This is the code for paper: Multimodal connectivity-based individualized parcellation and analysis for humans and rhesus monkeys Now published on IEEE Transcations on Medical Imaging: https://ieeexplore.ieee.org/document/10508267
MCIP simultaneously optimizes a single subject’s within-region homogeneity with the fusion of functional and anatomical connectivity, spatial continuity, and the similarity to a reference atlas.
- Reading GIfTI files in MATLAB: https://www.gllmflndn.com/software/matlab/gifti/
- Reading CIfTI files in MATLAB: https://github.com/Washington-University/cifti-matlab
- GCoptimization - software for energy minimization with graph cuts: https://github.com/nsubtil/gco-v3.0
Set rfMRI_dir as your rfMRI data directory, dMRI_dir as your dMRI data directory, and out_dir as your output directory in batch_mcip.sh.
Revise ref_atlas_file, LUT_file, and neighbor_file in batch_mcip.sh, if you want to use another reference atlas. neighbor_file can be automatically generated during the program running.
Run the code with a batch of subjects using Slurm: sbatch -a 1-${N} batch_mcip.sh $sub_list, in which $sub_list is a .txt file containing subject ID.
If you do not use Slurm, please change sub_num=... to be a for loop.
MCIP-derived individualized parcellations based on Glasser atlas and Brannetome atlas for HCP subjects is available on https://drive.google.com/file/d/1H0mV6Z4icdO9QSda7lPx0TgCO_1c_Dm8/view?usp=drive_link
Please cite the following paper when using MCIP:
Cui Yue, Li Chengyi, Lu Yuheng, et al. Multimodal Connectivity-based Individual Parcellation and Analysis for Humans and Rhesus Monkeys[J]. IEEE Transactions on Medical Imaging, 2024.