Imaging the retinal vasculature offers a surrogate view of systemic vascular health, allowing noninvasive and longitudinal assessment of vascular pathology. The earliest anomalies in vascular disease arise in the microvasculature, however current imaging methods lack the spatiotemporal resolution to track blood flow at the capillary level. We report here on novel imaging technology that allows direct, noninvasive optical imaging of erythrocyte flow in human retinal capillaries. This was made possible using adaptive optics for high spatial resolution (1.5 μm), sCMOS camera technology for high temporal resolution (460 fps), and tunable wavebands from a broadband laser for maximal erythrocyte contrast. Particle image velocimetry on our data sequences was used to quantify flow. We observed marked spatiotemporal variability in velocity, which ranged from 0.3 to 3.3 mm/s, and changed by up to a factor of 4 in a given capillary during the 130 ms imaging period. Both mean and standard deviation across the imaged capillary network varied markedly with time, yet their ratio remained a relatively constant parameter (0.50 ± 0.056). Our observations concur with previous work using less direct methods, validating this as an investigative tool for the study of microvascular disease in humans.

No abstract

Conventional adaptive optics enables correction of high-order aberrations of the eye, but only for a single retinal point. When imaging extended regions of the retina, aberrations increase away from this point and degrade image quality. The zone over which aberrations do not change significantly is called the "isoplanatic patch." Literature concerning the human isoplanatic patch is incomplete. We determine foveal isoplanatic patch characteristics by performing Hartmann-Shack aberrometry in 1 deg increments in 8 directions on 7 human eyes. Using these measurements, we establish the correction quality required to yield at least 80% of the potential patch size for a given eye. Single-point correction systems (conventional adaptive optics) and multiple-point correction systems (multiconjugate adaptive optics) are simulated. Results are compared to a model eye. Using the Marechal criterion for 555-nm light, average isoplanatic patch diameter for our subjects is 0.80+/-0.10 deg. The required order of aberration correction depends on desired image quality over the patch. For the more realistically achievable criterion of 0.1 mum root mean square (rms) wavefront error over a 6.0-mm pupil, correction to at least sixth order is recommended for all adaptive optics systems. The most important aberrations to target for a multiconjugate correction are defocus, astigmatism, and coma.

Neighboring cones in the fovea display high variability in their optical properties.