Cortical feedback to the vestibular brainstem

Baylor College Of Medicine, Houston TX

Investigators

Linked publications & trials

Abstract

? DESCRIPTION (provided by applicant): The goal of this R21 application is to understand the nature of feedback from cortex to early sensory areas. Specifically, our goal is the theoretical development and testing of computational models of how feedback from the vestibular cortex to the vestibular brainstem affects the perception of translation (heading). We propose to understand how inactivation of the parieto-insular vestibular cortex (PIVC), which increases perceptual heading thresholds, alters the response properties of neurons in the vestibular nuclei. The best known correlational link between sensory neural responses and perceptual choice is known as the `choice probability'. Relevant to the origin of choice probabilities is the amount of `shared noise' among neurons, which is measured as interneuronal correlation of spike rates (`noise correlations'). First, using mechanistic models, we will generalize previously identified relationships between neural tuning, noise correlations, and choice probabilities to network architectures that incorporate recurrent connectivity. Second, we will compare various normative models that specify the function and form of neural feedback for our tasks, and derive from them predictions for the same experimental quantities. We will then test the predictions of these theories on data (neural sensitivities, choice probabilities and noise correlations) recorded from vestibular nuclei neurons before and after PIVC inactivation. Results from these studies are critical for understanding how sensory signals contribute to perception. Showing that perceptual, choice-driven signals are fed back onto the early sensory areas that themselves provide the main contributions to the thalamo-cortical network will provide valuable new insights about `active sensing' and how perception arises from the coordinated activity of multiple interconnected loops and networks.