Innovators in Cognitive Neuroscience: December 2, 2020
Alexandra Keinath
Postdoctoral Fellow
Charitable cause: Mazzoni Center
Experience-dependent pattern separation and pattern completion in the mouse hippocampus
Abstract: Extensive research has implicated the hippocampus as a key locus of episodic memory. Critical among the computations governing hippocampal-dependent mnemonic processing are ‘pattern separation’ and its complement ‘pattern completion’. Pattern separation refers to the mapping of similar inputs, for example ‘car’ and ‘cat’, to dissimilar representations. Pattern completion refers to the mapping of dissimilar inputs, for example ‘car’ and ‘automobile’, to similar representations. A large body of work has fleshed out how these processes, instantiated at different stages within the hippocampal formation, may shape hippocampal representations and episodic memory alike. Crucially though, whether and how inputs are transformed throughout this circuit is hypothesized to depend on past experience, yet little is known about this dependency. Here I address this knowledge gap. In part one, I introduce a novel behavioral paradigm which induces hippocampal pattern separation on the basis of past experience alone, without changes to sensory cues or explicit task demands. Next, I show that selective chemogenetic inactivation of DG-CA3 circuitry in this paradigm completely and reversibly abolishes pattern separation in downstream CA1 while sparing other representational components. In part two, I adapt existing geometric morph paradigms, in which one familiar environment is progressively deformed to resemble another familiar environment, to assay the development of hippocampal pattern separation and completion across extended experience (>30 days). With this rich dataset I identify multiple previously unrecognized dynamics, including the evolution of pattern separation/completion over experience and contributions of individual cells to distinct components of contextual representation and representational drift over time. Together, these results characterize the critical and continued influence of past experience on hippocampal pattern separation and completion, demonstrate a causal and selective contribution of DG-CA3 circuitry to experience-dependent hippocampal pattern separation, and set the stage for further causal tests of the specific circuitry mediating these dynamics.