Email questions
or comments

Age Related Atrophy Dataset


The Age Related Atrophy dataset contains structural MR images of 70 subjects collected during 2008-2010 across a wide range of ages. The dataset also contains resting state fMRI for most subjects. The structural images are T1 weighted, T2 weighted-FLAIR, 25 direction DTI, and the T1 mapping DESPOT [1] sequence. Reconstructed T1 maps for each subject are also available.

The aquisition protocol was designed to study structural differences between young and older adults including both shape and intensity changes. Anonymized DICOM image sessions and processed images for each subject are available through the BSL XNAT server. The data is licensed under the Creative Commons Attribution License. It may be used freely for commercial, academic, or other use, as long as the original source is properly cited as follows:

C. L. Wyatt and P. J. Laurienti. Age Related Atrophy Dataset, Virginia Tech and Wake Forest University School of Medicine, 2011.


Subjects were recruited from the community by word of mouth, advertising, and using a participant database. All subjects were screened prior to inclusion to exclude anyone that has contraindications for entering the MRI scanner. To ensure normal cognitive function, participants with mini mental state exam scores greater than 3 standard deviations from the norm for age and education were excluded. Potential participants were also excluded for previous brain surgery, major CNS trauma, or prior documented history of stroke (not including transient ischemic attacks) that resulted in lasting sequela. Individuals diagnosed with active neurological dysfunction or who were taking antipsychotic and/or antiepileptic medications with known neurological side effects were excluded. A full medical history was collected on each subject using self-report forms and a medication review to identify major medical conditions subject to exclusion.  All subjects completed and signed a consent form approved by the IRBs at Virgina Tech and the Wake Forest University School of Medicine.

The dataset consists of 70 subjects:

Group 1, N = 23, ages 18 - 38 years, 11 male / 12 female

Group 2, N = 47, ages 65 - 90 years, 22 male / 25 female

Image Acquisition

All data were collected on a GE 1.5 T scanner with an eight-channel head coil. A conventional T1w image (3D IR-SPGR, flip angle = 20 degrees, inversion time (TI) = 600ms) and a pair of spoiled gradient recalled echo (SPGR) images at two programmed flip angles (angle 1 = 4 degrees and angle 2 = 15 degrees), the DESPOT sequence, were collected for each subject. In both cases the same slice prescription was used with repetition time / echo time (TR / TE) = 7.39ms / 2.04ms, and a slice thickness = 1.2mm. The total acquisition times were 7'28'' and 6'54'' for the IR-SPGR image and multi-flip angle SPGR image pair respectively.


Deidentification was performed by defacing the high-resolution structural images (T1w, DESPOT, and T2-FLAIR) using the Freesurfer mri_deface tool applied to the T1w image converted to the NIFTI format, creating a face mask. The original DICOM images were then defaced by resampling of the face mask and zeroing out those voxels using a custom ITK-based tools (source below). Anonymization of the DICOM fields was then performed using the DicomRemap tool from the Neuroinformatics Research Group.

The T1 maps were generated from the two SPGR images using the driven equilibrium single-pulse observation of T1 (DESPOT1) technique [1]. The two flip angle images are rigidly registered prior to pixelwise estimation of the T1 value to account for any patient motion between the aquisitions. The source code for the reconstruction is provided below.

Source Code

The C++ source code has ITK 3.20 as a dependency and uses CMake.

The DESPOT reconstruction code needs the GE-private.dic file provided on the command line using the -d option to pickup the private fields used to encode the flip angles. This file is included in the source distribution. Usage is available from the command line using the -h flag.

Obtaining the Data

All data is avaliable from the public ACE project in the BSL XNAT instance. You can download some or all of the data using the download images link under the actions tab.

The group designations (group1 and group2) corresponding the age ranges above can be obtained by adding the Group column to the display via the Options pulldown, or from this csv file: ARA_spreadsheet.csv.


[1] S. Deoni, B. Rutt, T. Peters, Rapid combined T1 and T2 mapping using gradient recalled acquisition in the steady state, Magnetic Resonance in Medicine 49 (3) (2003) 515–526. [pubmed link]


This research was funded by the National Institutes of Health Roadmap for Medical Research, Grant U54 RR021813.