neuropixels-analysis

You can run an end-to-end Neuropixels data analysis pipeline that handles preprocessing, motion correction, spike sorting with Kilosort4, quality metrics, and AI-assisted curation automatically. It processes raw recordings into curated neural units using SpikeInterface.

Yes, the spike sorting workflow supports loading raw Neuropixels 1.0 and 2.0 data directly from SpikeGLX, Open Ephys, and NWB formats. It processes these recordings through a standardized analysis pipeline.

Yes, you can use GPU-enabled Kilosort4 for spike sorting and utilize automated Phy export for curation. The workflow integrates AI-assisted curation and exports publication-ready results directly to Phy or NWB.

You need a Python environment with SpikeInterface, pandas, numpy, scipy, matplotlib, and probeinterface. For spike sorting, optional GPU-enabled sorters like Kilosort4 are supported, alongside optional AI tools for curation.

Yes, AI-assisted curation automates the refinement of sorted neural units after spike sorting. The workflow calculates quality metrics to evaluate units and supports automated export to Phy or NWB for publication-ready results.

The best way to apply motion correction is using the built-in preprocessing steps in this SpikeInterface pipeline. It standardizes raw Neuropixels recordings before spike sorting to improve data quality and reproducibility.