Welcome to the website of the Multiple Acquisitions for Standardization of Structural Imaging Validation and Evaluation (MASSIVE) database. MASSIVE is an initiative from the PROVIDI-Lab, (UMC Utrecht, The Netherlands), which is headed by Alexander Leemans.
The MASSIVE database was designed and acquired by Martijn Froeling, Alexander Leemans, Chantal Tax and Sjoerd Vos and provides a robust framework to evaluate novel diffusion processing techniques and reliably compare different approaches for diffusion modeling. The database consist of 8000 diffusion-weighted volumes (five shells and two Cartesian grids) and ten 3D fluid-attenuated inversion recovery (FLAIR), T1-, and T2-weighted datasets of a single healthy subject and may be used for your own research purposes (reference below).
For more details, see: M. Froeling, C.M.W. Tax, S.B. Vos, P.R. Luijten, A. Leemans, "MASSIVE" Brain Dataset: Multiple Acquisitions for Standardization of Structural Imaging Validation and Evaluation, Magnetic Resonance in Medicine, 2016
Currently, we are providing the raw, the normalized and the drift-corrected diffusion data in nifti file format. For each scan session the corresponding B0 and noise maps are available. In addition, the dataset contains ten anonymized (facial features were removed) volumetric FLAIR, T1-, and T2-weighted datasets of the same subject. Furthermore, we provide a fully processed dataset at 2.5 mm3 isotropoic resolution, which is corrected for subject motion, eddy current induced distortions, and susceptibility (EPI) artifactst.
supplemental.xls (all the information of the data)
bmatfiles-all.zip (all the b-value, b-vector and b-matrix files )
Diffusion data - full datasets unsorted
These are the nifti files containg all the raw, the intensity normalized, and the signal drift-corrected diffusion data in the order they were acquired. This data is not sorted or corrected for subject motion, eddy current distortions, or EPI artifacts.
These are the nifti files containing the B0-maps, FLAIR, T1-, and T2-weighted datasets, and the noise maps.
These are the nifti files containing the sorted, the intensity normalized, and the signal drift-corrected diffusion data grouped per shell and grid.
First, the data were correted for subject motion and eddy current distortions. Next the all data was tranformed to a T1-weighted anatomical target dataset using non rigid registration. All other anatomical datasets (FLAIR, T1-, and T2-weighted) were also tranfered to the same T1-weighted target dataset.
Diffusion data - sorted
These are the nifti files containing the registered (corrected for subject motion, eddy current distortions, and EPI artifacts), sorted, intensity normalized, and signal drift-corrected diffusion data grouped per shell and grid.
These are the nifti files containing the FLAIR, T1-, and T2-weighted datasets matching the registered diffusion data.
The gradients were generated using a tool (Figure 1) for mathematica cdf player developed by Martijn Froeling, PhD.
The tool is integrated in QMRITools or can be downloaded as a standalown file here.
For information or questions please send an e-mail to: firstname.lastname@example.org
Martijn Froeling, PhD
Alexander Leemans, PhD
Chantal Tax, MSc
Sjoerd Vos, PhD