Andreas A. Linninger

Mechanistic simulations of blood flow and oxygen exchange showed regions of cortical tissue tolerating substantial increase in local oxygen consumption (CMRO ₂ ) before reaching hypoxia (pO ₂ < 10 mmHg). The observed robustness in O ₂ supply was attributed to overcapacity in convective oxygen transport in the pial arterial network combined with a surplus in the number of capillary flow paths. Microcirculatory flux analysis suggests that network induced hemodynamic flow patterns impart intrinsic reserve to protect the brain against perfusion variances or metabolic demand surges during activation. Furthermore, oxygen transport in cortical tissue is characterized by two regimes: in the transport zone— centered on penetrating arteriole trees composed of a single penetrating vessel connected to the post-arteriole capillary transition zone—strong diffusion supports high oxygen tension with only modest contribution from capillaries. This regime transitions into the terminal/reactive zone where oxygenation is sensitive to capillary density and perfusion. Quasi-dynamic simulations also enabled reconstruction of the BOLD signal underlying functional imaging. Simulations at single micron resolution further show that age-related reductions in arterial saturation and systemic hematocrit were sufficient to induce hypoxic tissue pockets in the terminal zone at nominal perfusion (CBF) and metabolic activity (CMRO ₂ ), and neutrophil adhesion induced capillary flow stalling further exacerbates hypoxia.

Mechanistic simulations of blood flow and oxygen exchange showed regions of cortical tissue tolerating substantial increase in local oxygen consumption (CMRO2) before reaching hypoxia (pO22 supply was attributed to overcapacity in convective oxygen transport in the pial arterial network combined with a surplus in the number of capillary flow paths. Microcirculatory flux analysis suggests that network induced hemodynamic flow patterns impart intrinsic reserve to protect the brain against perfusion variances or metabolic demand surges during activation. Furthermore, oxygen transport in cortical tissue is characterized by two regimes: in the transport zone — centered on penetrating arteriole trees composed of a single penetrating vessel connected to the post-arteriole capillary transition zone — strong diffusion supports high oxygen tension with only modest contribution from capillaries. This regime transitions into the terminal/reactive zone where oxygenation is sensitive to capillary density and perfusion. Quasi-dynamic simulations also enabled reconstruction of the BOLD signal underlying functional imaging. Simulations at single micron resolution further show that age-related reductions in arterial saturation and systemic hematocrit were sufficient to induce hypoxic tissue pockets in the terminal zone at nominal perfusion (CBF) and metabolic activity (CMRO ₂ ), and neutrophil adhesion induced capillary flow stalling further exacerbates hypoxia.

While transcranial neuroimaging of individual capillary function holds transformative potential for diagnostics, it has proven difficult to achieve. Superresolution ultrasound, while capable of achieving micron-scale resolution, relies on the accumulation of multiple microbubble events, a method inherently limited by the exceedingly low probability of observing such events within capillaries. We present single capillary reporters (SCaRe), a paradigm-shifting approach that utilizes the complete flow trajectory information extracted from individual microbubbles to directly image single capillaries. This method allows for transcranial reconstruction and functional assessment of deep capillary networks in the entire brain. We employed computational simulations and pathological neuroinflammation models to quantify and validate metrics such as capillary transit-time and capillary stalling. Importantly, we demonstrated SCaRe’s ability to resolve immune responses to injury at the single capillary level, markedly broadening research avenues for exploring microvascular dysfunction across diverse neurological conditions.

Due to the scarcity of reliable in vivo data, the pharmacokinetics of intrathecally (IT) administered drugs remain inadequately quantified. Designing new therapies is further hindered by variability in experimental methods, inter-individual and inter-species differences, and poor reproducibility across animal and human studies. To address these limitations, we developed an anatomically accurate, subject-specific replica of the cerebrospinal fluid (CSF)-filled spaces of the human central nervous system (CNS) using a multistep mold/casting process. The 3D-printed, transparent, deformable CNS phantom enables precise control of the infusion and physiological parameters, allowing systematic generation of reliable and repeatable biodispersion data for lumbar IT infusion protocols. Pulsatile artificial CSF flow within the closed system was tuned to replicate subject-specific stroke volumes and flow rates observed in MRI. The model’s optical clarity facilitated high-speed visualization and tracking of tracer dispersion, exceeding the temporal resolution of current neuroimaging techniques. An experimental series spanning physiologically relevant CSF and infusion conditions enabled quantification of the spatiotemporal distribution of IT-administered tracers. Inversion of the parabolic diffusion equation provided estimates of the coefficient of effective dispersion . A distributed pharmacokinetic model was used to evaluate the influence of chemical kinetics and mass transfer on tracer behavior. The proposed experimental apparatus for in vitro design of IT therapies offers a complementary or alternative approach to traditional trial-and-error animal studies.

Brain microcirculation is the intricate network of blood vessels encompassing arterioles, capillaries, and venules that delivers oxygen and nutrition and removes waste products necessary for the proper functioning of brain tissue.^1,2 It plays a critical role in maintaining homeostasis associated with metabolic requirements and oxygen distribution to the tissue in the brain. In particular, brain capillaries play a pivotal role in facilitating the exchange of essential nutrients and oxygen to support neurons. However, in the presence of neurological disorders, such as Alzheimer’s disease³ or stroke,^{4 – 6} this microcirculatory system can be disrupted due to capillary stalling.

Quantifying anatomical and hemodynamical properties of the brain vasculature in vivo is difficult due to limited spatiotemporal resolution neuroimaging, variability between subjects, and bias between acquisition techniques. This work introduces a metabolically inspired vascular synthesis algorithm for creating a digital representation of the cortical blood supply in humans. Spatial organization and segment resistances of a cortical vascular network were generated. Cortical folding and macroscale arterial and venous vessels were reconstructed from anatomical MRI and MR angiography. The remaining network, including ensembles representing the parenchymal capillary bed, were synthesized following a mechanistic principle based on hydrodynamic efficiency of the cortical blood supply. We evaluated the digital model by comparing its simulated values with in vivo healthy human brain measurements of macrovessel blood velocity from phase contrast MRI and capillary bed transit times and bolus arrival times from dynamic susceptibility contrast. We find that measured and simulated values reasonably agree and that relevant neuroimaging observables can be recapitulated in silico. This work provides a basis for describing and testing quantitative aspects of the cerebrovascular circulation that are not directly observable. Future applications of such digital brains include the investigation of the organ-wide effects of simulated vascular and metabolic pathologies.

Background: Traditionally, there is a widely held belief that drug dispersion after intrathecal (IT) delivery is confined locally near the injection site. We posit that high-volume infusions can overcome this perceived limitation of IT administration.Methods: To test our hypothesis, subject-specific deformable phantom models of the human central nervous system were manufactured so that tracer infusion could be realistically replicated in vitro over the entire physiological range of pulsating cerebrospinal fluid (CSF) amplitudes and frequencies. The distribution of IT injected tracers was studied systematically with high-speed optical methods to determine its dependence on injection parameters (infusion volume, flow rate, and catheter configurations) and natural CSF oscillations in a deformable model of the central nervous system (CNS).Results: Optical imaging analysis of high-volume infusion experiments showed that tracers spread quickly throughout the spinal subarachnoid space, reaching the cervical region in less than 10 min. The experimentally observed biodispersion is much slower than suggested by the Taylor–Aris dispersion theory. Our experiments indicate that micro-mixing patterns induced by oscillatory CSF flow around microanatomical features such as nerve roots significantly accelerate solute transport. Strong micro-mixing effects due to anatomical features in the spinal subarachnoid space were found to be active in intrathecal drug administration but were not considered in prior dispersion theories. Their omission explains why prior models developed in the engineering community are poor predictors for IT delivery.Conclusion: Our experiments support the feasibility of targeting large sections of the neuroaxis or brain utilizing high-volume IT injection protocols. The experimental tracer dispersion profiles acquired with an anatomically accurate, deformable, and closed in vitro human CNS analog informed a new predictive model of tracer dispersion as a function of physiological CSF pulsations and adjustable infusion parameters. The ability to predict spatiotemporal dispersion patterns is an essential prerequisite for exploring new indications of IT drug delivery that targets specific regions in the CNS or the brain.

Intrathecal administration is an important mode for delivering biological agents targeting central nervous system (CNS) diseases. However, current clinical practices lack a sound theorical basis for a quantitative understanding of the variables and conditions that govern the delivery efficiency and specific tissue targeting especially in the brain. This work presents a distributed mechanistic pharmacokinetic model (DMPK) for predictive analysis of intrathecal drug delivery to CNS. The proposed DMPK model captures the spatiotemporal dispersion of antisense oligonucleotides (ASO) along the neuraxis over clinically relevant time scales of days and weeks as a function of infusion, physiological and molecular properties. We demonstrate its prediction capability using biodistribution data of antisense oligonucleotide (ASO) administration in non-human primates. The results are in close agreement with the observed ASO pharmacokinetics in all key compartments of the central nervous system. The model enables determination of optimal injection parameters such as intrathecal infusion volume and duration for maximum ASO delivery to the brain. Our quantitative model-guided analysis is suitable for identifying optimal parameter settings to target specific brain regions with therapeutic drugs such as ASOs.

Oxygen transfer from blood vessels to cortical brain tissue is representative of a class of problems with mixed‐domain character. Large‐scale efficient computation of tissue oxygen concentration is dependent on the manner in which the tubular network of blood vessels is coupled to the tissue mesh. Models which explicitly resolve the interface between the tissue and vasculature with a contiguous mesh are prohibitively expensive for very dense cerebral microvasculature. We propose a mixed‐domain mesh‐free technique whereby a vascular anatomical network (VAN) represented as a thin directed graph serves for convection of blood oxygen, and the surrounding extravascular tissue is represented as a Cartesian grid of 3D voxels throughout which oxygen is transported by diffusion. We split the network and tissue meshes by the Schur complement method of domain decomposition to obtain a reduced set of system equations for the tissue oxygen concentration at steady state. The use of a Cartesian grid allows the corresponding matrix equation to be solved approximately with a fast Fourier transform‐based Poisson solver, which serves as an effective preconditioner for Krylov subspace iteration. The performance of this method enables the steady‐state simulation of cortical oxygen perfusion for anatomically accurate vascular networks down to single micron resolution without the need for supercomputers.