Dataset Open Access

EDEN2020 Human Brain MRI Datasets for Healthy Volunteers

Castellano, Antonella; Pieri, Valentina; Galvan, Stefano; Rodriguez y Baena, Ferdinando; Falini, Andrea

High-resolution  MR datasets of a cohort of 15 healthy adult subjects acquired on  a  3T  scanner  at  the Neuroradiology Unit and CERMAC (Center  of Excellence for High Field Magnetic Resonance), Vita-Salute San Raffaele University and IRCCS San Raffaele Scientific Institute, Milano, Italy. The data includes:

  • T1_3D_PROSET_Sag:  T1-weighted  volumetric  sequence  acquired  on  the sagittal plane for  morphological  characterization.  This  sequence  demonstrates  difference  in  the  T1 relaxation time of tissues and provide excellent contrast between GM and WM.
  • 3D_FLAIR_Tra: Fluid‑Attenuated Inversion Recovery volumetric sequence acquired on the axial planefor morphological characterization. This is an inversion recovery sequence with a long inversion time (TI), which results in removing signal from the cerebrospinal fluid from the images.
  • SWIp_axial: Susceptibility‑Weighted Imaging sequence acquired on the axial plane.This  is  a three-dimensional  high-spatial  resolution  Gradient  Echo  MRI  sequence providing excellent contrast for venous vascular modeling.
  • s3DI_MC_HR: three‑dimensional   high‑resolution   time‑of‑flight   (TOF)   MR angiography  acquisition  to visualize  flow  within  the  arterial  vessel.  It  is  based  on  the phenomenon of flow-related enhancement of spins entering into an imaging slice. As a result of being unsaturated, these spins give more signal that surrounding stationary spins.
  • MIP_s3DI_MC_HR:  angiographic  3D  visualization  of  TOF  images  using  the maximum intensity projection (MIP) technique of reconstruction.
  • raw_data_DTI_32: Diffusion Tensor Imaging raw data. This is a diffusion-weighted Spin  Echo  EPI  single-shot  pulse  sequence  acquired  on  the axial  planealong  32  gradient directions at a b-value of 1000 s/mm2 and one volume without diffusion-weighting (b0 image).
  • raw_data_NODDI: multi-compartmental dMRI sequence for advanced tractography and  NODDI  analyses,  including  an  axial  high  angular  resolution  diffusion-weighted  imaging (HARDI) acquisition along 60 gradient directions at a b-value of 3000 s/mm2,a DTI acquisition along 35 directions at a b-value of 711 s/mm2 and 11 volumes without diffusion-weighting  (b0  images).  The  phase-encoding  direction was  anterior-to-posterior for all these acquisitions.
  • B0_reverse: a sequence without diffusion-weighting having the same geometrical parameters of the ‘raw_data_NODDI’ images, but acquired using a reversed phase-encoding direction (posterior-to-anterior).   This   volume   allowed   estimation   and   correction for susceptibility-induced distortions.

‘DTI’ Folder’: This folder contains the DTI-derived parametric maps calculated off-linefrom the ‘raw_data_DTI_32’ acquisition (32 gradient directions, b-value = 1000 s/mm2) and saved in the NIfTI-1 Data Format.

‘HARDI’ Folder: This folder contains the parametric maps calculated off-linefrom the HARDI acquisition (60 gradient directions, b-value = 3000 s/mm2) of the ‘raw_data_NODDI’ sequence. Maps are saved in the NIfTI-1 Data Format.

‘Tractography’ Folder: This  folder  contains  the  probabilistic  tractography  reconstructions  of  the  main  white  matter  fiber tracts,  calculated  from  the  HARDI  acquisition (60  gradient  directions, b-value  =  3000  s/mm2)  of  the ‘raw_data_NODDI’ sequence. Dipy has been used for q-ball residual-bootstrap fiber tracking. The folder contains a minimum number of two pair of tracts for each subjects.

‘NODDI’ Folder: This folder contains the Neurite orientation dispersion and density imaging (NODDI) parametric maps calculated off-line from the ‘raw_data_NODDI’ acquisition (60 gradient directions at b=3000 s/mm2, 35  gradient  directions  at b=711  s/mmand  11  b0  volumes)  and  saved  in  the  NIfTI-1  Data  Format. Maps have been generated using the NODDI Matlab Toolbox (

Note that all MRI data files were converted from DICOM series using Chris Rorden's dcm2niiX version v1.0.20200331.


Files (6.9 GB)
Name Size
491.5 MB Download
431.7 MB Download
480.5 MB Download
446.3 MB Download
451.1 MB Download
440.5 MB Download
456.0 MB Download
509.1 MB Download
545.4 MB Download
404.6 MB Download
436.2 MB Download
436.6 MB Download
443.7 MB Download
481.1 MB Download
467.7 MB Download
4.6 kB Download
All versions This version
Views 1,651927
Downloads 730729
Data volume 312.9 GB312.4 GB
Unique views 1,424807
Unique downloads 321320


Cite as