SPySort is a Python package for spikes sorting. The package has the ability to read raw neurophysiological data of extracellular recordings in binary format. Although, the user is not forced to use the modules of SPySort in order to read and preprocess their raw data. The spike sorting method that is mainly implemented is given in [1].

The modules of SPySort are:

• import_data.py : This module offers a class of functions for reading, normalizing, subsetting data, selecting specific recording channels, a summary functions that prints to the stdout all the important statistical numbers related to the raw data and an interactive plot function.

• spikes.py : This module includes all the necessary methods in order to filter the raw data and to detect the spike events. In addition, two plot functions are included. One for plotting the spike events and another one which plots the spike events over the raw data and the threshold values as well.

• events.py : In this module the user can find methods for extracting the spike events from the raw normalized data. There are two main methods that make the events and the noise. Moreover, two plot functions are included in order to facilitate the visualization of the events and some statistical measures over the created events.

• clusters.py : PCA based clustering is included in this module. A Principal Component Analysis is applied on the events in order to reduce the dimension of the problem and then a plethora of clustering methods are implemented. The basic clustering method is the k-means algorithm. In addition, a Gaussian Mixture model and a Bagged clustering algorithm have been implemented, as well. Plot functions for the PCAs and the clustered events are also provided.

• alignment.py : Because the spike events are not aligned the most of the times a brute-force algorithm of spike events alignment has been included in this module. Therefore, after the clustering of the spike events the user can refine the spike events and improve the clustering by using the methods provided by this module.

References:

[1] Pouzat, C., Mazor, O. & Laurent, G., "Using noise signature to optimize spike-sorting and to assess neuronal classification quality", Journal of neuroscience methods, 122, pp.43–57, 2002.