Skip to content
Open Imaging Index
MRI PET Brain

HABS

Harvard Aging Brain Study (HABS) public dataset

Brain MRI, amyloid, tau and FDG PET and up to 11 years of cognitive follow-up of older adults who were cognitively normal at entry, followed to study the earliest brain changes of preclinical Alzheimer's disease. 358 people from Massachusetts General Hospital.

Go to dataset Edit Report

Overview

The Harvard Aging Brain Study (HABS) is an NIA-funded longitudinal cohort at Massachusetts General Hospital that follows older adults who were cognitively normal at entry, to study the earliest brain changes of preclinical Alzheimer's disease. The study shares its data with outside researchers; release 3 (March 2025) is hosted on Synapse.

Composition

Release 3 covers 358 participants aged 50 to 90 at baseline (216 women, 142 men), followed for up to 11 years. By the last visit, 300 were still cognitively normal, 43 had mild cognitive impairment and 15 had dementia. Spreadsheets hold demographics, APOE status, clinical ratings (CDR, MMSE, GDS, Hachinski), neuropsychological tests, the PACC composite, plasma pTau217 and FreeSurfer and PET regional measures.

Release 3 started with ADNI-style T1w MPRAGE images and mean PiB and flortaucipir PET images. The Synapse roadmap lists T2w, FLAIR, FDG, resting-state fMRI, DTI and SWI images as later additions. The scan counts in the statistics come from release 2.0 (2020), which shared raw NIfTI images from the baseline, 3-year and 5-year visits, for example 508 SWI scans of 289 participants.

Acquisition

All MRI was acquired on two matched Siemens 3 T TIM Trio scanners at the Martinos Center: MEMPRAGE with matched T2w SPACE, ADNI-style MPRAGE, 3D FLAIR, SWI, 30-direction DTI and resting-state BOLD. SWI was collected for microbleed counts and clinical reads, and was dropped from the 5-year visit onward. PET was done on a Siemens ECAT HR+ with dynamic PiB (0 to 60 min), FDG and flortaucipir. Images are converted to NIfTI. Release 3 refaces T1w and PET images with mri_reface, and all dates are shifted by a per-subject offset.

Annotations

No manual image labels are shared. Derived data are FreeSurfer 6 regional measures and PET regional SUVR and DVR values from PETSurfer.

Known limitations

  • Single site, one scanner model, and a mostly white cohort (80% in release 3).
  • Imaging is collected only at selected visits, so the set of scans per participant varies.
  • The per-sequence image counts describe release 2.0. Counts for release 3 are not published.
  • Whether microbleed counts or SWI reads are part of the shared spreadsheets is not documented.

Cohort

Aggregate numbers from the sources below. Bars are relative to the 358 subjects.

Sex

  • Female 216 60%
  • Male 142 40%

Participants table in HABS public dataset on Synapse, Data Details page (release v3)

PET tracer

  • [11C]PiB 288 80%
  • [18F]FDG 285 80%
  • [18F]flortaucipir 195 54%

p. 4, FDG, p. 4, PiB, p. 5, FTP in HABS public data release, readme for raw neuroimaging data, release 2.0 (2020-07-11)

Condition

Groups can overlap

  • Healthy control 300 84%
  • Mild cognitive impairment 43 12%
  • Dementia 15 4%

Participants table in HABS public dataset on Synapse, Data Details page (release v3)

Age

mean 71.5 ± 7.9, range 50 to 90

No age bins reported.

Participants table, Participants in HABS public dataset on Synapse, Data Details page (release v3)

Contrast / sequence

Groups can overlap

  • Susceptibility-weighted 289 81%
  • FLAIR 289 81%
  • BOLD fMRI 289 81%
  • Diffusion-weighted 287 80%

p. 2, SWI, p. 3, FLAIR, p. 3, DTI, p. 3, Resting State fMRI in HABS public data release, readme for raw neuroimaging data, release 2.0 (2020-07-11)

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
Application

A research proposal is reviewed and approved

Access page

Harvard Aging Brain Study (HABS) Data Use Agreement

Click-through agreement on Synapse after a reviewed request with an intended data use statement. Data may only be used for the stated study, stay inside the recipient's institution and must be destroyed at the end. The data may not be sold or commercialized, publications must acknowledge and cite HABS and be reported to the HABS data committee, and use from listed countries of concern is barred.

Original license text Version read: HABS DUA Click Wrap v.2 on Synapse (access requirement 9606742, October 2025), with the HABS guidance on AI tools Checked 2026-10-08

What you can do

  • Conditional
  • Not stated
  • Conditional
  • Not stated
  • Conditional

What you can share

  • No
  • Not stated
  • Conditional

What you must do

  • Yes
  • Share alike No
  • Yes
  • Conditional
  • Conditional
  • Release code No
  • No
  • Yes

Limits

  • Yes
  • Yes

Commercial license: Not stated

Citation

Dagley A, LaPoint M, Huijbers W, Hedden T, McLaren DG, Chatwal JP, Papp KV, Amariglio RE, Blacker D, Rentz DM, Johnson KA, Sperling RA, Schultz AP. Harvard Aging Brain Study: Dataset and accessibility. NeuroImage 144(Pt B):255-258 (2017). doi:10.1016/j.neuroimage.2015.03.069

Sources

Every number on this page comes from one of these documents. Each chart names the table or page it is taken from. The raw numbers are in stats.csv.