Performs FSL's dual_regression for CIFTI files in addition to permutation based analyses (via FSL's PALM).
This script requires that FSL, connectome workbench (wb_command), FSL's PALM, and GNU parallel to be installed and added to the system path for this script to work correctly.
Note:
- Intended to run on LSF platforms (i.e. jobs for PALM are submitted via
bsubto run in parallel).
Usage: dual_regression_cifti.sh -i <image> -o <out_dir> -f <list.txt> -tsL <left_surf_template> -tsR <right_surf_template>
Performs FSL's dual regression for CIFTI files in addition to permutation based analyses.
Required arguements:
-i, -ics, --ica-maps Input CIFTI dscalar IC maps
-f, -files, --file-list Text file list of CIFTI dtseries subject files
-o, -out, --out-dir Output directory
-tsL, --template-surf-L Template midthickness (left) surface
-tsR, --template-surf-R Template midthickness (right) surface
Optional arguements:
-j, -jobs, --jobs Number of jobs that can be run in parallel.
The default varies per system and is defined as
N-1 the maximum number of cores available. [default: 11]
-sL, --surf-list-L Text file list of subject left midthickness surface files
-sR, --surf-list-R Text file list of subject right midthickness surface files
--atlas-dir Surface template atlas directory. If specified, then the
'--template-surf-L' and '--template-surf-R' options do not
need to be specified. Atlas directory name and layout are expected
to be similar to that of the HCP S1200 fs_LR atlas(es).
--no-stats-cleanup No clean-up of PALM's IC map sub-directory will be done
--convert-all Converts all output files from (fake) NIFTI-1 images to CIFTI (recommended)
Dual Regression specific arguements:
-des, --des-norm Whether to variance-normalise the timecourses used as the stage-2 regressors (recommended)
-d, -design, --design Design matrix for final cross-subject modelling with PALM
-c, -contrast, --t-contrast Design t-contrast for final cross-subject modelling with PALM
-f, --f-contrast Design F-contrast for final cross-subject modelling with PALM
-n, -nperm, --permutations Number of permutations for PALM. If not set or set to 0, then
permutations are performed exhaustively.
--thr Perform thresholded dual regression to obtain unbiased timeseries for connectomics
analyses (e.g., with FSLnets)
PALM specific arguements:
--fdr Produce FDR-adjusted p-values.
--f-only Run only the F-contrasts, not the t-contrasts.
--save1-p Save (1-p) instead of the actual p-values (mutually exclusive with '--log-p').
--log-p Save the output p-values as -log(p) (mutually exclusive with '--save1-p', recommended).
--two-tail Run two-tailed tests for all the t-contrasts instead of one-tailed.
--demean Mean center the data, as well as all columns of the design matrix. If the design has
an intercept, the intercept is removed.
--sig Significance threshold for statistical thresholding [default: 0.05]
--method Method used for determining corrected cifti-stat threshold.
Valid options include: Bonferroni ('bonf') and Šidák ('sid') [default: bonf]
LSF specific arguements:
--mem, --memory The amount of memory to be used when submitting jobs for PALM (in MB) [default: 5000]
--wall The amount of wall-time to be allocated to each job for PALM (in hours) [default: 100]
-q, --queue LSF queue name to submit jobs to (, look up queue names with the command 'bqueues') [default: normal]
----------------------------------------
-h,-help,--help Prints usage and exits.
NOTE:
- Requires FSL v5.0.11+
- Requires Connectome Workbench v1.3.2+
- Requires FSL's PALM version alpha115+
- Requires GNU parallel to be installed and
added to system path
- Default LSF arguements are unlikely to result
in all PALM jobs running to completion.
----------------------------------------
Adebayo B. Braimah - 2020 02 18 17:31:22
dual_regression_cifti.sh v0.0.1
----------------------------------------
Usage: dual_regression_cifti.sh -i <image> -o <out_dir> -f <list.txt> -tsL <left_surf_template> -tsR <right_surf_template>