waveform_classification

Discription: Waveform classifcation with extracellular recordings in mouse brain using Neuropixels probes

This repo include supplemental materials, data and code used in JNP paper:

Source: Jia et al. (2019) "High-density extracellular probes reveal dendritic backpropagation and facilitate neuron classification." J Neurophys 121: 1831-1847

Data includes:

1. multi-channel waveforms
2. single-channel waveforms
3. extracted features and cluster labels
4. velocity profiles

1D waveform features: Waveform duration, peak-trough ratio, repolarization slope, and recovery slope.

Example 2D waveform: Signals from channels along one edge of the probe are smoothed with a moving average to create the 2D waveform. Green dots indicate the location of the waveform trough on each channel.

2D waveform features: Waveform spread, velocity above the soma, and velocity below the soma.

Codes used to process waveform features and quality metrics are in: https://github.com/jiaxx/waveform_feature_extraction and in the ecephys_spike_sorting repo: https://github.com/AllenInstitute/ecephys_spike_sorting/tree/master/ecephys_spike_sorting/modules/mean_waveforms

The quality metrics and waveform features developed in this paper are included in the unit_table of publicly released Neuropixels dataset, which can be accessed through AllenSDK: https://github.com/AllenInstitute/AllenSDK/tree/master/allensdk/brain_observatory/ecephys