Deep learning-enabled volumetric cone photoreceptor segmentation in adaptive optics optical coherence tomography images of normal and diseased eyes.

Objective quantification of photoreceptor cell morphology, such as cell diameter and outer segment length, is crucial for early, accurate, and sensitive diagnosis and prognosis of retinal neurodegenerative diseases. Adaptive optics optical coherence tomography (AO-OCT) provides three-dimensional (3-D) visualization of photoreceptor cells in the living human eye. The current gold standard for extracting cell morphology from AO-OCT images involves the tedious process of 2-D manual marking. To automate this process and extend to 3-D analysis of the volumetric data, we propose a comprehensive deep learning framework to segment individual cone cells in AO-OCT scans. Our automated method achieved human-level performance in assessing cone photoreceptors of healthy and diseased participants captured with three different AO-OCT systems representing two different types of point scanning OCT: spectral domain and swept source.

Evolution of adaptive optics retinal imaging [Invited].

This review describes the progress that has been achieved since adaptive optics (AO) was incorporated into the ophthalmoscope a quarter of a century ago, transforming our ability to image the retina at a cellular spatial scale inside the living eye. The review starts with a comprehensive tabulation of AO papers in the field and then describes the technological advances that have occurred, notably through combining AO with other imaging modalities including confocal, fluorescence, phase contrast, and optical coherence tomography. These advances have made possible many scientific discoveries from the first maps of the topography of the trichromatic cone mosaic to exquisitely sensitive measures of optical and structural changes in photoreceptors in response to light. The future evolution of this technology is poised to offer an increasing array of tools to measure and monitor in vivo retinal structure and function with improved resolution and control.

Cone photoreceptor dysfunction in retinitis pigmentosa revealed by optoretinography.

Retinitis pigmentosa (RP) is the most common group of inherited retinal degenerative diseases, whose most debilitating phase is cone photoreceptor death. Perimetric and electroretinographic methods are the gold standards for diagnosing and monitoring RP and assessing cone function. However, these methods lack the spatial resolution and sensitivity to assess disease progression at the level of individual photoreceptor cells, where the disease originates and whose degradation causes vision loss. High-resolution retinal imaging methods permit visualization of human cone cells in vivo but have only recently achieved sufficient sensitivity to observe their function as manifested in the cone optoretinogram. By imaging with phase-sensitive adaptive optics optical coherence tomography, we identify a biomarker in the cone optoretinogram that characterizes individual cone dysfunction by stimulating cone cells with flashes of light and measuring nanometer-scale changes in their outer segments. We find that cone optoretinographic responses decrease with increasing RP severity and that even in areas where cone density appears normal, cones can respond differently than those in controls. Unexpectedly, in the most severely diseased patches examined, we find isolated cones that respond normally. Short-wavelength-sensitive cones are found to be more vulnerable to RP than medium- and long-wavelength-sensitive cones. We find that decreases in cone response and cone outer-segment length arise earlier in RP than changes in cone density but that decreases in response and length are not necessarily correlated within single cones.

Revealing How Color Vision Phenotype and Genotype Manifest in Individual Cone Cells.

Purpose: Psychophysical and genetic testing provide substantial information about color vision phenotype and genotype. However, neither reveals how color vision phenotypes and genotypes manifest themselves in individual cones, where color vision and its anomalies are thought to originate. Here, we use adaptive-optics phase-sensitive optical coherence tomography (AO-PSOCT) to investigate these relationships. Methods: We used AO-PSOCT to measure cone function-optical response to light stimulation-in each of 16 human subjects with different phenotypes and genotypes of color vision (five color-normal, three deuteranopic, two protanopic, and six deuteranomalous trichromatic subjects). We classified three spectral types of cones (S, M, and L), and we measured cone structure-namely cone density, cone mosaic arrangement, and spatial arrangement of cone types. Results: For the different phenotypes, our cone function results show that (1) color normals possess S, M, and L cones; (2) deuteranopes are missing M cones but are normal otherwise; (3) protanopes are missing L cones but are normal otherwise; and (4) deuteranomalous trichromats are missing M cones but contain evidence of at least two subtypes of L cones. Cone function was consistent with the subjects' genotype in which only the first two M and L genes in the gene array are expressed and was correlated with the estimated spectral separation between photopigments, including in the deuteranomalous trichromats. The L/M cone ratio was highly variable in the color normals. No association was found between cone density and the genotypes and phenotypes investigated, and the cone mosaic arrangement was altered in the dichromats. Conclusions: AO-PSOCT is a novel method for assessing color vision phenotype and genotype in single cone cells.

Multi-reference global registration of individual A-lines in adaptive optics optical coherence tomography retinal images.

SIGNIFICANCE: Adaptive optics optical coherence tomography (AO-OCT) technology enables non-invasive, high-resolution three-dimensional (3D) imaging of the retina and promises earlier detection of ocular disease. However, AO-OCT data are corrupted by eye-movement artifacts that must be removed in post-processing, a process rendered time-consuming by the immense quantity of data. AIM: To efficiently remove eye-movement artifacts at the level of individual A-lines, including those present in any individual reference volume. APPROACH: We developed a registration method that cascades (1) a 3D B-scan registration algorithm with (2) a global A-line registration algorithm for correcting torsional eye movements and image scaling and generating global motion-free coordinates. The first algorithm corrects 3D translational eye movements to a single reference volume, accelerated using parallel computing. The second algorithm combines outputs of multiple runs of the first algorithm using different reference volumes followed by an affine transformation, permitting registration of all images to a global coordinate system at the level of individual A-lines. RESULTS: The 3D B-scan algorithm estimates and corrects 3D translational motions with high registration accuracy and robustness, even for volumes containing microsaccades. Averaging registered volumes improves our image quality metrics up to 22 dB. Implementation in CUDA™ on a graphics processing unit registers a 512 × 512 × 512 volume in only 10.6 s, 150 times faster than MATLAB™ on a central processing unit. The global A-line algorithm minimizes image distortion, improves regularity of the cone photoreceptor mosaic, and supports enhanced visualization of low-contrast retinal cellular features. Averaging registered volumes improves our image quality up to 9.4 dB. It also permits extending the imaging field of view (∼2.1 × ) and depth of focus (∼5.6 × ) beyond what is attainable with single-reference registration. CONCLUSIONS: We can efficiently correct eye motion in all 3D at the level of individual A-lines using a global coordinate system.

Weakly supervised individual ganglion cell segmentation from adaptive optics OCT images for glaucomatous damage assessment.

Cell-level quantitative features of retinal ganglion cells (GCs) are potentially important biomarkers for improved diagnosis and treatment monitoring of neurodegenerative diseases such as glaucoma, Parkinson's disease, and Alzheimer's disease. Yet, due to limited resolution, individual GCs cannot be visualized by commonly used ophthalmic imaging systems, including optical coherence tomography (OCT), and assessment is limited to gross layer thickness analysis. Adaptive optics OCT (AO-OCT) enables in vivo imaging of individual retinal GCs. We present an automated segmentation of GC layer (GCL) somas from AO-OCT volumes based on weakly supervised deep learning (named WeakGCSeg), which effectively utilizes weak annotations in the training process. Experimental results show that WeakGCSeg is on par with or superior to human experts and is superior to other state-of-the-art networks. The automated quantitative features of individual GCLs show an increase in structure-function correlation in glaucoma subjects compared to using thickness measures from OCT images. Our results suggest that by automatic quantification of GC morphology, WeakGCSeg can potentially alleviate a major bottleneck in using AO-OCT for vision research.

Adaptive optics for high-resolution imaging.

Adaptive optics (AO) is a technique that corrects for optical aberrations. It was originally proposed to correct for the blurring effect of atmospheric turbulence on images in ground-based telescopes and was instrumental in the work that resulted in the Nobel prize-winning discovery of a supermassive compact object at the centre of our galaxy. When AO is used to correct for the eye's imperfect optics, retinal changes at the cellular level can be detected, allowing us to study the operation of the visual system and to assess ocular health in the microscopic domain. By correcting for sample-induced blur in microscopy, AO has pushed the boundaries of imaging in thick tissue specimens, such as when observing neuronal processes in the brain. In this primer, we focus on the application of AO for high-resolution imaging in astronomy, vision science and microscopy. We begin with an overview of the general principles of AO and its main components, which include methods to measure the aberrations, devices for aberration correction, and how these components are linked in operation. We present results and applications from each field along with reproducibility considerations and limitations. Finally, we discuss future directions.

Cellular-Scale Imaging of Transparent Retinal Structures and Processes Using Adaptive Optics Optical Coherence Tomography.

High-resolution retinal imaging is revolutionizing how scientists and clinicians study the retina on the cellular scale. Its exquisite sensitivity enables time-lapse optical biopsies that capture minute changes in the structure and physiological processes of cells in the living eye. This information is increasingly used to detect disease onset and monitor disease progression during early stages, raising the possibility of personalized eye care. Powerful high-resolution imaging tools have been in development for more than two decades; one that has garnered considerable interest in recent years is optical coherence tomography enhanced with adaptive optics. State-of-the-art adaptive optics optical coherence tomography (AO-OCT) makes it possible to visualize even highly transparent cells and measure some of their internal processes at all depths within the retina, permitting reconstruction of a 3D view of the living microscopic retina. In this review, we report current AO-OCT performance and its success in visualizing and quantifying these once-invisible cells in human eyes.

Suite of methods for assessing inner retinal temporal dynamics across spatial and temporal scales in the living human eye.

Significance: There are no label-free imaging descriptors related to physiological activity of inner retinal cells in the living human eye. A major reason is that inner retinal neurons are highly transparent and reflect little light, making them extremely difficult to visualize and quantify. Aim: To measure physiologically-induced optical changes of inner retinal cells despite their challenging optical properties. Approach: We developed an imaging method based on adaptive optics and optical coherence tomography (AO-OCT) and a suite of postprocessing algorithms, most notably a new temporal correlation method. Results: We captured the temporal dynamics of entire inner retinal layers, of specific tissue types, and of individual cells across three different timescales from fast (seconds) to extremely slow (one year). Time correlation analysis revealed significant differences in time constant (up to 0.4 s) between the principal layers of the inner retina with the ganglion cell layer (GCL) being the most dynamic. At the cellular level, significant differences were found between individual GCL somas. The mean time constant of the GCL somas ( 0.69 ± 0.17 s ) was ∼ 30 % smaller than that of nerve fiber bundles and inner plexiform layer synapses and processes. Across longer durations, temporal speckle contrast and time-lapse imaging revealed motion of macrophage-like cells (over minutes) and GCL neuron loss and remodeling (over one year). Conclusions: Physiological activity of inner retinal cells is now measurable in the living human eye.

In vivo measurement of organelle motility in human retinal pigment epithelial cells.

Retinal pigment epithelial (RPE) cells are well known to play a central role in the progression of numerous retinal diseases. Changes in the structure and function of these cells thus may serve as sensitive biomarkers of disease onset. While in vivo studies have focused on structural changes, functional ones may better capture cell health owing to their more direct connection to cell physiology. In this study, we developed a method based on adaptive optics optical coherence tomography (AO-OCT) and speckle field dynamics for characterizing organelle motility in individual RPE cells. We quantified the dynamics in terms of an exponential decay time constant, the time required for the speckle field to decorrelate. Using seven normal subjects, we found the RPE speckle field to decorrelate in about 5 s. This result has two fundamental implications for future clinical use. First, it establishes a path for generating a normative baseline to which motility of diseased RPE cells can be compared. Second, it predicts an AO-OCT image acquisition time that is 36 times faster than used in our earlier report for individuating RPE cells, thus a major improvement in clinical efficacy.

Cone photoreceptor classification in the living human eye from photostimulation-induced phase dynamics.

Human color vision is achieved by mixing neural signals from cone photoreceptors sensitive to different wavelengths of light. The spatial arrangement and proportion of these spectral types in the retina set fundamental limits on color perception, and abnormal or missing types are responsible for color vision loss. Imaging provides the most direct and quantitative means to study these photoreceptor properties at the cellular scale in the living human retina, but remains challenging. Current methods rely on retinal densitometry to distinguish cone types, a prohibitively slow process. Here, we show that photostimulation-induced optical phase changes occur in cone cells and carry substantial information about spectral type, enabling cones to be differentiated with unprecedented accuracy and efficiency. Moreover, these phase dynamics arise from physiological activity occurring on dramatically different timescales (from milliseconds to seconds) inside the cone outer segment, thus exposing the phototransduction cascade and subsequent downstream effects. We captured these dynamics in cones of subjects with normal color vision and a deuteranope, and at different macular locations by: (i) marrying adaptive optics to phase-sensitive optical coherence tomography to avoid optical blurring of the eye, (ii) acquiring images at high speed that samples phase dynamics at up to 3 KHz, and (iii) localizing phase changes to the cone outer segment, where photoactivation occurs. Our method should have broad appeal for color vision applications in which the underlying neural processing of photoreceptors is sought and for investigations of retinal diseases that affect cone function.

Combined hardware and computational optical wavefront correction.

In many optical imaging applications, it is necessary to overcome aberrations to obtain high-resolution images. Aberration correction can be performed by either physically modifying the optical wavefront using hardware components, or by modifying the wavefront during image reconstruction using computational imaging. Here we address a longstanding issue in computational imaging: photons that are not collected cannot be corrected. This severely restricts the applications of computational wavefront correction. Additionally, performance limitations of hardware wavefront correction leave many aberrations uncorrected. We combine hardware and computational correction to address the shortcomings of each method. Coherent optical backscattering data is collected using high-speed optical coherence tomography, with aberrations corrected at the time of acquisition using a wavefront sensor and deformable mirror to maximize photon collection. Remaining aberrations are corrected by digitally modifying the coherently-measured wavefront during imaging reconstruction. This strategy obtains high-resolution images with improved signal-to-noise ratio of in vivo human photoreceptor cells with more complete correction of ocular aberrations, and increased flexibility to image at multiple retinal depths, field locations, and time points. While our approach is not restricted to retinal imaging, this application is one of the most challenging for computational imaging due to the large aberrations of the dilated pupil, time-varying aberrations, and unavoidable eye motion. In contrast with previous computational imaging work, we have imaged single photoreceptors and their waveguide modes in fully dilated eyes with a single acquisition. Combined hardware and computational wavefront correction improves the image sharpness of existing adaptive optics systems, and broadens the potential applications of computational imaging methods.

Measuring polarization changes in the human outer retina with polarization-sensitive optical coherence tomography.

Morphological changes in the outer retina such as drusen are established biomarkers to diagnose age-related macular degeneration. However, earlier diagnosis might be possible by taking advantage of more subtle changes that accompany tissues that bear polarization-altering properties. To test this hypothesis, we developed a method based on polarization-sensitive optical coherence tomography with which volumetric data sets of the macula were obtained from 10 young (<25 years) and 10 older (>54 years) subjects. All young subjects and 5 of the older subjects had retardance values induced by the retinal pigment epithelium and Bruch's membrane (RPE-BM) complex that were just above the noise floor measurement (5°-13° at 840 nm). In contrast, elevated retardance, up to 180°, was observed in the other 5 older subjects. Analysis of the degree of polarization uniformity (DOPU) demonstrates that reduced DOPU (<0.4) in the RPE is associated with elevated double pass phase retardation (DPPR) below the RPE-BM complex, suggesting that the observed elevated DPPR in older subjects is the result of increased scattering or polarization scrambling. Collectively, our measurements show that the outer retina can undergo dramatic change in its polarization properties with age, and in some cases still retain its clinically normal appearance.

Imaging and quantifying ganglion cells and other transparent neurons in the living human retina.

Ganglion cells (GCs) are fundamental to retinal neural circuitry, processing photoreceptor signals for transmission to the brain via their axons. However, much remains unknown about their role in vision and their vulnerability to disease leading to blindness. A major bottleneck has been our inability to observe GCs and their degeneration in the living human eye. Despite two decades of development of optical technologies to image cells in the living human retina, GCs remain elusive due to their high optical translucency. Failure of conventional imaging-using predominately singly scattered light-to reveal GCs has led to a focus on multiply-scattered, fluorescence, two-photon, and phase imaging techniques to enhance GC contrast. Here, we show that singly scattered light actually carries substantial information that reveals GC somas, axons, and other retinal neurons and permits their quantitative analysis. We perform morphometry on GC layer somas, including projection of GCs onto photoreceptors and identification of the primary GC subtypes, even beneath nerve fibers. We obtained singly scattered images by: (i) marrying adaptive optics to optical coherence tomography to avoid optical blurring of the eye; (ii) performing 3D subcellular image registration to avoid motion blur; and (iii) using organelle motility inside somas as an intrinsic contrast agent. Moreover, through-focus imaging offers the potential to spatially map individual GCs to underlying amacrine, bipolar, horizontal, photoreceptor, and retinal pigment epithelium cells, thus exposing the anatomical substrate for neural processing of visual information. This imaging modality is also a tool for improving clinical diagnosis and assessing treatment of retinal disease.

Adaptive optics optical coherence tomography angiography for morphometric analysis of choriocapillaris [Invited].

Histological studies have shown that morphometric changes at the microscopic level of choriocapillaris (CC) occur with aging and disease onset, and therefore may be sensitive biomarkers of outer retinal health. However, visualizing CC at this level in the living human eye is challenging because its microvascular is tightly interconnected and weakly reflecting. In this study, we address these challenges by developing and validating a method based on adaptive optics optical coherence tomography with angiography (AO-OCTA) that provides the necessary 3D resolution and image contrast to visualize and quantify these microscopic details. The complex network of anastomotic CC capillaries was successfully imaged in nine healthy subjects (26 to 68 years of age) and at seven retinal eccentricities across the macula. Using these images, four fundamental morphometric parameters of CC were characterized: retinal pigment epithelium-to-CC depth separation (17.5 ± 2.1 µm), capillary diameter (17.4 ± 2.3 µm), normalized capillary density (0.53 ± 0.08), and capillary length per unit area (50.4 ± 9.5 mm-1). AO-OCTA results were consistent with histologic studies and, unlike OCTA, showed clear delineation of CC capillaries, a requirement for measuring three of the four morphometric parameters. Success in younger and older eyes establishes a path for testing aging and disease effects in larger populations. To the best of our knowledge, this is the first quantitative morphometry of choriocapillaris at the level of individual capillaries in the living human retina.

Vision science and adaptive optics, the state of the field.

Adaptive optics is a relatively new field, yet it is spreading rapidly and allows new questions to be asked about how the visual system is organized. The editors of this feature issue have posed a series of question to scientists involved in using adaptive optics in vision science. The questions are focused on three main areas. In the first we investigate the use of adaptive optics for psychophysical measurements of visual system function and for improving the optics of the eye. In the second, we look at the applications and impact of adaptive optics on retinal imaging and its promise for basic and applied research. In the third, we explore how adaptive optics is being used to improve our understanding of the neurophysiology of the visual system.

Photoreceptor disc shedding in the living human eye.

Cone photoreceptors undergo a daily cycle of renewal and shedding of membranous discs in their outer segments (OS), the portion responsible for light capture. These physiological processes are fundamental to maintaining photoreceptor health, and their dysfunction is associated with numerous retinal diseases. While both processes have been extensively studied in animal models and postmortem eyes, little is known about them in the living eye, in particular human. In this study, we report discovery of the optical signature associated with disc shedding using a method based on adaptive optics optical coherence tomography (AO-OCT) in conjunction with post-processing methods to track and monitor individual cone cells in 4D. The optical signature of disc shedding is characterized by an abrupt transient loss in the cone outer segment tip (COST) reflection followed by its return that is axially displaced anteriorly. Using this signature, we measured the temporal and spatial properties of shedding events in three normal subjects. Average duration of the shedding event was 8.8 ± 13.4 minutes, and average length loss of the OS was 2.1 µm (7.0% of OS length). Prevalence of cone shedding was highest in the morning (14.3%) followed by the afternoon (5.7%) and evening (4.0%), with load distributed across the imaged patch. To the best of our knowledge these are the first images of photoreceptor disc shedding in the living retina.

3D Imaging of Retinal Pigment Epithelial Cells in the Living Human Retina.

PURPOSE: Dysfunction of the retinal pigment epithelium (RPE) underlies numerous retinal pathologies, but biomarkers sensitive to RPE change at the cellular level are limited. In this study, we used adaptive optics optical coherence tomography (AO-OCT) in conjunction with organelle motility as a novel contrast mechanism to visualize RPE cells and characterize their 3-dimensional (3D) reflectance profile. METHODS: Using the Indiana AO-OCT imaging system (λc = 790 nm), volumes were acquired in the macula of six normal subjects (25-61 years). Volumes were registered in 3D with subcellular accuracy, layers segmented, and RPE and photoreceptor en face images extracted and averaged. Voronoi and two-dimensional (2D) power spectra analyses were applied to the images to quantify RPE and cone packing and cone-to-RPE ratio. RESULTS: Adaptive optics OCT revealed two distinct reflectance patterns at the depth of the RPE. One is characterized by the RPE interface with rod photoreceptor tips, the second by the RPE cell nuclei and surrounding organelles, likely melanin. Increasing cell contrast by averaging proved critical for observing the RPE cell mosaic, successful in all subjects and retinal eccentricities imaged. Retinal pigment epithelium mosaic packing and cell thickness generally agreed with that of histology and in vivo studies using other imaging modalities. CONCLUSIONS: We have presented, to our knowledge, the first detailed characterization of the 3D reflectance profile of individual RPE cells and their relation to cones and rods in the living human retina. Success in younger and older eyes establishes a path for testing aging effects in larger populations. Because the technology is based on OCT, our measurements will aid in interpreting clinical OCT images.

A Review of Adaptive Optics Optical Coherence Tomography: Technical Advances, Scientific Applications, and the Future.

PURPOSE: Optical coherence tomography (OCT) has enabled "virtual biopsy" of the living human retina, revolutionizing both basic retina research and clinical practice over the past 25 years. For most of those years, in parallel, adaptive optics (AO) has been used to improve the transverse resolution of ophthalmoscopes to foster in vivo study of the retina at the microscopic level. Here, we review work done over the last 15 years to combine the microscopic transverse resolution of AO with the microscopic axial resolution of OCT, building AO-OCT systems with the highest three-dimensional resolution of any existing retinal imaging modality. METHODS: We surveyed the literature to identify the most influential antecedent work, important milestones in the development of AO-OCT technology, its applications that have yielded new knowledge, research areas into which it may productively expand, and nascent applications that have the potential to grow. RESULTS: Initial efforts focused on demonstrating three-dimensional resolution. Since then, many improvements have been made in resolution and speed, as well as other enhancements of acquisition and postprocessing techniques. Progress on these fronts has produced numerous discoveries about the anatomy, function, and optical properties of the retina. CONCLUSIONS: Adaptive optics OCT continues to evolve technically and to contribute to our basic and clinical knowledge of the retina. Due to its capacity to reveal cellular and microscopic detail invisible to clinical OCT systems, it is an ideal companion to those instruments and has the demonstrable potential to produce images that can guide the interpretation of clinical findings.

Modal content of living human cone photoreceptors.

Decades of experimental and theoretical investigations have established that photoreceptors capture light based on the principles of optical waveguiding. Yet considerable uncertainty remains, even for the most basic prediction as to whether photoreceptors support more than a single waveguide mode. To test for modal behavior in human cone photoreceptors in the near infrared, we took advantage of adaptive-optics optical coherence tomography (AO-OCT, λc = 785 nm) to noninvasively image in three dimensions the reflectance profile of cones. Modal content of reflections generated at the cone inner segment and outer segment junction (IS/OS) and cone outer segment tip (COST) was examined over a range of cone diameters in 1,802 cones from 0.6° to 10° retinal eccentricity. Second moment analysis in conjunction with theoretical predictions indicate cone IS and OS have optical properties consistent of waveguides, which depend on segment diameter and refractive index. Cone IS was found to support a single mode near the fovea (≤3°) and multiple modes further away (>4°). In contrast, no evidence of multiple modes was found in the cone OSs. The IS/OS and COST reflections share a common optical aperture, are most circular near the fovea, show no orientation preference, and are temporally stable. We tested mode predictions of a conventional step-index fiber model and found that in order to fit our AO-OCT results required a lower estimate of the IS refractive index and introduction of an IS focusing/tapering effect.