Item

Abstract

In the mouse olfactory bulb glomerular layer, computations are largely mediated by a set of interneurons, known as the periglomerular cells (PGCs). I hypothesized that PGC subtypes exhibit differences in synaptic connectivity with mitral (MCs) and tufted cells (TCs), and therefore, have functionally-distinct roles in odor processing. Through a correlative electron microscopy (EM) reconstruction of a genetically-identified glomerulus, I investigated the differences in synaptic connectivity between neurochemically-distinct PGC subtypes and projection neurons. My aim was to explain functional differences of olfactory projection neurons by the intraglomerular connectivity. I combined cellular-resolution protein expression of neuronal cell types with synaptic connectivity and physiology in order to generate network-level computational models. By preserving extracellular space in acutely fixed sections, I developed a permeabilization-free immunohistochemistry protocol compatible with EM, which permits deep antibody penetration and preserves ultrastructure. This approach allows for the correlation of morphological and physiological properties of neurons through a dense reconstruction of the synaptic connectivity. To investigate the PGC-subtype specific involvement in olfactory bulb computations, I generated a rate-based computational model using the measured synaptic connectivity. I have demonstrated that a subset of PGCs selectively inhibits MCs and allows for differences in projection neuron phase preference. The connectivity-based model also predicts the mechanism underlying MC odor concentration encoding. This correlative glomerular layer reconstruction and the subsequent computational modeling reveals previously unidentified wiring specificity underlying OB computations.