HCP Pipeline

HCP pipelines

• A commonly used series of neuroimaging pipelines built around FreeSurfer, FSL and wb_command

• Designed for HCP-like data, good for processing multi-band images

• Steps are explained here. In brief:

  • It removes the skull from images using FSL's brain extraction tool

  • It aligns T1w, T2w, and fMRI using FLIRT - linear imaging registration tool - and BBRegister

  • It performs bias field correction on structural images to remove gradients of voxel intensity difference

  • It realign fMRI to each individual's first volume using FLIRT

  • It registers structural images to MNI152 space usinf FLIRT and FNIRT - non-linear imaging registration tool

  • It coregisters fMRIs to individual's T1w in native and MNI152 (nonlinear asymmetric template)

  • It intensity normalises the fMRI data to the value of 1000

  • Denoises images with FSL-FIX

BIDS App

The original HCP Pipeline does not take BIDS as an input, but there is a BIDS App wrapper to use the HCP Pipeline with data in the BIDS format

We have a Singularity container for the BIDS App on MASSIVE at /projects/kg98/Priscila/honours/script/bids-hcppipelines--4.3.0-3.sing. This runs the HCP Pipline up to ICA-FIX. A sample script used to run the Singularity on the cluster can be foun at /projects/kg98/Priscila/honours/script/sbatch_hcp.sh

Note that this App does not include denoising, please refer to ICA-FIX for more information.