A synthetic 3D human cerebrovascular model informed by histology for simulating the cortical depth-dependent BOLD fMRI signal.

Recent advances in functional magnetic resonance imaging (fMRI) using the blood oxygenation level-dependent (BOLD) signal at ultra-high field (≥7T) permit mesoscopic investigations of neurovascular function. However, interpreting the BOLD signal remains challenging because it is an indirect measure of neuronal activity influenced by complex vascular architectures and hemodynamic changes. Existing biophysical models often rely on rodent data, limiting their accuracy for human neuroimaging. We introduce 3D VAMOS (three-dimensional VAscular MOdel based on Statistics), a computational framework that generates synthetic 3D vascular networks for specific human cortical regions. By incorporating histological features, such as vessel volume fractions, tortuosity, and artery-to-vein ratios, 3D VAMOS integrates hemodynamic parameters and biophysical processes to estimate depth-dependent BOLD contributions. To ensure physiological plausibility, we validated 3D VAMOS by comparing simulations of synthetic mouse cortex against realistic models derived from two-photon microscopy. Results showed that regional variability in vessel architecture and cortical thickness significantly modulates laminar BOLD responses. Comparisons between human (visual and motor cortices) and mouse models revealed distinct BOLD profiles reflecting species-specific vascular distributions, where superficial vessels disproportionately influence signal detection. Furthermore, simulations demonstrated that gradient-echo BOLD signals emphasize large-vessel contributions, while spin-echo signals better capture microvascular effects. This highlights a critical sequence-dependent sensitivity in laminar fMRI. Additionally, localized simulations of neuronal activity showed that BOLD profiles depend on vessel-specific changes in blood volume and oxygenation. Thus, 3D VAMOS provides a robust computational framework for understanding BOLD changes across cortical depth for both microvessels and larger intracortical veins. As a computationally efficient and scalable tool, it offers a biologically informed framework to interpret cortical depth-dependent signals, to refine high-resolution imaging protocols, and to explore healthy and pathological brain functions. By bridging the gap between microscopic histology and macroscopic neuroimaging, 3D VAMOS provides a robust foundation for decoding the human brain function.