Project Abstract The question of why animals sleep has puzzled scientists for over a century. Contemporary research suggests that sleep serves as a compensatory mechanism for brain plasticity, restoring the network to a homeostatic set- point known as criticality. However, the precise mechanisms between sleep and neural plasticity are still subjects of intense debate, and the information processing aspect of this problem hasn’t been addressed experimentally. This proposal aims to address this gap by positing the hypothesis that experience-dependent plasticity in neuronal circuitry progressively undermines indices of an optimal computational regime, and proportionally increases the likelihood of sleep. I will adopt a two-pronged approach, leveraging the diverse expertise of my mentorship team. The first objective is to develop a machine learning algorithm that predicts the mechanisms of neural criticality and its homeostasis. This model will challenge existing machine learning approaches, which often lack explanatory power, by extracting dynamical rules from existing datasets using cutting-edge modeling and optimization techniques. The second objective involves empirical testing to ascertain whether a plasticity threshold exists at the network level that triggers sleep. This will be conducted through continuous neuronal recordings in freely behaving mice, coupled with methods to induce plastic change in the network. The third objective is to construct a biophysically grounded mathematical model to identify mechanisms that stabilize network information processing in the face of plastic changes. This model will incorporate methods from control theory and dynamical systems and will be validated against the rules derived in the first objective and the data collected in the second objective. This interdisciplinary project is groundbreaking in its attempt to offer a dynamical and mathematical explanation for why plasticity requires sleep. Furthermore, in contrast to many other works that focus on either theory or experiment, here I will support my theoretical work with empirical data. Overall, the project promises to make significant contributions to both theoretical and applied neuroscience. It seeks to develop a predictive model for sleep/wake behavior based on fundamental rules and to elucidate the functional mechanisms by which sleep serves the healthy brain. By directly testing a dynamical/mathematical explanation of why plasticity requires sleep, the project stands to resolve a long- standing question in neuroscience. Furthermore, uncovering these mechanisms will fill the important gap of adaptation for a possible fundamental framework of brain dynamics.

Project Narrative The complex relationship between sleep and neural plasticity remains a subject of debate. In this project I hypothesize that experience-dependent changes in neuronal circuits undermine optimal information processing, thereby increasing the need for sleep. To support this hypothesis, a machine learning algorithm will be developed to extract the dynamical rules for sleep homeostasis, experimental work will be conducted to test the plasticity thresholds in neural circuits, and a model will be developed to build predictions and test the mathematical limits of such a problem.