AFNI program: 3dAutoTcorrelate
Output of -help
Usage: 3dAutoTcorrelate [options] dset
Computes the correlation coefficient between the time series of each
pair of voxels in the input dataset, and stores the output into a
new anatomical bucket dataset [scaled to shorts to save memory space].
*** Also see program 3dTcorrMap ***
Options:
-pearson = Correlation is the normal Pearson (product moment)
correlation coefficient [default].
-eta2 = Output is eta^2 measure from Cohen et al., NeuroImage, 2008:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2705206/
http://dx.doi.org/10.1016/j.neuroimage.2008.01.066
** '-eta2' is intended to be used to measure the similarity
between 2 correlation maps; therefore, this option is
to be used in a second stage analysis, where the input
dataset is the output of running 3dAutoTcorrelate with
the '-pearson' option -- the voxel 'time series' from
that first stage run is the correlation map of that
voxel with all other voxels.
** '-polort -1' is recommended with this option!
-spearman AND -quadrant are disabled at this time :-(
-polort m = Remove polynomical trend of order 'm', for m=-1..3.
[default is m=1; removal is by least squares].
Using m=-1 means no detrending; this is only useful
for data/information that has been pre-processed.
-autoclip = Clip off low-intensity regions in the dataset,
-automask = so that the correlation is only computed between
high-intensity (presumably brain) voxels. The
mask is determined the same way that 3dAutomask works.
-mask mmm = Mask of both 'source' and 'target' voxels.
** Restricts computations to those in the mask. Output
volumes are restricted to masked voxels. Also, only
masked voxels will have non-zero output.
** A dataset with 1000 voxels would lead to output of
1000 sub-bricks. With a '-mask' of 50 voxels, the
output dataset have 50 sub-bricks, where the 950
unmasked voxels would be all zero in all 50 sub-bricks
(unless option '-mask_only_targets' is also used).
** The mask is encoded in the output dataset header in the
attribute named 'AFNI_AUTOTCORR_MASK' (cf. 3dMaskToASCII).
-mask_only_targets = Provide output for all voxels.
** Used with '-mask': every voxel is correlated with each
of the mask voxels. In the example above, there would
be 50 output sub-bricks; the n-th output sub-brick
would contain the correlations of the n-th voxel in
the mask with ALL 1000 voxels in the dataset (rather
than with just the 50 voxels in the mask).
-prefix p = Save output into dataset with prefix 'p'
[default prefix is 'ATcorr'].
-time = Mark output as a 3D+time dataset instead of an anat bucket.
-mmap = Write .BRIK results to disk directly using Unix mmap().
This trick can speed the program up when the amount
of memory required to hold the output is very large.
** In many case, the amount of time needed to write
the results to disk is longer than the CPU time.
This option can shorten the disk write time.
** If the program crashes, you'll have to manually
remove the .BRIK file, which will have been created
before the loop over voxels and written into during
that loop, rather than being written all at once
at the end of the analysis, as is usually the case.
** If the amount of memory needed is bigger than the
RAM on your system, this program will be very slow
with or without '-mmap'.
** This option won't work with NIfTI-1 (.nii) output!
Notes:
* The output dataset is anatomical bucket type of shorts
(unless '-time' is used).
* Values are scaled so that a correlation (or eta-squared)
of 1 corresponds to a value of 10000.
* The output file might be gigantic and you might run out
of memory running this program. Use at your own risk!
++ If you get an error message like
*** malloc error for dataset sub-brick
this means that the program ran out of memory when making
the output dataset.
++ If this happens, you can try to use the '-mmap' option,
and if you are lucky, the program may actually run.
* The program prints out an estimate of its memory usage
when it starts. It also prints out a progress 'meter'
to keep you pacified.
* This is a quick hack for Peter Bandettini. Now pay up.
* OpenMP-ized for Hang Joon Jo. Where's my baem-sul?
-- RWCox - 31 Jan 2002 and 16 Jul 2010
=========================================================================
* This binary version of 3dAutoTcorrelate is NOT compiled using OpenMP, a
semi-automatic parallelizer software toolkit, which splits the work
across multiple CPUs/cores on the same shared memory computer.
* However, the source code is modified for OpenMP, and can be compiled
with an OpenMP-capable compiler, such as gcc 4.2+, Intel's icc, and
Sun Studio.
* If you wish to compile this program with OpenMP, see the man page for
your C compiler, and (if needed) consult the AFNI message board, and
http://afni.nimh.nih.gov/pub/dist/doc/misc/OpenMP.html
++ Compile date = May 2 2012
This page auto-generated on
Thu May 3 04:28:08 EDT 2012