SUMMARYWe present a method for automated spike sorting for recordings with high-density, large-scale multielectrode arrays. Exploiting the dense sampling of single neurons by multiple electrodes, an efficient, lowdimensional representation of detected spikes consisting of estimated spatial spike locations and dominant spike shape features is exploited for fast and reliable clustering into single units. Millions of events can be sorted in minutes, and the method is parallelized and scales better than quadratically with the number of detected spikes. Performance is demonstrated using recordings with a 4,096-channel array and validated using anatomical imaging, optogenetic stimulation, and model-based quality control. A comparison with semi-automated, shape-based spike sorting exposes significant limitations of conventional methods. Our approach demonstrates that it is feasible to reliably isolate the activity of up to thousands of neurons and that dense, multi-channel probes substantially aid reliable spike sorting.

Fast and accurate spike sorting invitroandin vivofor up to thousands of electrodes

“…Datasets from thousands of electrodes necessitate a spike sorting algorithm that is extensively automated. Furthermore, the few algorithms that have been designed to process large-scale recordings have not been tested on data where one neuron is recorded by the large-scale recordings and simultaneously by another technique, so that the success rate of the spike sorting algorithm can be measured [Pachitariu et al, 2016, Leibig et al, 2016, Hilgen et al, 2016.…”

SpikeInterface, a unified framework for spike sorting

“…For this analysis, we select six different spike sorters: HerdingSpikes2 [36], Kilosort2 [58], IronClust [39], SpyKING Circus [74], Tridesclous [31], and HDSort [23] 3 . As most of these algorithms have been tuned rigorously on multiple ground truth datasets (including the recent large-scale evaluation from [47]), we fix their parameters to default values to allow for straightforward comparison.…”

“…To alleviate this issue, spike sorting has seen decades of algorithmic and software improvements to increase both the accuracy and automation of the process [62]. This progress has accelerated in the past few years as high-density devices [27,12,28,10,55,75,46,40,24,9], capable of recording from hundreds to thousands of neurons simultaneously have made manual intervention impractical, increasing the demand for both accurate and scalable spike sorting algorithms [65,59,44,20,74,36,39,23].…”

Recent progress in multi-electrode spike sorting methods

“…Nowadays, new devices with CMOS components now allow recordings from thousands electrodes simultaneously (Berdondini et al (2005); Fiscella et al (2012); Müller et al (2015); Hilgen et al (2016)), and it remains to be seen it these algorithms can scale up and process such a large amount of data. We need to be sure that the time spent on manual curation can remain small enough that we can get thousands of spike trains in a decent amount of time (see preliminary evidence that it might be the case by Yger et al (2016)).…”

“…tetrodes), methods that could be seen as adaptations of single electrode sorting worked very well (McNaughton et al, 1983;Harris et al, 2000;Gao et al, 2012), this is not the case with new devices designed with hundreds of electrodes all densely packed. CMOS-based devices with thousands of electrodes have been tested and are now frequently used (Berdondini et al (2005); Fiscella et al (2012); Müller et al (2015); Hilgen et al (2016)), calling for new algorithmic methods, largely different from the usual sorting methods.…”