AMYPAD PNHS
Amyloid Imaging to Prevent Alzheimer's Disease (AMYPAD) Prognostic and Natural History Study
European cohort of 3,368 older adults without dementia from 10 parent cohorts, with florbetaben or flutemetamol amyloid PET for 1,620 and longitudinal multimodal brain MRI (T1w, FLAIR, T2*, SWI, diffusion, fMRI, ASL), plus clinical, cognitive, CSF and genetic data. Access by reviewed request.
Overview
The AMYPAD Prognostic and Natural History Study added amyloid PET to people already followed in European cohorts of adults without dementia, to study how amyloid builds up in the preclinical phase of Alzheimer's disease. It was part of the IMI 2 project AMYPAD, and Amsterdam UMC holds the data. Participants stay in their parent cohort, which supplies cognition, biomarkers, MRI and risk factors, while AMYPAD provides the PET scans. The study is moving into the Euro-PAD community.
Composition
The harmonised release integrates 3,368 participants from 10 parent cohorts, among them EPAD LCS, EMIF-AD 60++ and 90+, ALFA+, FACEHBI, F-PACK, DELCODE and the AMYPAD diagnostic study. Of these, 1,620 had a baseline amyloid PET and 888 at least one follow-up PET. The MRI data descriptor covers 2,759 cognitively normal or mildly impaired people with longitudinal scans. A separate raw release holds 3,266 participants from nine cohorts. Tables cover demographics, cognition, CSF, APOE and polygenic scores, vital signs, medical history and lifestyle.
Acquisition
Sites are in Belgium, France, Germany, Spain, Sweden, Switzerland, the Netherlands and the United Kingdom. PET used [18F]florbetaben or [18F]flutemetamol, as static or dual-time-window dynamic scans, with reconstructions harmonised using Hoffman phantom scans across the 28 PET scanners of the AMYPAD studies. MRI comprises core T1w, T2w, T2* and FLAIR and advanced SWI, resting-state fMRI, diffusion and ASL sequences. Raw and processed images are shared as NIfTI in BIDS layout.
Annotations
PET scans carry blinded visual reads and Centiloid, SUVR and kinetic values. MRI comes with visual ratings (atrophy scales, Fazekas, lacunes, microhaemorrhages, infarcts), BaMoS white matter hyperintensity segmentations, FreeSurfer measures and diffusion, perfusion and connectivity phenotypes. No voxel-level microbleed labels are described.
Known limitations
- Data come from 17 sites, and participants entered at different points of their parent cohort follow-up.
- Raw data need approval from every contributing parent cohort, and only group-level derived data may be shared.
- Scanner vendors, field strengths and per-sequence counts are not stated in the public sources used here.
Cohort
Aggregate numbers from the sources below. Bars are relative to the 3,368 subjects.
Modality
subjects
- MRI 2,759 82%
- PET 1,620 48%
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
A research proposal is reviewed and approved
AMYPAD PNHS Data Use Agreement
Accepted with a data access request that the AMYPAD Data Sharing and Publication Committee reviews. Users act as data processors, must not re-identify participants or pass the data on, may share derived data only at group level, and must credit AMYPAD as set out in the PNHS Publication Policy.
Requests need a one-page proposal reviewed by the AMYPAD Data Sharing and Publication Committee. Raw data and single-cohort requests also need approval from each parent cohort. Tables are analysed in the AD Data Initiative's AD Workbench, whose own terms of use are not covered here; images are provided through the Health-RI XNAT platform.
What you can do
- Not stated
- Train ML models Not stated
- Yes
- Yes
What you can share
- No
- Conditional
- Not stated
What you must do
- Yes
- Share alike No
- Yes
- Conditional
- Manuscript review No
- Release code No
- Return results No
- Delete after use No
Limits
- Yes
- No
Citation
Add "On behalf of the AMYPAD Consortium*" to the byline and include the EudraCT number 2018-002277-22, the dataset version and DOI, and the IMI 2 grant 115952 funding statement (AMYPAD PNHS Publication Policy, doi:10.5281/zenodo.21032996). MRI reference: Pieperhoff L, et al. AMYPAD-PNHS: A Pan-European, Multi-Site & Multimodal MRI Dataset of Older Adults Without Dementia. Scientific Data (2026). doi:10.1038/s41597-026-07988-5
All numbers
Every number on this page, as stored in stats.csv, with its source.
| Measure | Breakdown | Value | Source |
|---|---|---|---|
| Subjects | total harmonised dataset from 10 parent cohorts | 3,368 | amypad-data About the AMYPAD PNHS Data |
| Subjects | modality=PT participants with a baseline amyloid PET (visual read; 1476 also have a Centiloid value) | 1,620 | amypad-data About the AMYPAD PNHS Data |
| Subjects | modality=MR cognitively normal or MCI participants in the MRI dataset; longitudinal | 2,759 | pieperhoff2026 Abstract |
Sources
The keys used in the table above.
- amypad-data AMYPAD website, Data page (harmonised PNHS dataset) website
- pieperhoff2026 Pieperhoff L et al. 2026, AMYPAD-PNHS multimodal MRI dataset, Scientific Data (accelerated article preview) paper
- bollack2025 Bollack A et al. 2025, Implementing robust amyloid PET quantification in multi-center studies, the AMYPAD approach paper
- amypad-dar AMYPAD PNHS Data Access Request procedure, June 2026 (Zenodo) website
- lopesalves2020 Lopes Alves I et al. 2020, Quantitative amyloid PET in Alzheimer's disease, the AMYPAD PNHS, Alzheimer's & Dementia paper