Optimising light-tissue interaction to enable multiscale imaging of neuronal dynamics deep within the neocortex

The neocortex plays a central role in learning new motor skills, such as typing, driving a car and playing tennis. Neocortical circuits receive subcortical input and are highly interconnected within and across layers, so their activity arises from both extrinsic input and intrinsic sources. The process of learning is thought to involve modification in the strength of synaptic connections between neurons and changes in neuronal excitability. But how principal neocortical neurons combine 'top down' self generated predictions with 'bottom up' extrinsic feedforward inputs to improve task performance during learning is unknown. A key reason for this is that such information is encoded in brief synaptic signals that are distributed across their large 3D dendritic trees, which span both superficial and deep cortical regions, which are inaccessible to current imaging methods. Although synaptic activity can be detected using optical microscopes that measure fluorescence from genetically encoded reporters, no available imaging technology can monitor synaptic activity that is distributed across all neocortical layers at high spatiotemporal resolution, because image quality is degraded by scattering of light as it passes through brain tissue. The major physical factor limiting imaging in deep tissue is scattering of light. Understanding how light interacts with brain tissue and developing strategies to compensate for the optical distortions caused by scattering are necessary for extending the depth at which optical microscopes can operate effectively. Collaboration at the interface between physics and biology is therefore essential for addressing neocortical processing at the synaptic level, since it requires deeper, faster imaging than currently possible. This project will bring together teams of leading physicists, microscope developers and neuroscientists at UCL and Oxford University with expertise in modelling light-tissue interactions, optical wavefront shaping and in vivo imaging. This cross-discipline collaboration will push the frontiers of deep tissue multiphoton imaging by experimentally measuring and simulating light-tissue interactions and developing strategies for correcting the resulting optical distortions. Predictions from models that link light-tissue interactions to circuit structure will inform optimal strategies for monitoring neural activity. Deeper and higher spatiotemporal resolution of 3D multiphoton imaging will be achieved by novel combinations of two- and three photon microscopy, high speed spatial light modulators and acousto-optic lens 3D scanning. This will allow synaptic population dynamics to be mapped at high spatiotemporal resolution across all the layers of the neocortex for the first time. This research will provide fundamental new insights into how the neocortical neurons contribute to motor learning by imaging the synaptic input across the entire dendritic tree of deep pyramidal cells. This will show how extrinsic feedforward information arriving onto the basal dendrites in deep layers is combined with intrinsic information from cortex conveyed by synaptic inputs in more superficial layers.This will reveal the nature of the information available to pyramidal cells during learning, the dendritic computations performed and provide new insight into the 'learning rules' that could be employed to adjust their synaptic weights during learning. Development of novel multiphoton methods for imaging deeper and faster than is currently possible will enable researchers to investigate the properties of brain and other tissues that were previously inaccessible. Extending the amount of information that we can acquire through microscopic observations requires an understanding of tissue properties, optics and neural dynamics. Advancing our understanding of neocortical function therefore requires a fully integrated approach and cannot be answered if the biological and physical aspects are considered separately.