braintools.analysis.spike.auto_correlogram

braintools.analysis.spike.auto_correlogram#

braintools.analysis.spike.auto_correlogram(trains, bin_size, max_lag)#

Compute the spike-time auto-correlogram of every neuron.

This is cross_correlogram() of trains with itself, with the zero-lag pairs of a spike with itself removed. Distinct spikes at the same time are still counted at lag 0.

Parameters:

  • trains (SpikeTrains) – Spike trains with N neurons.

  • bin_size (Any) – Width of the lag bins, static and positive. A plain number is in milliseconds.

  • max_lag (Any) – Largest lag of interest, static and non-negative. A plain number is in milliseconds.

Return type:

Tuple[Array, Any]

Returns:

  • counts (jax.Array) – int32 pair counts of shape (N, n_bins) with n_bins = 2 * ceil(max_lag / bin_size) + 1; symmetric about the centre bin.

  • lags (jax.Array or Quantity) – Bin centres, shape (n_bins,).

Raises:

See also

cross_correlogram

Notes

Equal to elephant.spike_train_correlation.cross_correlation_histogram of a binned train with itself minus the spike count in the centre bin, for grid-aligned spikes. Works under jax.jit and jax.vmap; not differentiable.

Examples

>>> import jax.numpy as jnp
>>> from braintools.spike import SpikeTrains
>>> from braintools.analysis.spike import auto_correlogram
>>> raster = jnp.zeros((10, 1)).at[jnp.array([1, 3, 4]), 0].set(1)
>>> st = SpikeTrains.from_dense(raster, dt=1.)
>>> counts, lags = auto_correlogram(st, bin_size=1., max_lag=3.)
>>> counts
Array([[1, 1, 1, 0, 1, 1, 1]], dtype=int32)
>>> lags
Array([-3., -2., -1.,  0.,  1.,  2.,  3.], dtype=float32)