BRAIN CONNECTS: Photoemission Electron Microscopy for Connectomics

We propose to develop a new microscope for connectomes with the potential to reduce the price of a connectomic voxel by order(s) of magnitude. We will focus on a relatively unexplored type of EM microcopy based on excitation of photoelectrons by UV light and their detection with standard widefield EM optics (PEEM). Simply, PEEMs for connectomes can combine the

reliability of SEM imaging with the thruput of TEM imaging. We have demonstrated for the first time that PEEMs can see synapse and we have designed a tailored PEEM microscope with strong UV excitation from lasers and direct EM detectors that can work at Gigahertz imaging rates. We will explore collecting 1000s of ultra-thin brain slices (UTBS) using two state of the art

automated approaches, the ATUM and Mag-C. We will optimize this pipeline, extending our current sample preparation to support such imaging rates, adding Gas Cluster Ion milling to improve the Z-resolution of PEEM connectomes, and, by designing, developing, and integrating a new generation of fast stages that keep up with imaging rates. Finally, we will instantiate auto-

acquisition and quality assurance algorithms, e.g., auto-focus, auto-sigmate, etc. to facilitate the imaging and collection of 1000s of ultrathin brain sections. We provide clear metrics for success and achieving those milestones, we argue, could revolutionize how EM conenctomes are made.