Structural Correlates of Related Contextual Learning in Retrosplenial Cortex

Abstract

The discovery of dendritic spines by Santiago Ramón y Cajal is one of the most important chapters in the history of neuroscience. Apart from establishing the neuron doctrine, it presented dendritic spines as the information storage unit. More than century of further studies has only strengthened the postulate made by Cajal. With the advent of two photon microscopy, it has been found that dendritic spines are highly dynamic entities and displays a basal turnover over the temporal scale of days. Dendritic integration and the possible role of spines dynamics in memory allocation and organisation has only been postulated theoretically. Recent in vivo work using longitudinal two- photon microscopy has shown that repetitive motor and contextual learning induces clustered addition of new spines in primary motor cortex and retrosplenial cortex. In this dissertation, I further test the encoding of multiple related memories at the level of dendritic spines. I established and characterized the state of art two photon microscope and related laboratory protocols to allow for chronic in vivo imaging of dendritic spines in live mice undergoing training in multiple contextual memories across days. In this process, I developed novel optical methods to increase depth and contrast of two photon microscopy. For the first time, we have developed a novel method for quantitatively measuring and characterizing the extent of cooperativity/clustering seen among the spines present in the dendrites using spatial autocorrelation. This method enabled us to show that dendrites encode multiple memories through spine clusters and preserves their interrelation through spatially selective and clustered loss. Our study shows that contrary to common belief, spine loss is not necessarily an indicator of memory loss; instead, they are standard and an integral feature in encoding related memories. Using spatial autocorrelation, we show that spine loss occurs during related memory formation. In this process, we formulate a method for probabilistic...