This repository is divided into two sets of scripts. Preprocessing takes raw diffusion data through preprocessing, DTI, and NODDI model fits. Postprocessing includes comparative analyses aimed at evaluating the relative utility of 14 different diffusion metrics for developmental neuroimaging. Postprocessing also includes MAPL fits. Further detail below.
Raw DWI -> Quality assurance, motion correction, phase-encoding distortion correction, eddy current correction, coregistration to structurals, template, NODDI scalars via AMICO framework
Runs QA independent of preproc pipeline
Combines subject-level QA into one csv
rounds b-value files to multiple of provided integers (for Siemens scanners)
AMICO (Daducci et al., 2015) framework adapted for current dataset/file structure
value of 1 x number of TRs for eddy input
series of quality assurance scripts from www.med.upenn.edu/cmroi/qascripts.html, requires fsl and afni
just runs AMICO(Matlab) on an SGE
generates file with basic scan sequence info wrap_MultiShell_PreProc For parallel job submission to SGE
FSL (fsl.fmrib.ox.ac.uk/fsl/fslwiki), AFNI (afni.nimh.nih.gov/), NODDI via AMICO (github.com/daducci/AMICO_matlab), ANTs (stnava.github.io/ANTs/), mrtrix (www.mrtrix.org/). For more information on whose work we were heavily dependent on, please reference our manuscript (biorxiv.org/content/10.1101/611590v2).
generates r^2 difference maps from voxelwise (CRAN.R-project.org/package=voxel) outputs
subject-level squareformed (vectorized) tractographyt values into one csv
Brings subject scalars to standard space, needed for voxelwise analyses
correlates scalars maps within subjects within a white matter mask, obtains a mean correlation for each pair of scalars for each subject
single-shell iteration
deterministic tractography utilized
run generalized additive models on each edge (possible streamline b/w two ROIs) for age effects
run generalized additive models on each edge (possible streamline b/w two ROIs) for image quality effects
fits mapl model (Fick et al., 2016) to data
fits mapl model with bvalue extrapolation
comprehensive r markdown of analyses ran and figures created for Developmental Cognitive Neuroscience Manuscript
vectorizes structural connectivity matrices
uses AFNI to get mean white matter values for each diffusion metric of interest
wrapper for parallel job submission of MAPL fits
wrapper for parallel job submission of standard space affines + warps
wrapper for parallel job submission of scalar spatial correlations
wrapper for parallel job submission of deterministic tractography
dipy (dipy.org), camino (camino.cs.ucl.ac.uk/), and a ton of R packages including mgcv, visreg, ggplot2, scales, parallel, corrplot, gratia, dplyr, svglite, cowplot, oronifti, and voxelwise.