Time-Series Statistical Applications to Mammalian Cerebral Physiology for Understanding Network Relations and Building State-Space Projections

在人类和大型哺乳动物中，大脑生理过程很少孤立地发生。然而，我们现有的关于压力-流量动力学、脑血管控制、氧气和营养输送、自主神经系统调节、神经电生理学以及它们与心血管生理学的整合等脑生理系统之间的相互关系的知识受到了一些限制。特别是，以前的工作通常发生在人类或大型哺乳动物的小群体中，使用低分辨率的生理数据，在没有考虑大系统相互作用的情况下探索单个系统之间的关系。这导致我们对人类和大型哺乳动物的大脑生理学的理解存在巨大的知识鸿沟，缺乏对生理系统网络关系的全面了解。通过了解网络关系，在高分辨率下考虑整个大脑和心血管生理组，我们可以提高我们对更大的大脑生理系统的理解，并潜在地促进生理状态建模的生成，以及点和区间生理状态预测。NSERC联盟国际催化剂项目利用现有的全球唯一的高保真、高频多模式脑生理数据集和来自加拿大马尼托巴大学和瑞典卡罗林斯卡研究所的专业知识，旨在应用时间序列、向量和状态空间统计方法(通常应用于市场分析或天体物理学)，借助机器学习，了解网络生理关系，并生成可应用于人类和大型哺乳动物脑生理学的新型生理状态空间预测模型。这里的发现将推动脑生理学、统计学方法、数据科学和机器学习应用等自然科学和工程(NSE)领域的发展，同时促进两个卓越中心之间建立长期的国际联系，这些中心专注于弥合我们对哺乳动物大脑生理学的基本理解中的知识差距。

Cerebral physiologic processes rarely occur in isolation within humans and large mammals. However, our existing knowledge of the inter-relationships between cerebral physiologic systems of pressure-flow dynamics, cerebrovascular control, oxygen and nutrient delivery, autonomic nervous system modulation, neural electrophysiology, and their integration with cardiovascular physiology has suffered from several limitations. In particular, previous work has often occurred in small groups of humans or large mammals, with low-resolution physiologic data, exploring single system relationships without accounting for large systems interactions. This has led to wide knowledge gaps in our understanding of cerebral physiology in humans and large mammals, where comprehensive understanding of the physiologic systems network relationships is absent. By understanding network relationships, taking the entire cerebral and cardiovascular physiome into account in high-resolution, we can improve our understanding of the greater cerebral physiologic system and potentially facilitate the generation of physiology state modelling, with point and interval physiology state forecasting. Leveraging existing globally unique high-fidelity, high-frequency multi-modal cerebral physiologic data sets and expertise from both the University of Manitoba (Canada) and Karolinska Institute (Sweden), this NSERC Alliance International Catalyst project aims to apply time-series, vector and state-space statistical methodologies (commonly applied in market analysis or astrophysics), with the aid of machine learning, to understand the network physiologic relationships and generate novel physiologic state-space forecasting models that can be applied to both human and large mammal cerebral physiology. Discoveries here will advance the natural sciences and engineering (NSE) fields of cerebral physiology, statistical methodologies, data science and application of machine learning, while facilitating the building of long-term international links between two centers of excellence focused on bridging knowledge gaps in our fundamental understanding of cerebral physiology in mammals.