Skip to content
MRI Brain

GSP

Brain Genomics Superstruct Project

Structural and resting-state brain MRI with behavioral measures of healthy young adults, released for large-sample work on brain structure, functional networks and their link to personality and cognition. 1,570 people aged 18 to 35 from Boston, with FreeSurfer measures and 69 test-retest pairs.

Overview

The Brain Genomics Superstruct Project collected a short structural and resting-state MRI protocol, saliva for DNA and online behavioral tests from healthy volunteers who were already taking part in other studies on matched scanners in the Boston area. Its first open data release, from Harvard University and Massachusetts General Hospital, covers 1,570 young adults scanned between 2008 and 2012 and is meant for large-sample work on normal brain structure, functional networks and their relation to personality and cognition. It is hosted on Harvard Dataverse.

Composition

All 1,570 participants are 18 to 35 years old (mean 21.5), and about 58% are women. Each has one T1-weighted scan and at least one resting-state BOLD run; 1,139 have a second run. A separate archive holds 69 people scanned a second time within six months for test-retest work. Spreadsheets give demographics, fMRI quality metrics and FreeSurfer morphometrics computed before face blurring. Self-report measures exist for 926 people and cognitive tasks for 892. A set of more sensitive phenotypes sits on the LONI archive under stricter terms. Genotypes were planned but are not part of this release.

Acquisition

Scans came from Siemens 3T Tim Trio scanners at Harvard and MGH, all with the 12-channel head coil. The T1w image is the root mean square of a four-echo MPRAGE at 1.2 mm isotropic, taking about two minutes. Each BOLD run lasts about six minutes at 3 mm isotropic with a 3 s TR, eyes open. Images are NIfTI, and anatomical scans were defaced by face blurring.

Annotations

There are no manual labels. FreeSurfer 4.5.0 outputs were produced automatically and only checked visually.

Known limitations

  • A convenience sample: many are college students, education and estimated IQ are above the population average, and about 62% are white non-Hispanic.
  • Sessions with artifacts, low signal-to-noise or a large shift in head size after face blurring were dropped, so the sample is filtered on data quality.
  • The scanner software version changed during collection (B13, B15, B17); it is recorded per session.
  • Face blurring changes some regional morphometric estimates.

Cohort

Aggregate numbers from the sources below. Bars are relative to the 1,570 subjects.

Age

mean 21.5 ± 2.9, range 18 to 35

0
200
400
560
18-1920-2122-2324-2526-2728-2930-3132-3334-35

Table 3, Table 2 in Holmes et al. 2015, Brain Genomics Superstruct Project initial data release, Scientific Data

Contrast / sequence

Groups can overlap

  • T1-weighted 1,570 100%
  • BOLD fMRI 1,570 100%

Abstract, Table 2 in Holmes et al. 2015, Brain Genomics Superstruct Project initial data release, Scientific Data

Field strength

  • 3 T 1,570 100%

Methods, MRI data acquisition in Holmes et al. 2015, Brain Genomics Superstruct Project initial data release, Scientific Data

Age by sex

Reported cross table. Missing cells were not published (fewer than 10 or not reported).

Female Male
  • 2
    34-35
    2
  • 5
    32-33
    6
  • 7
    30-31
    13
  • 14
    28-29
    14
  • 40
    26-27
    26
  • 68
    24-25
    41
  • 149
    22-23
    114
  • 289
    20-21
    220
  • 331
    18-19
    229

Table 3 in Holmes et al. 2015, Brain Genomics Superstruct Project initial data release, Scientific Data

Contrast combinations

How many subjects have exactly each set of contrasts.

T1wboldSubjects with exactly this set
1,570

Abstract in Holmes et al. 2015, Brain Genomics Superstruct Project initial data release, Scientific Data

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
Signed agreement

Sign a data use agreement, often reviewed by the provider

Access page

Brain Genomics Superstruct Project Open Access Data Use Terms

Accepted on Harvard Dataverse before requesting the files. Users must not try to identify or contact participants or link the data to other data about them, must follow their institution's human-subjects rules and must acknowledge the GSP. Original and derived data may be passed on only under the same terms.

Covers the open access release on Dataverse. An extended set of more sensitive phenotypes is held on the LONI Image Data Archive and needs separately signed GSP Restricted Data Use Terms, whose text is not public.

Original license text Version read: 2014-Apr-22 (Terms of Use of the GSP Dataverse dataset, version 10.6) Checked 2026-10-09

What you can do

  • Not stated
  • Not stated
  • Not stated
  • Yes
  • Yes

What you can share

  • Conditional
  • Conditional
  • Share trained models Not stated

What you must do

  • Yes
  • Yes
  • Yes
  • Conditional
  • Manuscript review No
  • Release code No
  • Return results No
  • Delete after use No

Limits

  • Yes
  • No

Commercial license: Not stated

Citation

Holmes AJ, Hollinshead MO, O'Keefe TM, et al. Brain Genomics Superstruct Project initial data release with structural, functional, and behavioral measures. Scientific Data 2, 150031 (2015). doi:10.1038/sdata.2015.31

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.