SpikeInterface, a unified framework for spike sorting

Area of Science:

• Neuroscience
• Computational Neuroscience
• Bioinformatics

Background:

• The field of spike sorting has seen extensive development, leading to numerous tools for analyzing neural recordings.
• This proliferation of specialized, often incompatible, software hinders rigorous benchmarking and reproducible scientific analysis.
• A unified framework is needed to streamline the comparison and adoption of diverse spike sorting algorithms.

Purpose of the Study:

• To introduce SpikeInterface, a Python framework designed to consolidate existing spike sorting technologies.
• To facilitate the reproducible comparison and benchmarking of various spike sorting methods.
• To reduce the manual effort in curating and analyzing extracellular electrophysiological data.

Main Methods:

• Development of SpikeInterface, a unified Python framework for spike sorting.
• Integration of multiple existing spike sorting algorithms within the SpikeInterface codebase.
• Application of SpikeInterface to real and simulated extracellular electrophysiological datasets.

Main Results:

• SpikeInterface enables researchers to run, compare, and benchmark diverse spike sorting algorithms with minimal code.
• The framework supports pre-processing, post-processing, and visualization of extracellular datasets.
• Demonstrated reduction in manual curation burden and enhanced benchmarking of automated spike sorters.

Conclusions:

• SpikeInterface provides a unified solution to the challenges posed by the fragmentation of spike sorting tools.
• The framework promotes reproducibility and facilitates comprehensive benchmarking in neural data analysis.
• SpikeInterface empowers researchers to efficiently analyze and interpret complex electrophysiological recordings.