UK Biobank: Predict functional connectivity from white matter microstructure in the brain

The spatial correspondence and genetic influence of inter-hemispheric connectivity with white matter microstructure.
Jeroen Mollink, Stephen M. Smith, Lloyd T. Elliott, Michiel Kleinnijenhuis, Marlies Hiemstra, Fidel Alfaro-Almagro, Jonathan Marchini, Anne-Marie van Cappellen van Walsum, Saad Jbabdi, Karla L. Miller. 

Contact: mollink.jeroen@gmail.com


Description: 
Functional connectivity is estimated from resting-state fMRI data between homotopic region pairs. ICA spatial maps are estimated with FSL's MELODIC tool to define these homotopic regions. Probabilistic tractography was run between a pair using diffusion MRI data to delineate the white matter tract connecting them. Microstructure metrics (derived from diffusion MRI) were obtained from the white matter tract connecting a homotopic pair. Tract-based spatial statistics (TBSS) was used to project the microstructure metrics onto a white matter skeleton. For a given homotopic pair, a GLM was constructed to predict functional connectivity from the microstructure of the white matter skeleton voxels connecting them. 

 
############################# #############################
###################### Prerequisites ######################
############################# ############################# 
# Software
 - Matlab r2012a or later (Mathworks Inc)
 - FMRIB's Software Library (FSL): https://fsl.fmrib.ox.ac.uk/
 - FSLNets Matlab toolbox: https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FSLNets

# Data
Pre-processed data for each subject in UK Biobank (not included here), see F. Alfaro-Almagro et. al. 2018, Neuroimage for processing steps. For data access, visit http://www.ukbiobank.ac.uk/register-apply.
 - Motion corrected and cleaned timeseries data (resting-state fMRI).
 - Microstructure parameter maps: MD, MO, FA, ICVF, ISOVF, OD (diffusion MRI). 
 - Microstructure projected onto a TBSS white matter skeleton.
 - Covariates for each subject (age, age^2, sex, age*sex, age^2*sex, resting-state fMRI head motion, and head size)
 

 
############################# #############################
###################### Data overview ######################
############################# ############################# 

~/Biobank/Data/rfMRI:
 - melodic_IC.nii.gz				ICA spatial maps that were split between the left and right cerebral hemispheres.
 - rfMRI_IC_cluster_LR_goodcomp.txt		Nodes corresponding to homotopic region pairs.
 - rfMRI_IC_cluser_LR_nonusedcomp.txt		Non-used (but not noise) components. For example cerebellar or sub-cortical nodes.
 
 
~/Biobank/Data/dMRI:
 - TBSS_mask.nii.gz 				TBSS white matter skeleton mask.
 - tbss_tdt.nii.gz				TBSS voxels connecting a corresponding homotopic region pair (derived from tractography).
 - seeds_ICA.nii.gz				Binarized ICA spatial maps (good comps only) corresponding to all homotopic regions. Used as seed and waypoints for probabilistic tractography (not included here).

~/Biobank/Data/Model:
 - beta.mat					Regression coefficients for each microstructure function model. However, unless the exact same subject cohort is used as in the paper, it is likely that these regression coefficients are meaningless due to the principal components analysis that was run on each microstructure matrix before feeding it in the GLM. The principal components analysis yields components specific to that cohort. 
 - TVE.mat					Total variance explained by each model.




############################# #############################
######################### Scripts #########################
############################# ############################# 

# Scripts (execute in this order):
 - rfMRI_get_timeseries.sh			Generate timeseries from the ICA spatial maps at single-subject level. 
 - rfMRI_processing.m 				Clean timeseries and estimate functional connectivity .
 - dMRI_get_micro.m 				Extract microstructure metrics from white matter skeleton voxels (TBSS).
 - micro_func_model.m				Estimate functional connectivty from white matter microstructure.