Synaptic integration and visual processing in health and disease, Ikuko Teramitsu Smith, University of California, Santa Barbara

Image

Ikuko Smith
Assistant Professor
Psychological & Brain Science
University of California, Santa Barbara

Synaptic integration and visual processing in health and disease

Neurons process tens of thousands of synaptic inputs arriving at their dendritic arbor to reliably compute appropriately tuned outputs to support sensory processing. My lab seeks to understand the role of active neuronal dendrites in this process. Dendrites can fire regenerative electrical spikes much like axons, and the underlying voltage-dependent nonlinear mechanisms increase the neuron’s computational capacity. How such active mechanisms are engaged and to what extent they play a functional role in an intact neural circuitry, however, remain elusive. Layer 2/3 pyramidal neurons in the mouse visual cortex, with their broad dendritic arbors that span across multiple cortical layers, offer an opportunity for studying the dendritic mechanisms by which a myriad of synaptic inputs is integrated to generate stimulus feature-specific responses. One of the fundamental computations known to be performed by these neurons is orientation tuning. Using direct dendritic patch-clamp recordings in awake mice, we have previously found that locally generated NMDA receptor-dependent dendritic spikes are not only visually evoked, but also exhibit reliable orientation tuning. Moreover, these dendritic spikes enhance the neuronal output tuning. However, the spatial organization of dendritic spiking, and their responses compared to local synaptic inputs, have remained unclear, and we are addressing these questions with our current studies. We employ in vivo two-photon calcium (2p-Ca²⁺) imaging with a genetically encoded calcium indicator (GECI), GCaMP8m, to map hundreds of synaptic inputs and simultaneously detect dendritic spikes, which are identified by their characteristic local calcium dynamics. We find that a simple linear sum of the individual synaptic inputs can account for the preferred orientation of the neuronal output, but the resulting curves overestimate the tuning width. By contrast, local dendritic Ca²⁺ spikes can provide a higher fidelity prediction of both the preferred orientation and tuning width of neuronal output. Taken together, our data provide synaptic-scale physiological evidence supporting the two-layer input integration model in which active conductances in the local dendrites nonlinearly enhance feature-specific inputs while other inputs are subject to attenuation due to the passive electrical properties of dendrites. I will also share some data from our longitudinal studies of dendritic function during pathological progression in a mouse model of a neurodegenerative disease, P301S.