ProbioticFarmer/oasis-1-lab-processed

OASIS-1 Cross-Sectional Structural Brain MRI — Complete Lab-Processed & Raw Release

Complete 436-session structural brain MRI release from the Open Access Series of Imaging Studies (OASIS-1) collected and processed by the Washington University in St. Louis Knight ADRC & Neuroinformatics Research Group (NRG) (Marcus et al., 2007), organized as losslessly compressed NIfTI-1 (.nii.gz) volumes with rich Parquet manifests for high-throughput compute on Hugging Face Jobs.

Cohort Overview

• 416 unique participants aged 18 to 96 years, including 100 older adults clinically diagnosed with very mild to moderate Alzheimer's disease (CDR = 0.5, 1.0, 2.0).
• 20 short-interval test-retest reliability sessions (MR2): 20 nondemented participants re-imaged on a subsequent visit within < 90 days using the identical 1.5T Siemens Vision protocol (436 total sessions).
• Multi-stage lab processing preserved per session (sessions/OAS1_XXXX_MRY/):
  1. raw/ — 3 to 4 individual raw 1.5T T1-weighted MPRAGE acquisitions (256 × 256 × 128, 1.0 × 1.0 × 1.25 mm).
  2. processed/subj_111/ — WashU NRG motion-corrected, coregistered, averaged 1.0 mm isotropic native-space structural MRI (256 × 256 × 160).
  3. processed/t88_111/ — WashU NRG 1.0 mm atlas-registered (711-2C Talairach space, 176 × 208 × 176), gain-field inhomogeneity-corrected (*_t88_gfc.nii.gz), and brain-masked (*_t88_masked_gfc.nii.gz) volumes, plus all t4_files/ rigid & affine transformation matrices.
  4. fsl_seg/ — WashU NRG 1.0 mm FSL FAST HMRF-EM tissue segmentations (*_fseg.nii.gz: 0=Background, 1=CSF, 2=GM, 3=WM) and convergence/volume logs (*_fseg.txt).
  5. nilearn_cache/ — Drop-in SPM8/DARTEL MNI152 2mm modulated gray/white matter probability maps (mwrc1*, mwrc2* across 403 QC-passed subjects) and Harvard-Oxford 1mm/2mm Cortical & Subcortical MNI152 Atlases.

Fast Usage on Hugging Face Jobs

from huggingface_hub import snapshot_download
import pandas as pd
import nibabel as nib

# Pull metadata + T88 skull-stripped 1mm volumes + FSL segmentations at multi-Gbps on HF Jobs
root = snapshot_download(
    repo_id="ProbioticFarmer/oasis-1-lab-processed",
    repo_type="dataset",
    allow_patterns=["metadata/*", "sessions/*/processed/t88_111/*", "sessions/*/fsl_seg/*"],
)

df = pd.read_parquet(f"{root}/metadata/sessions.parquet")
pairs = pd.read_parquet(f"{root}/metadata/test_retest_pairs.parquet")
img = nib.load(f"{root}/{df.loc[0, 't88_masked_gfc_path']}")

Terms of Use & Citation (OASIS Data Usage Agreement)

Redistribution of OASIS-1 data is permitted under the OASIS Data Usage Agreement. Users must make no attempt to identify participants and must acknowledge OASIS and the supporting NIH grants (P50 AG05681, P01 AG03991, R01 AG021910, P20 MH071616, U24 RR021382) in publications:

Marcus, D. S., Wang, T. H., Parker, J., Csernansky, J. G., Morris, J. C., & Buckner, R. L. (2007). Open Access Series of Imaging Studies (OASIS): Cross-sectional MRI data in young, middle aged, nondemented, and demented older adults. Journal of Cognitive Neuroscience, 19(9), 1498–1507.