Overview
The Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset is a longitudinal, multi-center study designed to develop clinical, imaging, genetic, and biochemical biomarkers for the early detection and tracking of Alzheimer's Disease (AD).
Official Website: adni.loni.usc.edu
Data Available in Mind Data Hub
Below, we summarize the modalities and number of subjects processed and available for request in Mind Data Hub:
Structural MRI (T1-weighted & T2-weighted)
Preprocessing Steps:
- Conversion to NIfTI format
- Brain Extraction
- Registration to MNI Space using ANTs (Rigid + Affine transformations)
- Quality Control (QC)
- Considers and specifies the best scan when multiple scans in the same session using SynthSeg.
Data Available:
- T1-weighted scans: 15,529 records from 2,744 unique subjects (~5.66 sessions per subject)
- T2-weighted scans: Limited subjects (insufficient for extensive analysis)
Metadata:
Comprehensive demographic, clinical, and scanner metadata, QC scores based on SynthSeg.
Metadata Source Files:
- Primary Demographics: PTDEMOG_17Jun2025.csv
- Age/Sex/Research Group: original.csv
- Scanner Metadata: MRIQC_17Jun2025.csv
- QC Results: SynthSeg QC results
Processed data are organized in a CSV file with the following columns:
| Column Name | Description |
|---|---|
| subjectId | Subject identifier (XXX_S_XXXX format) |
| sessionName | Session date (YYYY-MM-DD format) |
| MNI_Warped | Path to MNI warped brain file |
| MNI_ZSCORE | Path to z-score normalized file |
| MNI_Z_Cropped | Path to cropped z-score file |
| age | Age in years (range: 51-104, mean: ~73.8) |
| sex | Sex encoded (1=male, 2=female) |
| handedness | Handedness (0=unknown, 1=right, 2=left, 3=ambidextrous) |
| ethnicity | Ethnicity (0=unknown, 1=non-hispanic, 2=hispanic) |
| race | Race (standardized 0-7 encoding) |
| education | Education level in years |
| marital_status | Marital status (1-6 encoding) |
| work_status | Work status (1=no, 2=yes, 0=unknown) |
| home_status | Home living situation (1-10 encoding) |
| research_group | Research classification (CN, EMCI, LMCI, MCI, AD, SMC, Patient) |
| scanner_manufacturer | Scanner brand (GE, Siemens, Philips, etc.) |
| overall_qc_score | SynthSeg quality score (0-1 scale, higher=better) |
Encoding Dictionaries:
- Sex: 1 = Male, 2 = Female, 0 = Unknown
- Handedness: 0 = Unknown, 1 = Right, 2 = Left, 3 = Ambidextrous
- Ethnicity: 0 = Unknown, 1 = Non-Hispanic, 2 = Hispanic
- Race: 0 = Unknown, 1 = White, 2 = Black, 3 = Asian, 4 = Pacific Islander, 6 = Multiple, 7 = Native American
- Work Status: 0 = Unknown, 1 = No, 2 = Yes
- Research Group: CN (Cognitive Normal), EMCI (Early MCI), LMCI (Late MCI), MCI, AD (Alzheimer's Disease), SMC (Significant Memory Concern), Patient
Note: Due to the high volume of T1 scans per subject, the best-quality scan per subject/session was selected based on ADNI QC recommendations. Also, if a subject has more than one scan in the same session, we ranked them considering SynthSeg.
Download Data Date: February 2025
Functional MRI (Resting-state fMRI)
Preprocessing Steps (fMRIPrep):
- Conversion to NIfTI (BIDS format)
- Slice-timing and motion correction
- Brain extraction and normalization to MNI
- Denoising and smoothing
- Quality Control (QC) and time-series extraction
Data Available:
- fMRI scans processed using fMRIPrep: 1219 subjects, with considering all task resting sessions.
- Comprehensive QC files.
Dedicated scripts are available for both BIDS conversion and parcellation:
- An automated script converts raw imaging data to the standardized BIDS format. Access BIDS conversion script (GitHub)
- A separate script is provided for parcellation calculations. Access parcellation script (GitHub)
Download Data Date: February 2025
Diffusion Tensor Imaging (DTI)
Preprocessing Steps:
- Convert DICOM to NIfTI
- Topup Correction (Using a top-up tool from FSL)
- Eddy Correction (Using eddy tool from FSL)
- Brain Extraction (Using Bet tool from FSL)
- Merging of Acquisitions (If the session is split into multiple scans)
- DTI Model Fitting(Using the Dtifit tool from FSL)
- Registration to MNI Space using ANTs (Rigid + Affine transformations)
- QC
Data Available:
DTI scans: All Adni 1, 2,Go, 3 and 4 DTI scans, and Metadata.
Download Data Date: July 2025
Request Access to ADNI Data
To request access to ADNI data, please fill out our internal request form. Requests will be reviewed by our data management team.
Citation and Acknowledgment
When using ADNI data, please cite the official ADNI project and acknowledge the Mind Data Hub preprocessing pipeline:
“Data used in the preparation of this article were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). The preprocessing and harmonization of data were performed by Mind Data Hub at Stanford University.”
Useful Links