Significance - Cerebral autoregulation is the remarkable control task of maintaining constant cerebral blood flow over a wide range of disturbances such as posture change, physical activity, or changes in cardiac output. In addition, the brain is capable of dynamically up-regulating blood flow to very specific, locally confined territories of the brain to support the metabolic reactions of neuronal firing, referred to as functional hyperemia. This servo problem is accomplished in fractions of a second in a small confined cortical domain without affecting blood perfusion to other cortical regions. Though physiological and anatomical details of cerebral autoregulation and functional hyperemia have been well researched, the systems control behavior is not well understood. We propose a holistic systems approach for the investigation of the fundamental principles of decentralized distributed blood flow control in the entire human brain. Our project plan foresees a non-conventional pioneering approach integrating control theory, computational fluid dynamics, biochemistry, neurophysiology, and biomedical imaging. While our interdisciplinary approach entails a high risk approach, novel insights from research on complex dynamics and regulation of the human brain may be transformative, making the project an excellent candidate for the EAGER funding.
We have already constructed a morphologically accurate, physiologically consistent, multi-scale computer network model of the entire cerebral vasculature to predict the functional interaction of structural and hemodynamic parameters in tissue oxygen perfusion. Systems engineering methods will be used to adapt this model for the investigation of autoregulatory control and functional hyperemia of the human brain. To research the distributed decentralized control functions of the brain, we will incorporate the several biochemical and physiological principles: (i) active vasodilation through vasoactive signaling molecules, (ii) feed forward signal processing, and (iii) distributed oxygen-sensing chemo-receptors in the brain tissue.
This research will introduce spatially distributed, time-dependent simulations based on first principles of fluid dynamics and passivity control theory of the entire brain1. The decentralized distributed control mechanisms will be shown to require only biochemical sensors sensitive to neural tissue oxygenation and local wall shear stress without the need for centralized supervision. We will investigate the brain?s remarkable stability and specificity in achieving highly localized blood flow distribution without altering flow to adjacent cortical territories. This localized specificity of cortical blood flow has been observed in functional Magnetic Resonance Imaging (fMRI), but the hemodynamic control principles are not known. We will predict and explain the time delay between neuronal firing, changes in relative cerebral blood flow (rCBF) and tissue oxygen perfusion. This first principles model will explain the physical and chemical kinetic principles underlying fMRI, which are currently under debate in the medical imaging community. The final systems model of cerebral blood flow control will predict the decentralized, distributed, dynamic behavior of cerebral blood flow in response to local neuronal firing and stable rCBF despite inlet arterial blood pressure fluctuations. Model predictions will be validated using advanced distributed mathematical programming techniques to match a spectrum of temporally and spatially distributed data acquired in vivo by medical imaging modalities.
Broader Impact:
This project offers -for the first time- a dynamic computer model to elucidate the principles of cerebral autoregulation by integrating control theory, computational fluid mechanics and medical imaging into a single visionary project plan. The insights from investigating rCBF dynamics of the entire brain will unravel nature?s design for robust, distributed and decentralized control. Due to the complex distributed blood flow demand in the human cortex, a systems approach is needed to quantify and characterize the underlying dynamic mechanisms. The knowledge gain is expected to create new opportunities for controlling distributed technical systems such as artificial organs, dialysis machines or process engineering to hypothermal stroke treatments. To achieve a broader impact, the final computational model will be disseminated to the research community via a comprehensive data sharing plan and distribution via the lectures in the UIC CAVE-2 virtual reality environment at UIC. Undergraduate students and high school teachers will benefit from the intellectual core created in this through the NSF-sponsored REU and RET programs directed by the PI.
