You are here
Usage: cluster_prob clusterSize area t_val -[sld | fwhm] val For implementation details, see cluster_prob. "clusterSize" is the "Min Cnt" of Stimulate's cluster option 2D4N, (2 Dimensional - 4 Neighbor) "area" is either the entire image volume: Width x Height x Planes, or the size of a ROI which is given in Stimulate's Timecourse Full Plot summary window. "t_val" is Stimulates's "T-test threshold", not the Probability Confidence! "val" is the image smoothness. It is recommended that the "fwhm" option be given for use of the theoretical smoothness value. Then the "val" is the "Image Space" gaussian FWHM which/is equal to:
0.44 * (N / r_k)
where N is the image width and r_k is the radius of the
K-space Gaussian LPF.
Here is the source code: cluster_prob.c.