Spike sorting for large, dense electrode arrays

Abstract

Developments in microfabrication technology have enabled the production of neural electrode arrays with hundreds of closely-spaced recording sites, and electrodes with thousands of sites are currently under development. These probes in principle allow the simultaneous recording of very large numbers of neurons. However, use of this technology requires the development of techniques for decoding the spike times of the recorded neurons, from the raw data captured from the probes. Here, we present a set of novel tools to solve this problem, implemented in a suite of practical, user-friendly, open-source software. We validate these methods on data from the cortex, hippocampus, and thalamus of rat, mouse, macaque, and marmoset, demonstrating error rates as low as 5%.Users may view, print, copy, and download text and data-mine the content in such documents, for the purposes of academic research, subject always to the full Conditions of use

A Fully Automated Approach to Spike Sorting

“…Wilson; MClust, A.D. Redish; Offline Sorter, Plexon). In other situations, clustering is automated, but the user must curate the results by selecting which clusters to reject, merge, or even split (Hill et al, 2011; Kadir et al, 2014; Rossant et al, 2016). There also exist post-processing steps that resolve overlapping spikes (Ekanadham et al, 2014; Franke et al, 2015; Pachitariu et al, 2016; Pillow et al, 2013), and algorithms based on independent component analysis (ICA) that do not explicitly involve clustering (Takahahi et al, 2002).…”

“…Manual sorting can have error rates in excess of 20% (Wood et al, 2004) and there is substantial variability in labeling across different sorting sessions (Harris et al, 2000; Pedreira et al, 2012; Rossant et al, 2016). Furthermore, the human spike sorter could never keep up with the increasing volume of data arising from increasingly large electrode arrays applied over increasingly long durations (Berenyi et al, 2014; Dhawale AK, 2015; Du et al, 2011; Lopez et al, 2016; Santhanam et al, 2007; Shobe et al, 2015).…”

Why sharing matters for electrophysiological data analysis

“…A good example of this is filtering (Quian Quiroga and Panzeri, 2009), where the type of filter is shown to matter, and the use of spike shapes in spike sorting (Rossant et al, 2015). Workflows often include loops, whether for repeating an analysis across a number of datasets, or for repeating an analysis while varying some of the analysis parameters.…”

“…There is a huge and active literature on the analysis tools, whether for electrophysiology or other forms of neuroscience dataset (for example, Lewicki (1998), Harris et al (2000), Hulata et al (2002), Takahashi et al (2003), Pouzat et al (2004), Mizuseki et al (2014), Rossant et al (2015), to name but a few purely in the area of spike sorting). However, simply noting that a particular type of analysis (based on some equations in a paper, for example) has been carried out is not the same as sharing the analysis tools.…”

State-of-the-art MEMS and microsystem tools for brain research

“…This same study also found as many as 24 single units in a virtual tetrode in the relatively sparse visual cortex. More recently, software algorithms that utilize precise spatial information as part of the sorting logic show great promise for improving both speed and accuracy 30,[80][81][82] . Investigating the optimal microelectrode size and pitch for speed and accuracy in spike sorting is an important research area.…”