AFNI program: @simulate.motion
Output of -help
---------------------------------------------------------------------------
@simulate.motion - create simulated motion time series
This program is meant to simulate an EPI time series based only on the
motion parameters and an input volume.
The main action is to take the EPI (motion base) volume and warp it
according to the motion parameters. In theory, the result could be run
through 3dvolreg to generate a similar set of motion parameters.
Note: if slice timing is provided (via the -epi or -epi_timing datasets),
then slices will be generated individually at the interpolated offset
into each TR.
Purpose:
The resulting time series can be used to create regressors of no
interest, when trying to regress out motion artifacts (from either
task or resting state analysis). Ways it can be used:
a. Grab the first N (e.g. 6) principle components, and use them along
with other motion paramters. To do this, just run 3dpc with the
simulated time series and an appropriate mask.
b. First make the time series orthogonal to the motion parameters, and
only then take the first N principle components. For example, run
3dDeconvolve to remove the original motion parameters, and use the
resulting errts dataset as input to 3dpc.
c. Akin to ANATICOR, use locally averaged time series via 3dTfitter.
Run 3dLocalstat to generate a new volumetric time series from local
averages.
Alternatively and more easily, just blur the simulated time series.
d. Use the time series as is, using 3dTfitter.
Note that if censoring is being done, such TRs would have to be
removed, as 3dTfitter does not have a -censor option.
i) run '1d_tool.py -show_trs_uncensored ...' and extract those TRs
ii) pass the X-matrix and this time series to 3dTfitter
Eventually these methods will be put into afni_proc.py. Please pester
Rick if you have interest in any method that has not been implemented.
usage: @simulate.motion [options] -epi EPI_DSET -motion_file MOTION_PARAMS
needed inputs: EPI volume, motion parameters
output: motion simulated EPI time series
examples:
@simulate.motion -epi pb02.FT.r01.volreg+tlrc -motion_file dfile.r01.1D
@simulate.motion -epi pb02.FT.r01.volreg+tlrc"[2]" -motion_file dfile_rall.1D
@simulate.motion -epi pb02.FT.r01.volreg+tlrc -motion_file dfile_rall.1D \
-epi_timing pb00.FT.r01.tcat+orig -prefix sim.mot.FT
informational options:
-help : show this help
-hist : show program modification history
-ver : show program version
required parameters:
-epi EPI : provide input volume or time series
(only a volreg base is needed, though more is okay)
If slice timing is to be used, the number of slices
must match that of the -epi_timing dataset. So it
should not be the case where one view is +orig and
the other +tlrc, for example.
-motion_file MOTFILE : specify motion parameter file (as output by 3dvolreg)
options:
-epi_timing DSET : provide EPI dataset with slice timing
(maybe -epi no longer has slice times)
-save_workdir : do not remove 'work' directory
-prefix PREFIX : prefix for data results
(default = motion_sim.NUM_TRS)
-vr_base INDEX : 0-based index of volreg base in EPI dataset
-test : only test running the program, do not actually
create a simulated motion dataset
-verb LEVEL : specify a verbose level (default = 1)
WILL BE IMPLEMENTED SOON:
-warp_1D : specify a 12 parameter affine transformation,
presumably to go from orig space to standard space
e.g. -warp_1D mat_rall.warp.aff12.1D
This command must be paired with -warp_master.
-warp_master DSET : specify a grid master dataset for the -warp_1D xform
e.g. -warp_master pb02.FT.r01.volreg+tlrc
This DSET should probably be one of the volreg+tlrc
results from an afni_proc.py script.
-------------------------------------------------------
R Reynolds May, 2013
---------------------------------------------------------------------------
This page auto-generated on
Fri Oct 19 17:22:23 EDT 2018