Dissecting the entorhinal cortex target signal for behavioral timescale synaptic plasticity - Project Summary How individual neurons adaptively update their outputs to respond to a changing environment presents a unique learning challenge for the brain. Salient cues and events evolve over seconds to minutes, while standard Hebbian synaptic plasticity occurs on the scale of milliseconds, leaving the mechanism for forming memories of salient experiences unclear. A novel synaptic learning rule that bridges this gap, behavioral timescale synaptic plasticity (BTSP), was recently demonstrated in the entorhinal cortex (EC)-to-hippocampus circuit. To connect cues with new neuronal responses, BTSP relies on an upstream target signal from EC layer 3 – the first of its kind observed in the mammalian brain. This target signal opens a seconds-long time window for integrating behaviorally-relevant inputs within hippocampal region CA1, yet its mechanistic underpinnings remain unexplored. This proposal will test critical hypotheses about the formation, sculpting and maintenance of the BTSP target signal by specific layers and interneuron classes of the entorhinal cortex in awake, behaving animals. Using high-yield linear probes, we will simultaneously measure the activity of neural ensembles across EC layers to dissect the components of the instructive signal before, during, and after learning of novel cue-reward relationships in a head-fixed spatial learning task. We hypothesize that at both the single cell and population level different EC layers will remap their spatial representations during learning to produce a decodable target signal for hippocampus. Combining transgenic mouse lines and optogenetic photo- tagging with my recording will enable precise identification of EC layer 3 neurons during natural behavior. Inhibition of EC outputs to CA1 is known to impair spatial learning, so we will test whether inhibition of CA1 feedback to EC is also necessary for natural learning and the production of the target signal. Finally, the relative specificity of the target signal to behaviorally salient cues suggests intrinsic circuit mechanisms for stabilizing EC output, likely mediated by interneurons. We will therefore use optogenetic suppression of specific interneuron subtypes to acutely disrupt the circuit during learning. We predict that suppressing parvalbumin interneuron function will disinhibit principal cell output and decrease signal to noise ratio of the target signal, while suppressing somatostatin interneuron function will reduce the goal-targeted accuracy of the target signal. As learning is crucial to nearly all brain regions, these results will have a wide impact on cellular, systems, and cognitive neuroscience in addition to generating new hypotheses about neurodegenerative diseases targeting memory. The fellowship will also provide crucial training opportunities for Dr. Wilmerding in the collection, analysis, and modeling of high-dimensional time series data collected using cutting-edge electrophysiology tools.
