UK Biobank: Brain imaging protocols and first data release
Karla L Miller1, Neal K Bangerter2, Fidel Alfaro Almagro1, David L Thomas3, Essa Yacoub4, Junqian Xu5, Andreas J Bartsch1,6, Saad Jbabdi1, Stamatios N Sotiropoulos1, Mark Jenkinson1, Jesper Andersson1, Ludovica Griffanti1, Peter Weale7, Iulius Dragonu7, Steve Garratt8, Sarah Hudson8, Rory Collins8,9, Paul M Matthews10, and Stephen M Smith1

1FMRIB Centre, University of Oxford, Oxford, United Kingdom, 2Electrical and Computer Engineering, Brigham Young University, Provo, UT, United States, 3Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, University College London, London, United Kingdom, 4Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, MN, United States, 5Icahn School of Medicine at Mount Sinai, New York, NY, United States, 6Department of Neuroradiology, University of Heidelberg, Heidelberg, Germany, 7Siemens Healthcare (UK), London, United Kingdom, 8UK Biobank Ltd, Stockport, United Kingdom, 9Nuffield Department of Population Health, University of Oxford, Oxford, United Kingdom, 10Department of Medicine, Imperial College London, London, United Kingdom


UK Biobank is a prospective epidemiological study of 500,000 participants consisting of extensive questionnaires, physical measures and biological samples, linking to long-term health outcomes. The imaging extension for the UK Biobank ultimately aims to image 100,000 subjects from this cohort, including brain, cardiac and body MRI, bone scans and carotid ultrasound. We overview the brain imaging component, which includes structural, functional and diffusion MRI. The value of this open resource arises not only from multi-modal/multi-organ imaging, but also from the depth of other demographic, phenotypic and exposure data, and will increase over time as clinical outcomes are realized in the population.


To present the brain imaging protocols included in UK Biobank, a large prospective epidemiological study, and provide an initial overview of data quality.


UK Biobank is a prospective epidemiological study of 500,000 participants consisting of extensive questionnaires, physical measures and biological samples, including genetics1. The study will link against UK National Health Service records to relate these measures to long-term health outcomes. All data from UK Biobank are available to researchers world-wide upon application for access.

An imaging extension for the UK Biobank has recently been launched that, if fully funded, ultimately aims to image 100,000 subjects from the original cohort (currently aged 45-78)2. Imaging includes brain, cardiac and body MRI, DEXA bone scans and carotid ultrasound. The pilot phase has scanned 6500 participants3 in 18 months, and produced a data release of 5000 subjects, recently made available to researchers3. Here, we overview the brain imaging component of UK Biobank.

Protocol Overview

Imaging of 100,000 subjects requires three dedicated imaging centres operating 7 days/week with throughput of 18 subjects/day over 5 years. This daily throughput places tight timing constraints, corresponding to one subject completed every 36 minutes. As a prospective study, the UK Biobank does not target any particular disease or hypothesis, and hence the imaging protocol must be as broadly useful as possible.

Following optimization of acquisition protocols, streamlining of subject preparation and minimization of scanner dead time, Biobank is now achieving its target participant throughput. The brain imaging protocol (Table 1) includes T1-weighted MPRAGE (T1) and T2-weighted FLAIR (T2FLAIR), diffusion MRI (dMRI), task-based fMRI (tfMRI), resting-state fMRI (rfMRI) and susceptibility-weighted gradient echo (SWI).

Protocol Considerations

Many considerations went into protocol design. Here, we highlight decisions relating to the high throughput nature of Biobank.

° Tight FOVs minimize scan time, but in large studies these restrictions exclude subjects (e.g. 99% population compatibility loses 1000 participants). We conducted a study of population brain size4 that enables our FOVs to target ≥99.9%.

° Calibration scan times accumulate. By minimizing this “dead time” (e.g. altered shim defaults) we re-gained several minutes scan time.

° While T1 scan times could have been reduced, the central role of the T1 to cross-subject and cross-modal alignment for most processing pipelines made this an unacceptable risk.

° Simultaneous multi-slice (multiband, MB) acquisition enabled short fMRI5 and dMRI6 runs without sacrificing statistical robustness or directions/b-values, respectively.

° Field map acquisitions were replaced with blip-reversed spin echo acquisition as part of dMRI.

° We are piloting a potential protocol change, reducing task fMRI to 2 minutes with the goal of including a short ASL perfusion scan.

Image Processing and Associated Phenotypes:

Automated processing pipelines were developed, primarily based on tools from FSL, with future pipelines expanding the range of toolkits (e.g. FreeSurfer-based cortical and subcortical/cerebellar modeling). A streamlined set of image outputs is available for each modality. Example images for several modalities from the initial 5000 participant data release are shown in Figs 1-4.

Imaging-derived phenotypes (IDPs) were estimated from the processed outputs. IDPs are single values that would be appropriate for use by non-imaging experts. Examples of the extracted IDPs currently include (listed by modality, number of IDPs in parentheses):

° T1: volumetric estimates of tissue types and grey matter structures (33)

° dMRI: tensor- and NODDI-derived estimates in major white matter tracts (675)

° rfMRI: network-edge connectivity estimates for two network decompositions (3,390)

° tfMRI: summary statistics on BOLD effect size and significance values (18)

° SWI: median T2* values in major deep grey structures (14)

The set of available IDPs is expected to expand with improved and novel functionality of processing pipelines.


If funding is awarded for the full study, the UK Biobank will comprise the largest cohort of subjects ever imaged comprehensively. The value of the resource arises not only from multi-modal/multi-organ imaging, but also from the depth of other demographic, phenotypic and exposure data. Discovery of disease risk factors will increase over time as clinical outcomes are realized in the population. For example, 5000 cohort participants are expected to develop Alzheimer’s disease by 2020, rising to 10,000 by 2030.


The UK Biobank Imaging Enhancement is funded by the Medical Research Council UK. The work presented here was also supported by the Wellcome Trust (KLM and SMS). We would like to acknowledge the valuable contributions of members of the UK Biobank Imaging Working Group and the UK Biobank coordinating centre. Finally, we are extremely grateful to all UK Biobank study participants, who have made this study possible.


1. Sudlow, Gallacher, Allen, Beral, Burton, Danesh, Downey, Elliott, Green, Landray, Liu, Matthews, Ong, Pell, Silman, Young, Sprosen, Peakman, Collins (2015). UK Biobank: An Open Access Resource for Identifying the Causes of a Wide Range of Complex Diseases of Middle and Old Age. PLoS Med 12:3.

2. Petersen, Matthews, Bamberg, Bluemke, Francis, Friedrich, Leeson, Nagel, Plein, Rademakers, Young, Garratt, Peakman, Sellors, Collins, Neubauer (2013). Imaging in population science: cardiovascular magnetic resonance in 100,000 participants of UK Biobank - rationale, challenges and approaches. JCMR 15:46.


4. Mennes, Jenkinson, Valabregue, Buitelaar, Beckmann, and Smith (2014). Optimizing full-brain coverage in human brain MRI through population distributions of brain size. NeuroImage, 98: 513.

5. Moeller, Yacoub, Olman, Auerbach, Strupp, Harel, Ugurbil (2010). Multiband multislice GE-EPI at 7 tesla, with 16-fold acceleration using partial parallel imaging with application to high spatial and temporal whole-brain fMRI. Magn Reson Med. 63:1144.

6. Xu, Moeller, Auerbach, Strupp, Smith, Feinberg, Yacoub, Ugurbil (2013). Evaluation of slice accelerations using multiband echo planar imaging at 3 T. NeuroImage. 83 :991

7. Barch, D., Burgess, G., Harms, M., Petersen, S., Schlaggar, B., Corbetta, M., Glasser, M., Curtiss, S., Dixit, S., Feldt, C., Nolan, D., Bryant, E., Hartley, T., Footer, O., Bjork, J., Poldrack, R., Smith, S., Johansen-Berg, H., Snyder, A., Van Essen, D. - for the WU-Minn HCP Consortium (2013). Function in the Human Connectome: Task-fMRI and individual differences in behavior. NeuroImage, 80:169–189.

8. Daducci, A., Canales-Rodríguez, E. J., Zhang, H., Dyrby, T. B., Alexander, D. C., and Thiran, J.-P. (2015). Accelerated microstructure imaging via convex optimization (AMICO) from diffusion MRI data. NeuroImage, 105:32–44.


Table 1. Overview of the UK Biobank brain MRI protocol (31 minutes total scan time). R = in-plane acceleration factor, MB = multiband factor, PF=partial Fourier. All non-EPI scans were pre-scan normalized.

Figure 1. Cross-subject (n~4500) average T1 and T2-FLAIR structural scans. Non-linear registration using FNIRT to the MNI152 space enables excellent quality alignment, with clear delineation of deep grey nuclei and good agreement of sulcal folding patterns despite well established variation in these features across subjects.

Figure 2. Resting-state fMRI analyzed using group-level (n~4200) independent component analysis. Depicted are the two components accounting for greatest variance explained, corresponding to the default mode network (left) and the dorsal attention network (right). Resting-state maps depict excellent localization to sulcal folding patterns, despite being a composite across thousands of subjects.

Figure 3. Task fMRI for the Hariri faces/shapes “emotion” task7. The contrast depicted here is Faces-Shapes, which as expected elicits amygdala activation (group-level fixed-effect z-statistic, threshoded at Z>80).

Figure 4. Group-level diffusion analyses. On left, delineation of large white matter tracts based on diffusion tractography. On right, orientation dispersion (OD) derived from the NODDI model reflecting degree of microstructural order8.

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)