Connectomics of cerebral cortex: multiscale studies of cell types, circuits, and projections

Over the past decade, new tools have emerged for studying the structure of networks in the brain, a field now known as connectomics. Connectomes can be studied at different scales, from the synaptic connections between individual neurons—microscale connectomics—to brain-wide maps of neuronal projections between areas: macroscale connectomics. Microscale connectomics is typically performed with serial-section electron microscopy, a technique that creates volumetric images of synaptic circuits with resolution on the scale of ~10 nanometers. Our group is finishing data collection on an extremely large data set, encompassing a local circuit of the cerebral cortex, with ~100,000 neurons and 109interconnections. I’ll discuss the technical challenges in creating this petascale data set, using transmission electron microscopy to image 25,000 serial sections of brain tissue. Initial analysis of smaller data sets shows that microscale connectomics with dense segmentation (performed by our collaborators at Princeton) is a powerful tool for analyzing cortical cell types and their connection onto distinct targets within the local circuit.
 
I will also discuss using light microscopy to capture macroscale connectomes of large brains, including the human. This was previously unrealistic because data sizes would be intractable. With the tools developed for microscale connectomics—from data acquisition through computational analysis—a macroscale connectomics that traces individual axons throughout the brain is now a possibility.