UK Biobank Brain Imaging - Acquisition Protocol

...back to main FMRIB UK Biobank resource page

Current protocol (v4+v5) - from 23 September 2014

The protocol requires the CMRR multiband sequences.

Some of the protocol details below (e.g., the PDF) are out of date purely with respect to the order of the modalities. The acquired order is: T1, rfMRI, T2FLAIR, dMRI, swMRI, tfMRI, ASL. (ASL was added recently - more information about this to come.)

Protocol PDF

Protocol EDX (importable binary protocol file) and Protocol XML files. These may only import correctly on a Skyra running VD with the CMRR multiband already installed; feedback of experiences on other platforms would be welcome.

Protocol screenshots (single tarfile download and online views of all screenshot images). These may contain some settings not set by looking at the protocol PDF or loading the EDX. We would appreciate it if people could email us details of which settings are not set via the PDF and/or EDX so that we can include that information here.

dMRI directions - main file (the scanner will add on an extra b=0 volume at the start).
Some scanners have problems achieving all gradient directions - in some cases this is easily fixable.

dMRI secondary directions - short acquisition to get blip-reversed b=0 scans for distortion correction.

Fixation cross-hair (for use during all modalities except task-fMRI).

Task-fMRI stimuli and scripts for ePrime (4minutes, courtesy of Nik Bloom and Deanna Barch, WashU/HCP).

Example rfMRI header file containing exact slice timing information (though we do not generally recommend applying slice-timing-correction due to the low TR).

Shimming information: Shim field accuracy is critical for data quality. We recommend that one of the two following options is used: 1) Shim quality can be improved by manually iterating the shimming process 3 times. 2) Alternatively, the scanner "measperm" settings can be changed to reduce the shimming field-of-view, improving resolution of the acquired field map; however this "hard coding" is probably only desirable on scanners that are only running the UK Biobank protocol, and has to be set by Siemens engineers, and so is probably not in general to be recommended. In either case it is important that the scans are prescribed to avoid any subsequent re-shimming during the entire protocol (in particular because EPI distortion correction, derived from the dMRI, is applied also to the fMRI data).

Feedback from other scanners running this protocol

Prisma running VE11C - the above EDX imported fine. All settings appear to have been set correctly, with just the following very minor parameter changes happening:
T1: Rise time went from 10.5; TE increased from 2.01ms to 2.03ms
T2-FLAIR: Echo train duration increased from 876ms to 879ms; TE up from 395ms to 397ms