Spike Sorting
Also known as: unit isolation, single-unit recording, electrophysiology clustering
Spike sorting is an electrophysiological technique for identifying and isolating action potentials of individual neurons from extracellular electrical recordings. Central to single-unit neurophysiology, spike sorting assigns spikes recorded on electrode arrays to their neuron of origin, enabling study of individual neuron firing patterns, timing, and network interactions with single-cell resolution.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Spike sorting is appropriate for studying single-neuron firing dynamics, neuronal coding, and network interactions when cellular-level resolution is essential. Use spike sorting for behavioral correlation studies, circuit mapping, and neural computation research. Avoid spike sorting when only population-level activity suffices (alternative: local field potentials) or when temporal resolution of seconds is acceptable.
Strengths & limitations
- Provides single-cell resolution: can identify firing patterns of individual neurons
- High temporal precision (millisecond scale); resolves spike timing relationships and synchrony
- Allows study of neurons across all brain regions; not limited to accessible surface cortex
- Enables recording from many neurons simultaneously; modern arrays record from hundreds to thousands of units
- Biased toward neurons near electrodes; nearby neurons overrepresented, distant neurons underdetected
- Spike isolation ambiguous; contamination (false positives: noise misclassified as spikes) and merging (false negatives: true spikes missed) common
- Quality assessment difficult; no ground truth in vivo; unit quality metrics inconsistent across labs
- Computationally demanding; high-channel recording (>1000 channels) requires sophisticated algorithms and hardware
Frequently asked
What waveform features should I use for clustering?
Common features: peak amplitude, trough amplitude, action potential width, principal components of waveform, energy. No consensus on optimal features. Best practice: use multiple feature representations (raw waveforms, PCA, templates) and validate clustering with multiple algorithms. Redundancy provides robustness.
How do I estimate unit quality and spike contamination?
Contamination estimated from refractory period violations: spikes <2 ms apart suggest two spikes erroneously assigned to one unit (contamination). Interspike interval (ISI) violation rate indicates contamination. Other metrics: isolation distance (distance from cluster center), L-ratio (noise overlap). Report all quality metrics; units with >5% contamination should be excluded or downweighted.
Can automated spike sorting replace manual curation?
Partially. Automated methods (Kilosort, Ironclust) achieve ~90% accuracy compared to expert curation. Manual inspection remains gold standard, catching overclustering and underclustering. Best practice: automated sorting + expert validation; full manual curation is tedious and becomes impractical with >100 units.
How do I handle spike amplitude drift over recording time?
Drift occurs as electrodes move slightly during recording. Monitor spike waveforms over time; if clear amplitude changes occur, re-sort recording in shorter blocks (e.g., 10–15 minute epochs). Some algorithms (Kilosort) account for drift. Alternatively, compute drift offline using template matching and correct waveforms before final spike sorting.
Sources
- Harris, K. D., Csicsvari, J., Hirase, H., et al. (2016). Accuracy of tetrode spike separation as determined by simultaneous intracellular and extracellular recordings. Journal of Neurophysiology, 84(1), 401–414. link ↗
- Yger, P., Spampinato, G. L., Esposito, E., et al. (2018). A spike sorting toolbox for up to thousands of electrodes validated with ground truth recordings in vitro and in vivo. eLife, 7, e34518. DOI: 10.7554/eLife.34518 ↗
How to cite this page
ScholarGate. (2026, June 3). Spike Sorting. ScholarGate. https://scholargate.app/en/neuroimaging/spike-sorting
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- eLORETANeuroimaging↔ compare
- Event-Related Potential AnalysisNeuroimaging↔ compare
- MEG Source LocalizationNeuroimaging↔ compare