Spike Toolbox

STCrossCorrelation
Calcualte cross-correlation for instantiated spike trains

Syntax
[vCorrSmoothed, vCorrRaw, vtCorrTime] = ... STCrossCorrelation(stTrain1, stTrain2 <, tWindow, strKernel, tSmoothing>)

Description
STCrossCorrelation calculates the cross-correlation of two spike trains. Any mapping information is ignored; this function is designed to work with spike train instances. STFlatten and STExtract can be used to convert mapped trains to spike train instances. stTrain<1, 2> are valid spike train objects, which must be at least instantiated. Spike train definitions are not supported by STCrossCorrelation.

tWindow is an optional argument to specity the time range over which to calculate the cross correlation. The correlation will stretch over [-tWindow..tWindow]. tSmoothing is an optional argument that specifies the amount of smoothing to perform on the calculated correlation. A gaussian smoothing will be performed with a window width of tSmoothing. If tSmoothing is zero, smoothing will not be performed. Both time window parameters are in seconds. strKernel can optionally be used to specify the kernel used for smoothing the correlation function. Recognised kernels are gaussian, square and none. The defaults for these arguments can be set using STOptions.

vCorrRaw will be the raw calculated cross correlation. vCorrSmoothed will be the (optionally) smoothed correlation trace. vtCorrTime will be a vector of the corresponding time bins for each entry in vCorr.... If no output arguments are supplied, the (smoothed) correlation will be plotted in the current figure.

To perform an autocorrelation, just supply the same spike train twice. In this case, the zero point will be ignored for smoothing and will be replaced by a NaN.

$Id: STCrossCorrelation.html 1783 2005-08-31 21:05:51Z dylan $