Skip to main content

Python toolkit for computing spike sorting metrics

Project description

SpikeMetrics

Computes quality metrics for sorted units. This module can calculate metrics separately for individual epochs. If no epochs are specified, metrics are computed for the entire recording.

The base code (and portions of the README images/description) was ported from: https://github.com/AllenInstitute/ecephys_spike_sorting/tree/master/ecephys_spike_sorting/modules/quality_metrics

Copyright © 2019. Allen Institute. All rights reserved.

Included Metrics

Metric Icon Description Reference
Spike Count Spike count in an epoch
Firing rate Mean spike rate in an epoch
Presence ratio Fraction of epoch in which spikes are present
ISI violations Rate of refractory-period violations
Isolation distance The Mahalanobis distance from a specified unit within as many spikes belong to the specified unit as to other units Harris et al. Neuron 32.1 (2001): 141-149.
L-ratio The Mahalanobis distance and χ2 inverse cdf (given the assumption that the spikes in the cluster distribute normally in each dimension) are used to find the probability of cluster membership for each spike. Schmitzer-Torbert and Redish. J Neurophy 91.5 (2004): 2259-2272.
d' The classification accuracy between units based on linear discriminant analysis (LDA). Hill et al. (2011) J Neurosci 31, 8699-9705
Nearest-neighbors Non-parametric estimate of unit contamination using nearest-neighbor classification. Chung et al. (2017) Neuron 95, 1381-1394
Silhouette score A standard metric for quantifying cluster overlap
Maximum drift Maximum change in spike depth throughout recording
Cumulative drift Cumulative change in spike depth throughout recording

A Note on Calculations

For metrics based on waveform principal components (isolation distance, L-ratio, d', and nearest neighbors hit rate and false alarm rate), it is typical to compute the metrics for all pairs of units and report the "worst-case" value. We have found that this tends to under- or over-estimate the degree of contamination when there are large firing rate differences between pairs of units that are being compared. Instead, we compute metrics by sub-selecting spikes from all other units on the same set of channels, which seems to give a more accurate picture of isolation quality. We would appreciate feedback on whether this approach makes sense.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for spikemetrics, version 0.1.3
Filename, size File type Python version Upload date Hashes
Filename, size spikemetrics-0.1.3.tar.gz (12.4 kB) File type Source Python version None Upload date Hashes View hashes

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page