Cerebral Hemodynamics Integrated Estimation with PHysics (CHIEPH)

Long-duration spaceflight exposes astronauts to microgravity conditions that induce significant physiological adaptations, including cephalad fluid redistribution, altered venous drainage, changes in vascular resistance, and modifications of cerebral perfusion regulation. These mechanisms may affect the cerebrovascular system and contribute to adverse outcomes such as orthostatic intolerance, venous congestion, and neuro-ocular alterations. However, direct assessment of cerebral hemodynamics during spaceflight remains limited by the scarcity of in-flight measurements and by the operational constraints of advanced imaging techniques. In this context, this thesis presents CHIEPH, a physics-informed computational framework for the reconstruction and scenario-based analysis of cerebral hemodynamics under sparse-data conditions. A subject-specific CoW-centered cerebrovascular graph is reconstructed from three-dimensional magnetic resonance angiography data using SimVascular. Vessel paths, segmentation contours, and geometry-derived quantities are converted into a reduced-order vascular network suitable for one-dimensional hemodynamic modeling. On this image-derived vascular graph, a Physics-Informed Neural Network is used to reconstruct spatiotemporally resolved hemodynamic variables, including blood velocity, pressure, volumetric flow rate, and vessel cross-sectional area. The reconstruction problem is formulated as an inverse problem in which sparse synthetic observations are combined with conservation laws of mass and momentum, a pressure-area constitutive relation, and interface constraints at vascular junctions. By embedding the governing physical equations into the loss function, the framework reduces the dependence on fully prescribed boundary conditions and provides physically consistent estimates of unobserved variables. A controlled real-geometry numerical study is conducted to evaluate the baseline terrestrial state and a 14 cases matrix of microgravity-related physiological perturbations. The scenario matrix includes changes in effective cranial pressure, downstream venous pressure, distal vascular resistance, vessel stiffness, inlet flow magnitude, and pulsatility. The results show stable baseline reconstruction and indicate that inflow magnitude is the dominant driver of global velocity and flow-rate deviations in the simulated scenario space, while pressure and cross-sectional area remain comparatively constrained. This work contributes to the development of non-invasive, physics-informed computational tools for cerebrovascular analysis in data-limited environments. Beyond reconstructing hemodynamic fields, CHIEPH introduces a framework for virtual microgravity stress testing, with potential future applications in pre-flight susceptibility assessment, astronaut-specific monitoring, and early risk stratification for cerebrovascular and neuro-ocular complications.