neurotools.spikes package

Submodules

• neurotools.spikes.convolutional module
  • ccor()
  • ccm()
  • blockccm()
  • sta()
  • blocksta()
  • reconstruct()
  • cspect()
  • cspectm()
  • spike_lfp_filters()
  • spectreconstruct()
  • create_spectral_model()
  • construct_lowpass_operator()
  • autocorrelation_bayes()
• neurotools.spikes.ppc module
  • phase_randomize()
  • fftppc_biased()
  • fftppc()
  • fftppc_biased_multitaper()
  • fftppc_multitaper()
  • discard_spikes_closer_than_delta()
  • pairwise_phase_consistancy()
  • estimate_bias_in_uncorrected_ppc()
  • phase_randomized_bias_correction()
  • ppc_chance_level()
  • ppc_phase_randomize_chance_level_sample()
• neurotools.spikes.sta module
  • get_snips()
  • triggered_average()
  • get_triggered_stats()
  • fftsta()
• neurotools.spikes.waveform module
  • realign()
  • getFWHM()
  • getPVT()
  • getWAHP()
  • getPT()
  • getPTHW()
  • getPHP()
  • normalized_waveform()
  • is_thin()
  • is_thin_pvt()
  • process()

Module contents

Signal processing routines related to spike trains.

neurotools.spikes.pp_xcorr(t1, t2, maxlag)

Gets all pairwise relative time lags within maxlag. pp_xcorr stands for point-process cross correlation.

Parameters:

• t1 (1d arrays) – Lists of time points to compare

• t2 (1d arrays) – Lists of time points to compare

• maxlag (number) – maximum time lag to consider, in the same units as t1 and t2

neurotools.spikes.txcorr(t1, t2, maxlag, k=100, normalize=False, sampleat=None)

Computes cross correlation between two spike trains provided in terms of spike times, over a maximum range of lags. Uses nearest- neighbor density estimation to provide an adaptively smoothed cross-correlation function.

Parameters:

• t1 (1d arrays) – Lists of time points to compare

• t2 (1d arrays) – Lists of time points to compare

• maxlag (number) – maximum time lag to consider, in the same units as t1 and t2

• k (positive integer) – number of nearest neighbors to use in the density estimation

• normalize (boolean) – Normalize correlation by zero-lag correlation. Default False

• sampleat (int) – time lags to sample for the density estimation defaults to spanning +-lags with 1 time-unit bins if none

neurotools.spikes.pack_cross_correlation_matrix(xc)

Convert a 1D autocorrelogram into a two-timepoint correlation matrix.

This function is obsolete, use scipy.linalg.toeplitz instead.

Parameters:

xc (1D np.array)

Returns:

matrix

Return type:

2D np.array

neurotools.spikes.cut_spikes(s, cut)

Downsample a spike raster s by factor cut by symming adjacent bins. This may generate counts >1 when two spikes occur within the same downsampled bin.

Parameters:

• s (1D np.array) – Array of spike counts

• cut (positive int) – Downsampling factor

Returns:

downsampled

Return type:

np.array

neurotools.spikes.times_to_raster(spikes, duration=1000)

Convert spike times to a spike raster

Parameters:

• spikes (np.int32) – List of spike times as non-negative array indecies

• duration – Duration of the desired raster

Returns:

result

Return type:

1D np.array with length duration

neurotools.spikes.bin_spikes_raster(train, binsize=5)

Important! This accepts a spike raster, not spike times!

Parameters:

• train

• binsize

neurotools.spikes.bin_spike_times(times, binsize=5)

Important! This accepts spike times, not a raster

Parameters:

• train

• binsize