This repository contains the instructions & code to construct structural connectivity matrices and calculate network controllability and covers the analyses in the follwoing manuscript:
Drug-resistant focal epilepsy in children is associated with increased modal controllability of the whole brain and epileptogenic regions by Chari et al.
The structural imaging (3T T1 MPRAGE sequences) were organised in BIDS format and fed through Connectome Mapper 3. We chose the Lausanne Atlas at Scale 3 for this study.
Multi-shell multi-tissue diffusion imaging was preprocessed using standard preprocessing steps according to the script diffusion_preproc.sh. The multi-shell multi-tissue dicom files were stored in a folder called 'dwi' and the negative phase encoded series in a folder called 'negPE'. All tools were mrtrix in line with their recommended steps.
This was done by combining the parcellation scheme with the preprocessed diffusion imaging according to the script make_connectome.sh. Firstly the structural data was registered to diffusion space and the cortical target labels were amended according to Scale3_NewLabels.txt. For this, we fused the hippocampal subparcels into a single label, resulting in 253 cortical, subcortical and brainstem parcels per subject. The connectome used anatomically constrained tractography and the SIFT2 filtering step. The output is a 253x253 weighted, undirected structural connectivity matrix in .csv format.
The extracted connectomes were assembled as structures in Matlab v2020b using the scripts ExtractConnectomes.m and LoadConnectomes.m. Following this, the weighted degree, average and modal controllabilities for each node in each patient were calculated using the script CalculateControllabilities.m. These final data files with the connectivity matrices and controllabilities for each subject is avaiable within each group folder (Controls, Patients, PostopPatients and VNS).
The main analyses are encompassed by the file MainAnalyses.m. This file calculates the ranks, correlation coefficients (WD-AC, WD-MC and AC-MC), Z-scores etc that encompass all the patient data analysis elements of the manuscript. The GraphTheory.m file follows on from this and contains the script that calcualtes the graph theory metrics. It requires the BrainConnectivityToolbox to be installed.
GLMs were conducted by manually exporting the data to SPSS v24.
The modelling elements are contained in the folder Modelling. The code for first modelling exercise, adding random edges to the controls, is contained in ModellingEx1.m. The code for the second modelling exercise of adding low edge weights to thalamocortical edges is contained in ModellingEx2.m.
Elements required for visualising the atlas are contained in the folder SurfIce and requires dowloading SurfIce and the AtlasStatMap function. The folder also contains a version of the atlas in SurfIce mz3 format (merge.mz3). Use the code included to display any assigned values to the parcels. For more information, please see the SurfIce Wiki on viewing atlases.