Bibliography
References
33 sources in the bibliography, 32 cited in the current manuscript.
Each entry records what the source supplies and where it enters the model. The labels separate primary formulation, numerical method, parameter source, external validation and clinical context.
Parametercited 5×
Provides The Hill relation and retinal oxygen parameters: HD = 0.4, CHb,max = 0.5 mL(O₂)/mL, P50 = 26 mmHg, nH = 2.7, SH(Pa) = 0.92, P0 = 10 mmHg and χb = 3.1 × 10⁻⁵ mL(O₂) mL⁻¹ mmHg⁻¹.
In the manuscript Oxygen scenario parameters
Methodcited 1×
Provides Stability and convergence of the method of fundamental solutions on analytic domains, including source placement, coefficient growth and finite-precision effects.
In the manuscript Fundamental-solution kernels
Contextnot cited in v13
Provides Robust vessel segmentation in fundus images. Relevant to extracting vascular geometry from photographs, but not cited in the current manuscript.
In the manuscript Not cited in v13
Methodcited 2×
Provides Enriched spectral representations that incorporate a known leading singular structure to improve approximation of weakly singular solutions.
In the manuscript Singular enrichment
Modelcited 1×
Provides A duality formulation for convection–diffusion with mixed, Neumann and Fourier boundary conditions and measure data. Clarifies what a measure-data formulation does and does not imply for a singular load.
In the manuscript Weak formulation with singular loads
Methodcited 1×
Provides Error analysis with variational crimes. Controls the mismatch between an exact form and its quadrature-assembled counterpart alongside approximation and stability.
In the manuscript Assembly error analysis
Modelcited 1×
Provides Sobolev-space weak formulation for second-order elliptic problems. Supplies the energy-space background that lowers differentiability requirements for the coupled problem.
In the manuscript Function spaces; weak formulation
Validationcited 3×
Provides Retinal vessel oxygen saturation in healthy individuals. The reported mean venous saturation of 55.6% is used as an external descriptive comparator, not as a calibration target.
In the manuscript External cohort comparison
Modelcited 1×
Provides Singularity removal for coupled 1D–3D flow with line sources. Supplies the decomposition idea for embedding vessel sources in a three-dimensional tissue region.
In the manuscript Singular source treatment
Modelcited 1×
Provides A splitting method for elliptic equations with line sources. Separates the singular line contribution from the regular remainder, which is the structure used to organise the source terms here.
In the manuscript Singular source treatment
Methodcited 1×
Provides Perturbation analysis for linear systems and eigenvalue problems in finite precision. Supplies the language for separating coefficient sensitivity from assembly error.
In the manuscript Conditioning and perturbation bounds
Methodcited 1×
Provides Approximation results for the method of fundamental solutions, covering function and derivative convergence under stated conditions.
In the manuscript Fundamental-solution kernels
Contextcited 1×
Provides Evidence that retinal dopamine participates in light-adapted vision and that its loss impairs visual function in animals.
In the manuscript Clinical motivation
Modelcited 1×
Provides Classical thermodynamic analysis of membrane permeability. Treats solvent permeability, solute permeability and their coupling as distinct processes, which motivates the exchange law used here.
In the manuscript Wall exchange law
Methodcited 1×
Provides Quadrature by expansion for layer potentials. Separates a local expansion from quadrature error near singular kernels, a template for near-singular evaluation.
In the manuscript Near-singular kernel evaluation
Modelcited 1×
Provides Embedded mixed-dimension 1D–3D perfusion with distributed sources and a consistent flux correction. Informs the finite-width terminal source regions and their normalisation.
In the manuscript Terminal source regions
Methodcited 1×
Provides The reciprocal condition-number estimate computed from an LU factorization (ZGECON), used as the diagnostic for the assembled complex systems.
In the manuscript Condition-number estimates
Contextcited 1×
Provides Swept-source OCT angiography measurements of capillary plexus pattern and density from the fovea to the periphery in healthy eyes.
In the manuscript Heterogeneous capillary density
Contextcited 1×
Provides A survey of log-normal distributions across the sciences. Supplies the distributional convention for terminal-footprint radii; it is not a fitted measurement of retinal tissue.
In the manuscript Terminal-scale distribution
Modelcited 18×
Provides Spherical-cap pressure kernels, the vascular-network construction, elastic vessel-wall relations, terminal averaging rules and the disk image construction. The primary source for the geometry and pressure background, including equations (3.23), (3.35) and (3.58)–(3.63).
In the manuscript Geometry and flow inputs; pressure kernels; appendices
Modelcited 6×
Provides The published multiscale retinal-flow model and its analytic disk solution. Supplies the starting formulation and the singular-decomposition strategy that is extended here to the spherical cap.
In the manuscript Introduction; model formulation; analytic disk solution
Modelcited 1×
Provides A robustness study for capillary conductivity and local exchange impairment using an effective spectral–Galerkin resistance formulation. Remark III.1 records the weaker integrable singularity available in the variable-conductivity pressure problem.
In the manuscript Conductivity perturbation; singular remainder
Parametercited 2×
Provides A review of retinal oxygen transport; source for tissue solubility χt = 2.4 × 10⁻⁵ and the conventional tissue diffusion values used in the oxygen scenario.
In the manuscript Oxygen scenario parameters
Methodcited 2×
Provides Partition-of-unity finite elements with local, compactly supported enrichment. The model for restricting an additive correction to a source neighbourhood.
In the manuscript Local correction support
Contextcited 1×
Provides Clinical overview of diabetic retinopathy and its link to chronic hyperglycaemia and retinal microvascular damage.
In the manuscript Clinical motivation
Methodcited 1×
Provides Standard Ferrers-function conventions, including the Condon–Shortley phase, for Legendre expansions on the interval (−1, 1).
In the manuscript Legendre functions
Methodcited 1×
Provides Conservative transport of a regularised interface indicator. A reference for keeping a sharp indicator field conservative rather than merely bounded.
In the manuscript Interface treatment
Contextcited 1×
Provides Volume-based OCT detection of abnormal retinal reflectivity, nonperfusion and neovascularisation, using a distance-decaying Gaussian kernel for local perfusion effects.
In the manuscript Local perfusion modelling
Methodcited 1×
Provides Legendre spectral–Galerkin direct solvers for second- and fourth-order problems. The structured-matrix template for the radial basis and its assembly.
In the manuscript Radial basis construction
Contextcited 1×
Provides An annotated normal fundus photograph used as an anatomical reference; the original labels are retained in the manuscript figure.
In the manuscript Normal fundus figure
Contextcited 1×
Provides A normal adult fundus photograph (Jesse Vislisel, MD; Toni Venckus, CRA; CC BY-NC-ND 3.0) used as an anatomical reference without cropping or added annotation.
In the manuscript Normal fundus figure
Parametercited 1×
Provides Healthy human retinal oxygen extraction: mean blood flow 44.3 µL/min and oxygen extraction 2.33 µL/min. Their ratio defines the content-drop target 0.052596 mL(O₂)/mL(blood).
In the manuscript Oxygen calibration target
Methodcited 1×
Provides Maximum-principle-satisfying high-order finite-volume WENO schemes for convection–diffusion. Separates high-order accuracy from bound preservation.
In the manuscript Bound-preserving discretisation