Clusterless Decoding of Position from Multiunit Activity Using a Marked Point Process Filter

Abstract

Point process filters have been applied successfully to decode neural signals and track neural dynamics. Traditionally, these methods assume that multiunit spiking activity has already been correctly spike-sorted. As a result, these methods are not appropriate for situations where sorting cannot be performed with high precision such as real-time decoding for brain-computer interfaces. As the unsupervised spike-sorting problem remains unsolved, we took an alternative approach that takes advantage of recent insights about clusterless decoding. Here we present a new point process decoding algorithm that does not require multiunit signals to be sorted into individual units. We use the theory of marked point processes to construct a function that characterizes the relationship between a covariate of interest (in this case, the location of a rat on a track) and features of the spike waveforms. In our example, we use tetrode recordings, and the marks represent a four-dimensional vector of the maximum amplitudes of the spike waveform on each of the four electrodes. In general, the marks may represent any features of the spike waveform. We then use Bayes’ rule to estimate spatial location from hippocampal neural activity. We validate our approach with a simulation study and with experimental data recorded in the hippocampus of a rat moving through a linear environment. Our decoding algorithm accurately reconstructs the rat’s position from unsorted multiunit spiking activity. We then compare the quality of our decoding algorithm to that of a traditional spike-sorting and decoding algorithm. Our analyses show that the proposed decoding algorithm performs equivalently or better than algorithms based on sorted single-unit activity. These results provide a path toward accurate real-time decoding of spiking patterns that could be used to carry out content-specific manipulations of population activity in hippocampus or elsewhere in the brain.

A Fully Automated Approach to Spike Sorting

“…Cluster quality could also be used to weight the contribution of each unit to a given analysis, thereby ensuring that the best isolated clusters have a proportionally greater influence on the findings. Third, there are cases where including multiunit spiking improves the quality of the results, as is the case for clusterless decoding of animal position from unsorted hippocampal spiking activity (Deng et al, 2015; Kloosterman et al, 2013). Inclusion of all clusters regardless of quality, greatly simplifies the application of these techniques to the data.…”

Sums of Spike Waveform Features for Motor Decoding

“…Their approach is based on the spatial-temporal Poisson process, where a Poisson process describes the spike arrival time and a random vector describes the waveform features of the spike. Later, Deng et al (2015) presented a marked point-process decoder that uses waveform features as the marks of the point-process and tested it on hippocampal recordings. This approach is similar in spirit to the spatial-temporal Poisson process.…”

“…Several studies have used this approach (Chen et al, 2012; Kloosterman et al, 2014; Todorova et al, 2014; Deng et al, 2015; Ventura and Todorova, 2015). Particularly, Ventura and Todorova (2015) proposed an innovative, computationally efficient method for incorporating waveform information in linear decoders.…”

Extracellular voltage threshold settings can be tuned for optimal encoding of movement and stimulus parameters

“…With this knowledge, some researchers have investigated the possibility of moving away from using sorted units as inputs to BCI decoders and instead using threshold crossings (Fraser et al 2009). Many studies agree that BCI performance is substantially degraded when the non-spike parts of the signal are discarded (Kloosterman et al 2014, Todorova et al 2014, Deng et al 2015), raising the intriguing possibility that the threshold could be adjusted to maximize BCI performance.…”