UK Biobank Imaging
UK Biobank multi-modal imaging enhancement (brain, cardiac and abdominal MRI, DXA, carotid ultrasound)
Brain, heart and abdominal MRI, whole-body DXA and carotid ultrasound of over 100,000 UK Biobank participants aged about 44 to 87 at imaging, with repeat scans for a subset, linked to genetics and health records. Access by application and fee.
Overview
UK Biobank is a prospective cohort of about half a million adults recruited across the UK between 2006 and 2010. Its imaging enhancement, started in 2014, invited participants back for a single long visit with five imaging examinations. It is used for imaging-genetics, brain age, cardiac and body composition research and for training models applied to other cohorts.
Composition
Per the public Data Showcase, 102,521 participants attended a first imaging visit and 21,469 a repeat imaging visit (from 2019) by September 2025. Visits cover brain MRI (T1, T2 FLAIR, susceptibility-weighted, diffusion, resting and task fMRI and arterial spin labelling), cardiac MRI (cine long- and short-axis, tagging, aortic distensibility, flow and T1 mapping), abdominal MRI (neck-to-knee Dixon plus liver, kidney and pancreas sequences), whole-body DXA and carotid ultrasound. Images are released in tranches, so counts per image type lag behind attendance. Thousands of imaging-derived phenotypes accompany the images, and every participant links to the main cohort's genetic, biochemical and health-record data.
Acquisition
Imaging runs in four centres: Stockport, Newcastle, Reading and Bristol. Brain MRI uses a 3T Siemens Skyra with a 32-channel head coil in a 35-minute protocol; cardiac and abdominal MRI share a 1.5T Siemens Aera; DXA is done on a GE iDXA. Brain images come as DICOM and processed NIfTI, with T1 and FLAIR NIfTI defaced; undefaced DICOM needs a justification. The susceptibility scan is a 2.5-minute dual-echo gradient echo (TE 9.42 and 20 ms, 0.8x0.8x3 mm, 48 axial slices); the Oxford pipeline derives SWI images, QSM maps and regional T2* and susceptibility values from it.
Annotations
There are no manual labels for the full cohort. Automated pipelines produce brain, cardiac and abdominal imaging-derived phenotypes. Radiographers flag potentially serious incidental findings during scanning. Sundaresan et al. (2023) report microbleed masks for 78 preselected participants: a radiologist marked 186 microbleeds on SWI as points that region growing turned into masks. The paper does not say the masks are shared.
Known limitations
- The cohort is not representative of the UK population (5.5% of those invited joined), and the imaging subset is a further self-selected group.
- Participants with safety contraindications, or who cannot lie still or hold their breath, cannot complete the full protocol.
- Not every participant completes every examination; carotid ultrasound images are released for a smaller subset.
- Access requires an approved application, fees, and work inside the UK Biobank cloud platform.
Cohort
Aggregate numbers from the sources below. Bars are relative to the 102,521 subjects.
Age
mean 66.5± 7.8 · range 44 to 87No age bins reported.
Contrast / sequence
subjects, values can overlap
- T1-weighted 87,637 85%
- FLAIR 86,602 84%
- Diffusion-weighted 85,662 84%
- Susceptibility-weighted 74,493 73%
- QSM susceptibility map 73,820 72%
License and access
Our reading of the license, not legal advice. Before you use the data, read the original license and confirm that your use is allowed. We take no responsibility for how you use a dataset. Full disclaimer
Access costs money
UK Biobank Material Transfer Agreement (data only)
Access for an approved health-related research project in the public interest, open to academic and commercial applicants, after paying access charges and signing the MTA. Data are analysed on the UK Biobank Research Analysis Platform and may not be passed on. Results data, including trained model parameters, must be returned to UK Biobank, manuscripts must be sent two weeks before publication, and data must be destroyed when the project ends.
Access needs a registered researcher account, an approved application and payment of access charges. Individual data are normally analysed on the UK Biobank Research Analysis Platform (UKB-RAP); downloads need an exemption.
What you can do
- Conditional
- Conditional
- Yes
- Conditional
What you can share
- No
- Conditional
- Conditional
What you must do
- Yes
- No
- Yes
- Conditional
- Yes
- Conditional
- Yes
- Yes
Limits
- Yes
- Conditional
Commercial license: Not stated Commercial organisations use the same application, fee and MTA; no separate commercial license is described.
Citation
Include "This research has been conducted using the UK Biobank Resource under application number [ ]." in the acknowledgements and "UK Biobank" in the title or abstract (MTA clauses 5.14-5.15). Imaging reference: Littlejohns TJ, et al. The UK Biobank imaging enhancement of 100,000 participants: rationale, data collection, management and future directions. Nature Communications 11, 2624 (2020). doi:10.1038/s41467-020-15948-9
All numbers
Every number on this page, as stored in stats.csv, with its source.
| Measure | Breakdown | Value | Source |
|---|---|---|---|
| Subjects | total participants who attended the first imaging visit (2014+); Showcase version Sep 2025 | 102,521 | ukb-showcase Data-Field 53, instance 2 |
| Studies | total sum of 102,521 first imaging visits and 21,469 first repeat imaging visits (2019+); the two visit types are disjoint and are the only imaging instances listed | 123,990 | ukb-showcase Data-Field 53, instances 2 and 3 |
| Subjects | contrast=T1w participants with a released T1 NIfTI at either imaging visit (version Aug 2025) | 87,637 | ukb-showcase Data-Field 20252 |
| Subjects | contrast=FLAIR T2 FLAIR NIfTI | 86,602 | ukb-showcase Data-Field 20253 |
| Subjects | contrast=dwi multiband diffusion NIfTI | 85,662 | ukb-showcase Data-Field 20250 |
| Subjects | contrast=swi susceptibility-weighted brain images (NIfTI) at either imaging visit: 74,147 at the first and 4,648 at the repeat visit (version Aug 2025). The raw SWI DICOM, Data-Field 20219, covers 86,634 participants (version Sep 2025) | 74,493 | ukb-showcase Data-Field 20251 |
| Subjects | contrast=Chimap quantitative susceptibility mapping images (NIfTI) derived from the SWI scan, either imaging visit (version Aug 2025) | 73,820 | ukb-showcase Data-Field 26301 |
| Mean age | total age at the first imaging visit, 102,489 participants (Showcase: 66.471) | 66.5 | ukb-showcase Data-Field 21003, instance 2 |
| Age SD | total Showcase: 7.84919 | 7.8 | ukb-showcase Data-Field 21003, instance 2 |
| Median age | total | 67 | ukb-showcase Data-Field 21003, instance 2 |
| Minimum age | total | 44 | ukb-showcase Data-Field 21003, instance 2 |
| Maximum age | total | 87 | ukb-showcase Data-Field 21003, instance 2 |
Sources
The keys used in the table above.
- littlejohns2020 Littlejohns et al. 2020, The UK Biobank imaging enhancement of 100,000 participants, Nature Communications paper
- ukb-showcase UK Biobank Data Showcase, public data-field pages (versions Aug/Sep 2025) website
- ukb-community-imaging UK Biobank Community, Imaging Data article website
- sundaresan2023 Sundaresan et al. 2023, Automated detection of cerebral microbleeds on MR images using knowledge distillation framework, Frontiers in Neuroinformatics paper