Read the paper: Gumus, M., A. Bourganos, and M. L. Mack. 2025. In Vivo Quantification of White Matter Pathways in the Human Hippocampus. Human Brain Mapping 46, no. 17: e70417. https://doi.org/10.1002/hbm.70417.
HippCircuit is a resource for in vivo quantification of human white hippocampal white matter connections based on diffusion weighted imaging and various processing software. The processing pipeline includes 2 arms: a) quantification of white connections at the individual level, b) template creation based on the entire sample.
This processing pipeline was developed on the Human Connectome Project Young Adult Sample (N=831). Each participant's white matter connectivity matrices are available to download from this resource to answer various multimodal questions, leveraging the Human Connectome Project data. To explore the data, please check out our HippCircuit website! The pipeline can be applied to a sample of interest with certain modifications.
Scripts are organized into separate folders according to their specific roles in the pipeline:
- Hippocampal subfield segmentations:
maget - Entorhinal cortex segementations:
itk-snap - Quantification of individual white matter connections (a):
mrtrix - Creation of the template based on the entire sample (b):
fixel-based - Label files for the segmentations:
labels - Statistical analyses of connectivity matrices:
analyses
This pipeline was developed on Compute Canada's resources at SciNet, but can be adapted for other supercomputers. Most software is available as modules at SciNet. Python.2.7 was compatible with the software versions implemented here.
Software necessary for the processing pipeline:
- BIDS
- MRtrix3
- FSL
- Freesurfer
- MAGeTbrain
- ASHS
- ITK-SNAP
Data should be in BIDS format. Dicoms can be converted to BIDS format using Dcm2Bids.
To install & create basic BIDS files and directories:
pip install dcm2bids
dcm2bids_scaffold
MRtrix provides advanced tools for processing of diffusion weighted imaging. The dependencies can be loaded as the following:
module load NiaEnv/2018a
module load intel/2018.2
module load ants/2.3.1
module load openblas/0.2.20
module load fsl/.experimental-6.0.0
module load fftw/3.3.7
module load eigen/3.3.4
module load mrtrix/3.0.0
If using supercomputers other than SciNet, please follow the instructions on the MRtrix website for software installation.
FSL is loaded as part of the dependencies above. However, it can also be loaded separetely:
module load fsl/.experimental-6.0.0
The ideal version of Freesurfer 7.1. was not available in NiaEnv/2018a. Thus, we are using 6.0.0:
module load freesurfer/6.0.0
Hippocampus was segmentated into its subfields using MAGeTbrain. The CoBrALab has graciously given us access to their software stack on Niagara. If your Compute Canada account is associated with the mmack group, you will have access. You can activate MAGeTbrain as the following:
module load cobralab/2019b
source /project/m/mmack/software/MAGeTbrain/bin/activate
Entorhinal cortex was segmented using ASHS. Although not leveraged in this dataset, ASHS also provides hippocampal subfield segmentation. The segmentation software is user's choice. As long as there is no overlap between the segmentatations within an image, it will be compatible with the pipeline.
If you choose to use the cloud services of ASHS, you will also need to download ITK-SNAP. Users can submit their structural images to Distributed Segmentation System (DSS) for automatic segmentation.
If you are using any part of this resource, please cite us:
Gumus, M., A. Bourganos, and M. L. Mack. 2025. In Vivo Quantification of White Matter Pathways in the Human Hippocampus. Human Brain Mapping 46, no. 17: e70417. https://doi.org/10.1002/hbm.70417.