Center for a pipeline of high throughput integrated volumetric electron microscopy for whole mouse brain connectomics

Two fundamental components of the structural basis of brain function are cell type composition and the wiring diagram between those cells. Over the past decade there has been paradigm-shifting progress in understanding cell type composition of the brain. Now it’s time to systematically uncover the brain’s wiring diagram and place it into the context of cell types. 

Knowledge about the complete connectomes in C. elegans and Drosophila have revolutionized the understanding of cell types and circuit function in those systems. Transmission Electron Microscopy (TEM) has consistently led progress in that revolution and has the potential to scale up to the entire mouse brain with technical improvements in certain areas. During the IARPA MICrONS project, members of the HIVE team built a complete pipeline to section and image the mm3 and created the data processing, reconstruction and analysis infrastructure to make cells and connections analyzable. The result was a dataset with the largest EM level reconstructions of cells in any system, with neurons containing more than 14,000 inputs and 15,000 outputs. 

We accomplished this by applying a rigorous structured science process that is a hallmark of the Allen Institute’s team science approach. In this project we aim to improve our pipeline, developing critical technologies to tackle the challenges of scaling up to the whole mouse brain and linking to cell types. Our proposal will prepare and section an entire hemisphere, image it at 120 nm resolution, and image up to 10 mm3 at synaptic resolution within the Cortical Basal Ganglia Thalamic loop to provide key insights into circuit mechanisms within this circuit. 

To accomplish this, we will need to improve all individual aspects of the pipeline, while maintaining consistent integration tests that ensure that the pieces work together. 

• We will standardize a whole mouse brain staining protocol and advance the automation of serial sectioning to collection of serial TEM sections across a whole mouse brain. Developments of serial section tilt TEM tomography will allow to scale EM imaging to a whole mouse brain at multiple scales, leveraging the re-imaging capacity of serial section TEM. 

• We will develop open source data processing tools to bring down the cost of segmentation and while improving accuracy and integrating with a real-time globally accessible proofreading and analysis platform. 

• Finally, we will integrate our data with full morphology reconstructions linked to gene expression, allowing us to create an integrated atlas of cell types and connectivity. Our dissemination strategy will further amplify our impact by democratizing access for both the scientific and educational community.