The Brain/MINDS Marmoset Connectivity Atlas: exploring bidirectional tracing and tractography in the same stereotaxic space

Henrik Skibbe, Muhammad Febrian Rachmadi, Ken Nakae, Carlos Enrique Gutierrez, Junichi Hata, Hiromichi Tsukada, Charissa Poon, Kenji Doya, Piotr Majka, Marcello G. P. Rosa, Hideyuki Okano, Tetsuo Yamamori, Shin Ishii, Marco Reisert, Akiya WatakabebioRxiv 2022 · DOI

TL;DR

This paper introduces the Brain/MINDS Marmoset Connectivity Atlas (BMCA), a novel resource that integrates high-resolution anterograde and retrograde neural tracer data with diffusion MRI tractography in a common stereotaxic space. This atlas provides an unprecedented tool for studying primate brain connectivity, enabling direct comparisons between different individuals and facilitating cross-species analyses with human brain data.

What Problem Does This Paper Address?

Understanding the intricate structural connectivity of the primate brain is crucial for deciphering its complex functions and for developing treatments for neurological disorders. While rodent models have been extensively studied, their brain anatomy differs significantly from primates, limiting their utility for understanding human cognition. The marmoset, a small primate with a brain structure more akin to humans, has emerged as a valuable model organism. However, a comprehensive, publicly accessible, and high-resolution atlas of marmoset brain connectivity, particularly at the level of individual neuronal pathways, has been lacking. Existing resources often focus on specific modalities (e.g., only dMRI or retrograde tracing) or lack the resolution needed for detailed pathway analysis. This paper addresses this gap by creating a multimodal, high-resolution atlas that bridges different data types and species.

Methods

The BMCA pipeline involves several key steps:

  • Image Acquisition: This includes ex-vivo diffusion MRI (dMRI) for tractography, serial two-photon tomography (STPT) for anterograde tracing, and fluorescent microscopy for retrograde tracing. Nissl and backlit images are also acquired for anatomical reference and myelination visualization.
  • Image Processing: A fully automated pipeline handles image stitching, injection site detection, and segmentation of tracer signals from background noise. This involves sophisticated algorithms, including U-Nets for cell body detection and tracer segmentation.
  • Registration and Alignment: All imaging data is aligned to a common, population-average STPT template space with high resolution (3x3x50µm³). This is achieved using advanced registration techniques like ANTs.
  • Data Integration: The processed tracer data is integrated with dMRI tractography data, retrograde tracer data from the Marmoset Brain Connectivity Atlas (MBCA), and anatomical atlases.
  • Tools for Exploration: The BMCA-Explorer, an online viewer, allows for high-resolution visualization of the integrated data. The Nora-StackApp provides lower-resolution 3D volume data for offline exploration and comparative analysis.
  • Cross-Species Mapping: A diffeomorphic warp between marmoset and human brain anatomy is provided to facilitate translation of findings.

Datasets

  • Anterograde Tracer Data: Derived from 52 anterograde neural tracer injections in the prefrontal cortex of marmosets, collected by the Brain/MINDS project.
  • Retrograde Tracer Data: Integrated from the Marmoset Brain Connectivity Atlas (MBCA), comprising 145 retrograde tracer injections.
  • Diffusion MRI Data: High angular resolution diffusion imaging (HARDI) data from marmosets.
  • Anatomical Atlases: Includes the Brain/MINDS atlas (BMA), Marmoset Brain Connectivity Atlas (MBCA), and Marmoset Brain Mapping project (MBM) atlases.
  • Human Brain Data: Used for cross-species mapping, including human dMRI data and the JHU white matter atlas.

Key Insights and Ideas

  • Bridging Modalities: The BMCA successfully integrates diverse neuroimaging techniques (STPT, dMRI, fluorescent microscopy) into a unified framework, offering a more complete picture of brain connectivity than any single method alone.
  • High-Resolution Visualization: The BMCA-Explorer's ability to visualize data from different individuals in a common reference space at unprecedented resolution is a significant leap for comparative studies.
  • Ground Truth for dMRI: The high-resolution tracer data serves as invaluable "ground truth" for validating and refining dMRI-based tractography, highlighting both the strengths and limitations of dMRI in capturing fine-grained connectivity.
  • Cross-Species Translation: The development of a diffeomorphic warp between marmoset and human brain anatomy is a crucial step towards translating findings from this primate model to human neuroscience and clinical applications.
  • Reciprocity of Connectivity: The integration of anterograde and retrograde tracing data allows for the investigation of connectional reciprocity, a fundamental principle of brain organization.

Key Results

  • The BMCA provides access to data from 52 anterograde and 164 retrograde tracer injections in the marmoset cortex.
  • The BMCA-Explorer allows for direct comparison of data from different individuals in a common reference space at a high resolution of 3.0µm/px.
  • Visual comparisons between dMRI tractography and tracer data reveal remarkable similarities in pathway projections, but also highlight discrepancies, suggesting limitations in dMRI's specificity and sensitivity.
  • The integration of MBCA retrograde data shows strong spatial correspondence with BMCA anterograde data, supporting the alignment of these datasets.
  • Preliminary results of the human-marmoset warp show promising similarities in white matter structures, though challenges remain in gray matter registration.

Limitations and Future Work

  • Resolution Differences: While the BMCA offers high resolution for tracer data, dMRI data is inherently lower resolution, limiting direct comparison at the finest cellular level.
  • Manual Intervention: Despite efforts towards automation, some manual intervention was required for certain registration steps, particularly in the human-marmoset warp, due to significant anatomical differences.
  • Future Work: The authors plan to extend the BMCA to cover other cortical regions, include open-access raw dMRI data, and incorporate data from disease model marmosets. Further efforts are needed to improve and validate the human-marmoset mapping, accounting for differences in gray matter structure and evolutionary expansion patterns.

Summary updated

Abstract

Abstract We report on the implementation and features of the Brain/MINDS Marmoset Connectivity Atlas, BMCA, a new resource that provides access to anterograde neuronal tracer data in the prefrontal cortex of a marmoset brain. Neuronal tracers combined with fluorescence microscopy are a key technology for the systematic mapping of structural brain connectivity. We selected the prefrontal cortex for mapping due to its important role in higher brain functions. This work introduces the BMCA standard image preprocessing pipeline and tools for exploring and reviewing the data. We developed the BMCA-Explorer, which is an online image viewer designed for data exploration. Unlike other existing image explorers, it visualizes the data of different individuals in a common reference space at an unprecedented high resolution, facilitating comparative studies. To foster the integration with other marmoset brain image databases and cross-species comparisons, we added fiber tractography data from diffusion MRI, retrograde neural tracer data from the Marmoset Brain Connectivity Atlas project, and tools to map image data between marmoset and the human brain image space. This version of BMCA allows direct comparison between the results of 52 anterograde and 164 retrograde tracer injections in the cortex of the marmoset.
BibTeX
@misc{skibbe2022brain,
  author = {Henrik Skibbe and Muhammad Febrian Rachmadi and Ken Nakae and Carlos Enrique Gutierrez and Junichi Hata and Hiromichi Tsukada and Charissa Poon and Kenji Doya and Piotr Majka and Marcello G. P. Rosa and Hideyuki Okano and Tetsuo Yamamori and Shin Ishii and Marco Reisert and Akiya Watakabe},
  title = {{The Brain/MINDS Marmoset Connectivity Atlas: exploring bidirectional tracing and tractography in the same stereotaxic space}},
  year = {2022},
  doi = {10.1101/2022.06.06.494999},
  howpublished = {bioRxiv},
}
Figure