Mechanisms for Internal Models in a Cerebellum-like Circuit

点击翻译按钮获取中文摘要

Project Summary/Abstract Humans and other animals learn and store sophisticated models of the causal relationships that govern their interactions with the world. Such internal models are likely critical for transforming ambiguous and delayed sensory data into stable perceptions and coordinated movements. For example, distinguishing external sensory input from those that are self-generated could be accomplished via an internal model that predicts the sensory consequences of an animal’s own motor commands. Despite their potential importance for both normal brain function and neurological disorders, it has proven challenging to understand how internal models are actually implemented in neural circuits. This renewal proposal applies a combination of experimental and theoretical approaches to a model system—the weakly electric fish—with unique advantages for addressing this question. Our previous studies of electric fish were successful in developing a detailed mechanistic model of how neurons at the first stage of processing in the electrosensory lobe (ELL) predict and cancel out the effects of the fish’s own electric organ discharge (EOD). However, these studies considered a highly simplified version of the true problem facing the electrosensory system. Under natural conditions, electrosensory inputs vary moment-to-moment depending both on the movements of the fish (i.e. the position of the electric organ in the tail versus electroreceptors on the skin) and the temporal pattern of EOD motor commands emitted by the fish. Solving this problem requires a more complex internal model, akin to those believed to be generated in the mammalian brain. In addition, past models ignored key features of ELL circuitry, such as plasticity of inhibitory synapses, which likely play key functional roles (both in ELL and in other vertebrate brain circuits). By addressing these issues the proposed research will provide general insights into how neural circuits contribute to distinguishing self-generated from external stimuli. The proposed studies will also provide direct links between neural representations, well-defined circuitry, synaptic plasticity, and a behaviorally relevant systems level function. Though forging such links is a primary goal of neuroscience, there are still relatively few cases in which they can actually be made.