NextBrain: a next-generation, histological atlas of the human brain for high-resolution neuroimaging studies.

Videos

Some visualization material showing the 3D histology reconstruction of LFB, H&E and high resolution labels as well as its alignment to brain ex vivo MRI of case 5 (left) and the probabilistic atlas construction using all 5 cases (center) and a tutorial for the Data Explorer functionality (right)

The NextBrain data release includes a 200um isotropic labeling of the right hemisphere of the ex vivo scan publicly released by Edlow et al. in “7 Tesla MRI of the ex vivo human brain at 100 micron resolution”, Scientific data 6, 244 (2019). The following two videos fly over the coronal and axial slices of the dataset (download links for the imaging volumes can be found under “Code and Data” below):

Pipeline

The different steps of the pipeline from the fresh tissue donation to the joint multimodal 3D registration of LFB and HE stains, labels and MRI are summarized in the following figure:

Other publications

• The data acquisition and tissue processing pipeline are described in the following publication:

  A multimodal computational pipeline for 3D histology of the human brain.
  M. Mancini, A. Casamitjana, L. Peter, E. Robinson, S. Crampsie, D.L. Thomas, J.L. Holton, Z. Jaunmuktane, J.E. Iglesias
  Scientific Reports 10, 13839 (2020)
  [article]   [arxiv]   [citation]

• The histological sections are registered to the coordinate space of the ex vivo MRI using the pipeline described in the following publication:

  Robust joint registration of multiple stains and MRI for multimodal 3D histology reconstruction: Application to the Allen human brain atlas.
  A. Casamitjana, M. Lorenzi, S. Ferraris, L. Peter, M. Modat, A. Stevens, B. Fischl, T. Vercauteren, J.E. Iglesias
  Medical image analysis 75 (2022)
  [article]   [arxiv]   [citation]   [code]

• The histological sections are registered to the coordinate space of the ex vivo MRI using the pipeline described in the following publication:

  Synth-by-Reg (SbR): Contrastive learning for synthesis-based registration of paired images
  A. Casamitjana, M. Mancini, J.E. Iglesias
  Simulation and Synthesis in Medical Imaging Workshop (SASHIMI, 2021)
  [article]   [arxiv]   [citation]   [code]

Code and data

• The raw data of the project can be found here:

  Registered histology, MRI, and manual annotations of over 300 brain regions in 5 human hemispheres (data from ERC Starting Grant 677697 "BUNGEE-TOOLS")
  J.E Iglesias Gonzalez, A. Casamitjana, A. Atzeni, B. Billot, D. Thomas, E. Blackburn, J. Hughes, J. Althonayan, L. Peter, M. Mancini, N. Robinson, P. Schmidt, S. Crampsie
  [data repository]   [citation]

• Code used for the 3D histology reconstruction

  [github]   

• Bayesian segmentation using NextBrain probabilistic atlas

  [ready-to-use tool in FreeSurfer]   [github]   

• 200um isotropic labeling of the right hemisphere of the ex vivo scan publicly released by Edlow et al., (2019).
  We include the original scan, the manual segmentation and the look-up-table (LUT) in freeview format for label visualization.

  [data repository]   [original scan]   [manual segmentation]   [LUT]   

  You can open these files in Freeview with the command:

  freeview -v sub-001_ses-01_T2w.nii.gz sub-001_ses-01_desc-manual_dseg.nii.gz:colormap=lut:lut=sub-001_lut.txt

Authors

• James Hughes
  [Github]

• Peter Schmidt
  [Github]

• Adria Casamitjana
  a.casamitjana [at] ub.edu
  [Twitter  |  Scholar]

• Juan Eugenio Iglesias
  jiglesiasgonzalez [at] mgh.harvard.edu
  [Twitter  |  Linkedin  |  Scholar]