X-ray nanotomography for scalable connectomics

Current gold-standard methods for dense tissue reconstruction rely on volume electron microscopy. As electron penetration depth into tissue is limited to a few 100s of nm, this requires tissue sectioning or milling, presenting significant scalability challenges for reliable continuous image acquisition. Importantly, while automatic segmentation methods have seen substantial improvements in recent years, manual proof-reading still provides a significant cost and time overhead that is particularly driven by the interfaces of physical sections. X-rays penetrate macroscopic samples, allowing for largely non-destructive imaging of tissues. Large scale synchrotron facilities can provide high-brightness coherent X-rays at wavelengths in the sub-nanometer range, enabling coherent imaging with nanometer scale 3D resolution. Recent developments in synchrotron technology have seen further increase in coherent photon flux by orders of magnitude at “4th generation” synchrotrons. Theoretical calculations have demonstrated that this puts X-ray imaging of whole mouse brains with 20 nm isotropic resolution in reach. Here, we will leverage the first high-energy realisation of such a 4th generation source at the European Synchrotron Research Facility (ESRF). 

Our proposal is based on the recent decision of ESRF to dedicate a new beamline to the development of X-ray connectomics, aiming for routine acquisition of mm3 volumes with 20 nm resolution in <2 days and ultimately imaging of 500 mm3 with 20 nm resolution within <3 months. Our preliminary work shows that efficient full-field X-ray tomography of brain tissue can already today achieve a resolution sufficient to delineate synapses and fine neurites. In this proposal we will further develop and optimise full-field X-ray tomographic approaches including sample preparation, data acquisition, reconstruction and analysis strategies to further enhance resolving power, scalability and segmentability in order to enable whole mouse brain X-ray connectomics. We will use the existing nanoimaging beamline ID16A at ESRF to pilot these techniques and refine data acquisition and analysis protocols. We will then design and implement a new instrument including beamline optics, sample stage and detection systems to create the next generation, synchrotron-based microscope for connectomics at the new beamline. This will be a facility readily accessible to researchers world-wide and furthermore enable the efficient construction of X-ray connectomics beamlines at other synchrotrons that are undergoing the transition to 4th generation sources in the coming years such as APS, PSI, ALBA and Diamond.