Topup Pipeline

Topup

A big, but crucial challenge for laminar fMRI is the registration to an anatomical space. For that we have to do field map correction. We will use acquisitions with an inverted phase-encoding as compared to the actual BOLD images ("topup", "pepolar", or "blip-up/blip-down").

Possible pipelines

Gilles is currently implementing two candidate core pipelines:

1. A self-implemented workflow that uses FSL's topup, based on Tomas's work. See https://github.com/VU-Cog-Sci/hires_ODC_7T/tree/master/src/workflows/topup

Old version of fmriprep did something similar: https://github.com/chrisfilo/fmriprep/blob/295672888f468577aeeaf1924c0fc7cf42fbada3/fmriprep/workflows/fieldmap/unwarp.py#L221

2. The workflow baed on AFNI 3dQWARP used in fmriprep

Some issues that we want to study:

a) Influence of removing the skull (3dAutomask) b) Influence of removing "Noise slices" deeper in the brain with lots of SENSE artefacts c) When to do motion correction