Abstract
Spike sorting is the process of retrieving the spike times of individual neurons that are present in an extracellular neural recording. Over the last decades, many spike sorting algorithms have been published. In an effort to guide a user towards a specific spike sorting algorithm, given a specific recording setting (i.e., brain region and recording device), we provide an open-source graphical tool for the generation of hybrid ground-truth data in Python. Hybrid ground-truth data is a data-driven modelling paradigm in which spikes from a single unit are moved to a different location on the recording probe, thereby generating a virtual unit of which the spike times are known. The tool enables a user to efficiently generate hybrid ground-truth datasets and make informed decisions between spike sorting algorithms, fine-tune the algorithm parameters towards the used recording setting, or get a deeper understanding of those algorithms.
Footnotes
This work was carried out at the ESAT Laboratory of KU Leuven, in the frame of KU Leuven Special Research Fund projects C14/16/057, and the Research Foundation Flanders (FWO) project FWO G0D7516N. This project has received funding from the European Research Council (ERC) under the European Unions Horizon 2020 research and innovation programme (grant agreement No 802895). The scientific responsibility is assumed by its authors.