vbperminfo

vbperminfo examines a directory full of stat cubes and creates a distribution for use in a permutation test. The options for what distribution to tally are listed below in the help text:

VoxBo vbperminfo (v1.8.5/Jun 14 2011)
summary: display information about permutation testing results
usage:
  vbperminfo -c <dir> <outfile> <thresh> <alpha> <mask>
        creates a distribution of max cluster size
  vbperminfo -p <stem> <outfile> <alpha>
        creates a distribution of peak values
  vbperminfo -u -t <truevol> -i <iterationdir> -o output
  vbperminfo -r <file> [alpha]
        calculates criteria for an existing ResultValues.ref
  vbperminfo -k <stem> <outfile> <clustersize> <alpha>
        creates a distribution for cluster-corrected thresholds
  vbperminfo -d <dist> <stat value>
  vbperminfo -dr <dist> <stat value or ref>

other usage (these options may be outdated)
  vbperminfo -pm [-a <alpha>] [-m <mask>] [-o <outfile>] <dir> <dir> ...
        creates a distribution for peak minimum values across multiple maps
        the result is stored in outfile and sampled at alpha (default 0.05)

notes:
  The -p and -u options will accept a directory, filename, or stem as
  the data source.  A stem is just the part of the filename shared by
  all the permutation volumes.  E.g., "permdir/cube" matches all files
  in permdir that begin with "cube".  The -c and -k options currently
  require a stem (will be fixed soon).

  The -pm option can be slow and use up a lot of memory.  Only use it
  if you're sure you need it.

  The -u option calculates a separate permutation test in each voxel
  (using the same set of permutations) and produces three maps (in a
  single 4D volume).  The first is the p map, the second is a map of
  the original stat values, and the third contains Z scores (converted
  from the p values).

  The -d option takes a vector file that contains a distribution of
  values, and a stat value of interest, and returns a p value
  corresponding to the proportion of values in the distribution
  meeting or exceeding that value.  -dr returns the proportion of
  values <= the value you pass.

  The -r option allows you to examine existing ResultValues.ref files,
  trying out different alpha values.

  Note that the reported thresholds must be *exceeded* -- i.e., if the
  results file tells you that an alpha of 0.05 requires a threshold of
  t=3.17, voxels with a value of t=3.17 do not meet this standard.
  This is particularly important in cases where the permutation
  distribution may have duplicate values.