vbtmap

vbtmap is a command-line program for creating t (and z) maps, mostly for VLSM studies. Here's the help text:

VoxBo vbtmap (v1.8.5pre5/Oct 26 2009)
usage:
  vbtmap <4D file> <1D file> <outfile> <flags>
flags:
  -op <mat> <ind>    order permutation
  -m <maskfile>      specify inclusion mask file
  -n <min>           minimum number of lesions for inclusion (default:2)
  -q <q>             calculate FDR thresh for q
  -nodup             omit duplicate voxels from FDR calculation
  -f                 flip sign of stat values
  -2                 use two-tailed test when calculating ps
  -w                 use welch's t-test (unequal variances)
  -z                 convert t to z
  -h                 show help
  -v                 show version
notes:
  By default, vbtmap calculates t values for lesion minus no-lesion.
  I.e., a positive t or z means that patients with a lesion in that
  voxel had higher scores.  So you would typically expect positive
  values for error or response time measures, for example.  Use the -f
  flag to flip it.  p values are one-tailed.

  The minimum number of lesions can't be set to less than 2, because
  vbtmap is designed to carry out a 2-sample t-test.

  The -nodup flag allows groups of identical voxels (i.e., voxels
  lesioned in exactly the same set of patients) to be counted only
  once for the FDR comparison.  I'm not sure if this is a good, bad,
  or indifferent idea.  But it's important to consider that when you
  do this, the FDR rate applies to regions (contiguous or otherwise)
  of identical voxels, not to voxels per se.