neuropixels-analysis

To run spike sorting with drift correction on Neuropixels recordings, load raw SpikeGLX, OpenEphys, or NWB data, apply preprocessing and motion correction, then execute algorithms like Kilosort4, SpykingCircus2, or Mountainsort5 to isolate reliable units.

Yes, you can use SpikeInterface for automated unit curation on Neuropixels 2.0 data by computing post-processing quality metrics and applying Allen or IBL criteria to identify well-isolated neural units with minimal manual intervention.

The best way to preprocess raw SpikeGLX electrophysiology data involves filtering, detecting bad channels, and applying Common Average Reference (CAR) or Common Median Reference (CMR) to prepare clean signals for accurate spike sorting.

Yes, the workflow supports exporting sorted units, waveforms, templates, and quality metrics to Phy for manual inspection and NWB for standardized data sharing, while also generating publication-ready reports for visualizing drift and unit summaries.

Yes, you need Python data-science libraries like NumPy, pandas, and matplotlib alongside SpikeInterface to process Neuropixels recordings, compute quality metrics, visualize drift, and generate publication-ready unit summaries.

AI-assisted visual curation handles uncertain units in spike sorting outputs by providing an optional layer of automated review that evaluates waveforms and quality metrics to classify borderline units, reducing manual verification time while maintaining curation accuracy.