Autocorrelogram

An autocorrelogram plots the probability of observing a spike at various time lags relative to each spike in a spike train, revealing temporal patterns in a single neuron's firing.

Why it matters: Autocorrelograms reveal oscillatory firing (periodic peaks), bursting (peaks at short lags), and the refractory period (gap near zero lag). They are a standard tool for neuron classification and spike sorting quality assessment.

Construction: For each spike, count other spikes in the same train at each time lag (typically ±50–500 ms in 1 ms bins). The result is normalized and plotted as a histogram.

Interpretation: A clean refractory gap at ±1–2 ms confirms good single-unit isolation. Periodic peaks indicate oscillatory firing at the peak frequency. A broad peak at short lags indicates bursting.