Photoreceptor Mosaic Research Articles

Adaptive optics scanning laser ophthalmoscopy in a heterogenous cohort with Stargardt disease

Image based cell-specific biomarkers will play an important role in monitoring treatment outcomes of novel therapies in patients with Stargardt (STGD1) disease and may provide information on the exact mechanism of retinal degeneration. This study reports retinal image features from conventional clinical imaging and from corresponding high-resolution imaging with a confocal adaptive optics scanning laser ophthalmoscope (AOSLO) in a heterogenous cohort of patients with Stargardt (STGD1) disease. This is a prospective observational study in which 16 participants with clinically and molecularly confirmed STGD1, and 7 healthy controls underwent clinical assessment and confocal AOSLO imaging. Clinical assessment included short-wavelength and near-infrared fundus autofluorescence, spectral-domain optical coherence tomography, and macular microperimetry. AOSLO images were acquired over a range of retinal eccentricities (0°–20°) and mapped to areas of interest from the clinical images. A regular photoreceptor mosaic was identified in areas of normal or near normal retinal structure on clinical images. Where clinical imaging indicated areas of retinal degeneration, the photoreceptor mosaic was disorganised and lacked unambiguous cones. Discrete hyper-reflective foci were identified in 9 participants with STGD1 within areas of retinal degeneration. A continuous RPE cell mosaic at the fovea was identified in one participant with an optical gap phenotype. The clinical heterogeneity observed in STGD1 is reflected in the findings on confocal AOSLO imaging.

Morphology of the normative human cone photoreceptor mosaic and a publicly available adaptive optics montage repository.

Adaptive optics ophthalmoscopy has enabled visualization of the in vivo human photoreceptor mosaic in health, disease and its treatment. Despite this, the clinical utility of the imaging technology has been limited by a lack of automated analysis techniques capable of accurately quantifying photoreceptor structure and a lack of an available normative image database. Here, we present a fully automated algorithm for estimating cone spacing and density over a complete adaptive optics montage along with a database of normative images and cone densities. We imaged the cone mosaics surrounding the fovea and along the horizontal and vertical meridians of fifty normal-sighted controls with a custom-built, multimodal adaptive optics scanning light ophthalmoscope. Cone spacing was automatically measured in the frequency domain and spacing measurements were converted to estimates of cone density at all locations across the montage. Consistent with previous reports, cone density measurements were highest near fovea (152,906 ± 53,209 cones/mm2) and decreased exponentially with eccentricity. A 2.5-fold variation was found in cone density estimates at 0.1mm, this variation decreased to 1.75-fold at 1mm. We provide all images, mosaic quantifications, and automated software open source. This database will aid investigators in translating adaptive optics ophthalmoscopy to clinical applications.

Comparison of the Photoreceptor Mosaic Before and After Macular Hole Surgery with High Resolution Adaptive Optics Imaging

PurposeTo assess the photoreceptor mosaic in patients with idiopathic full thickness macular hole (MH) before and after pars plana vitrectomy (PPV) with adaptive optics enhanced retinal imaging (AO). DesignProspective case series. MethodsProspective cohort study of patients who presented at the Kensington Eye Institute, Toronto, Canada with a diagnosis of macular hole treated with PPV. Exclusion criteria: secondary MH, high myopia (axial length >26.5 mm), media opacity precluding OCT or AO imaging, previous intraocular surgery except for cataract extraction. Imaging using an AO fundus camera (Imagine Eyes, RTX1) was performed preoperatively and 3-months following successful MH repair in both eyes. Cone density (CD), regularity, dispersion, regularity, dispersion, and spacing were measured at 2° and 4° of eccentricity in 4 quadrants (superior, inferior, nasal, and temporal) with pre- and postoperative values compared. ResultsWe included 18 eyes of 9 patients. At 2° there was significant reduction in CD and increase in spacing and dispersion and a non-significant change in regularity post-operatively. Comparison between preoperative and postoperative measurements at 2° mean (Standard Error= SE) were: CD: 14612 ± 3003 and 12280 ± 4632 photoreceptors/mm2 [95%CIs= -2413 to -702] p=0.0004, regularity: 88 ± 7% and 84 ± 12% [95%CIs= -4.67 to 0.04] p=0.054, dispersion: 19 ± 6% and 23 ± 10% [95%CIs= 0.5 to 4.24] p=0.013, spacing: 9 ± 1 microns and 10 ± 2microns [95%CIs= 0.40 to 1.27] p=0.0002; at 4° was: CD: 13377 ± 4339 and 12770 ± 4391 photoreceptors/mm2 [95%CIs= -1368 to 252] p=0.176, regularity:87 ± 9% and 86 ± 12% [95%CIs= -4.65 to 0.08] p=0.74, dispersion: 20 ± 8% and 20 ±9 % [95%CIs= -2.11 to 1.5] p=0.74, spacing:10 ± 2 microns and 10 ± 3 microns [95%CIs= -0.23 to 0.58] p=0.39. ConclusionAdaptive optics imaging allows quantitative assessment of the photoreceptor mosaic pre- and post-PPV in patients with MH. There was a significant change to the photoreceptor mosaic related to the MH at 2° pre- and post-operatively. AO imaging enables high-resolution investigation of the photoreceptor remodeling process following surgery, which may allow for a more thorough assessment of surgical outcomes.

Adaptive optics retinal imaging in patients with usher syndrome.

To determine the structure of the cone photoreceptor mosaic in the macula in eyes with retinitis pigmentosa related to Usher syndrome using adaptive optics fundus (AO) imaging and to correlate these findings with those of the standard clinical diagnostics. Ten patients with a genetically confirmed retinitis pigmentosa in Usher syndrome due to biallelic variants in MYO7A or USH2A were enrolled in the study. All patients underwent a complete ophthalmological examination including best corrected visual acuity (BCVA), spectral-domain optical coherence tomography (SD-OCT) with fundus autofluorescence photography (FAF), full-field (ffERG) and multifocal electroretinography (mfERG) and Adaptive Optics Flood Illuminated Ophthalmoscopy (AO, rtx1™, Imagine Eyes, Orsay, France). The cone density was assessed centrally and at each 0.5 degree horizontally and vertically from 1-4 degree of eccentricity. In the AO images, photoreceptor cell death was visualized as a disruption of the cone mosaic and low cone density. In the early stage of the disease, cones were still visible in the fovea, whereas outside the fovea a loss of cones was recognizable by blurry, dark patches. The blurry patches corresponded to the parafoveal hypofluorescent ring in the FAF images and the beginning loss of the IS/OS line and external limiting membrane in the SD-OCT images. FfERGs were non-recordable in 7 patients and reduced in 3. The mfERG was reduced in all patients and correlated significantly (p <0.001) with the cone density. The kinetic visual field area, measured with III4e and I4e, did not correlate with the cone density. The structure of the photoreceptors in Usher syndrome patients were detectable by AO fundus imaging. The approach of using high-resolution technique to assess the photoreceptor structure complements the established clinical examinations and allows a more sensitive monitoring of early stages of retinitis pigmentosa in Usher syndrome.

Systematic analysis of ocular features and responses of cultured spotted wolffish (Anarhichas minor).

A better understanding of unique anatomical and functional features of the visual systems of teleost fish could provide key knowledge on how these systems influence the health and survival of these animals in both wild and culture environments. We took a systematic approach to assess some of the visual systems of spotted wolffish (Anarhichas minor), a species of increasing importance in North Atlantic aquaculture initiatives. The lumpfish (Cyclopterus lumpus) was included in these studies in a comparative manner to provide reference. Histology, light and electron microscopy were used to study the spatial distribution and occurrence of cone photoreceptor cells and the nature of the retinal tissues, while immunohistochemistry was used to explore the expression patterns of two photoreceptor markers, XAP-1 and XAP-2, in both species. A marine bacterial infection paradigm in lumpfish was used to assess how host-pathogen responses might impact the expression of these photoreceptor markers in these animals. We define a basic photoreceptor mosaic and present an ultrastructural to macroscopic geographical configuration of the retinal pigment tissues in both animals. Photoreceptor markers XAP-1 and XAP-2 have novel distribution patterns in spotted wolffish and lumpfish retinas, and exogenous pathogenic influences can affect the normal expression pattern of XAP-1 in lumpfish. Live tank-side ophthalmoscopy and spectral domain optical coherence tomography (SD-OCT) revealed that normal cultured spotted wolffish display novel variations in the shape of the retinal tissue. These two complementary imaging findings suggest that spotted wolffish harbour unique ocular features not yet described in marine teleosts and that visual function might involve specific retinal tissue shape dynamics in these animals. Finally, extensive endogenous biofluorescence is present in the retinal tissues of both animals, which raises questions about how these animals might use retinal tissue in novel ways for visual perception and/or communication. This work advances fundamental knowledge on the visual systems of two economically important but now threatened North Atlantic teleosts and provides a basic foundation for further research on the visual systems of these animals in health versus disease settings. This work could also be useful for understanding and optimizing the health and welfare of lumpfish and spotted wolffish in aquaculture towards a one health or integrative perspective.

Retinal Structure of Poecilia sphenops: Photoreceptor Mosaics, Synaptic Ribbon Patterns, and Glial Cell Expressions.

The specific arrangement and distribution of photoreceptors in the retina can vary among different fish species, with each species exhibiting adaptations related to its habitat, behavior, and visual requirements. Poecilia sphenops, a diurnal fish, was the focus of this study. The retinas of a total of eighteen Molly fish were investigated utilizing light and electron microscopy. The retina exhibited a square mosaic pattern of the inner segments of cones. This pattern comprised double cones positioned along the sides of a square, with two types of single cones situated at the center and corners of the square arrangement across the entire retina. The corner cones were slightly shorter than the central ones. Additionally, the outer plexiform layer contained both cone pedicles and rod spherules. The rod spherule consisted of a single synaptic ribbon arranged in a triad or quadrat junctional arrangement within the invaginating free ends of the horizontal and bipolar cell processes. On the other hand, cone pedicles have more than one synaptic ribbon in their junctional complex. The inner nuclear layer consisted of the amacrine, bipolar, Müller, and horizontal cell bodies. Müller cell processes, expressing GFAP, extended across all retinal layers, segmenting the deeper retina into alternating fascicles of optic axons and ganglion cells. The outer and inner plexiform layers showed many astrocyte cell processes expressing GFAP. In conclusion, the current study is the first record of the retinal structures of Molly fish. This study illustrated the mosaic arrangement of photoreceptors and GFAP expression patterns of astrocytes and Müller cells. The presence of three cone types, coupled with a sufficient number of rods, likely facilitates motion awareness for tasks like finding food and performing elaborate mating ceremonies.

Minimum intensity projection of embossed quadrant-detection images for improved photoreceptor mosaic visualisation.

Non-confocal split-detection imaging reveals the cone photoreceptor inner segment mosaic in a plethora of retinal conditions, with the potential of providing insight to ageing, disease, and response to treatment processes, in vivo, and allows the screening of candidates for cell rescue therapies. This imaging modality complements confocal reflectance adaptive optics scanning light ophthalmoscopy, which relies on the waveguiding properties of cones, as well as their orientation toward the pupil. Split-detection contrast, however, is directional, with each cone inner segment appearing as opposite dark and bright semicircles, presenting a challenge for either manual or automated cell identification. Quadrant-detection imaging, an evolution of split detection, could be used to generate images without directional dependence. Here, we demonstrate how the embossed-filtered quadrant-detection images, originally proposed by Migacz etal. for visualising hyalocytes, can also be used to generate photoreceptor mosaic images with better and non-directional contrast for improved visualisation. As a surrogate of visualisation improvement between legacy split-detection images and the images resulting from the method described herein, we provide preliminary results of simple image processing routines that may enable the automated identification of generic image features, as opposed to complex algorithms developed specifically for photoreceptor identification, in pathological retinas.

Poster Session II: Understanding scattering in high-resolution retinal imaging.

In conventional fundus images the contrast relates to scattering and absorption of light. Here, we report on modeling of high-resolution retinal imaging to understand the role of mitochondria light scattering in the ellipsoid of photoreceptors. We evaluate the directionality appearance of the photoreceptors, and its relation to drusen appearance in age-related macular degeneration (AMD). The refractive index contrast of the mitochondria collectively resembles that of micron-sized reflectors. Obliqueness of the scattering results in vignetting by the pupil that impacts on the brightness of the photoreceptors. Our hypothesis is that in early-stage AMD, drusen will perturb the photoreceptor alignment and thus cause oblique scattering and reduced brightness on the slope of the drusen in high-resolution retinal images. We model how this affects the observable photoreceptor mosaic both in the healthy eye and in eyes with AMD. The results confirm that the annular dark ring around drusen is likely caused by oblique scattering. We use a model eye to test the hypothesis with a reflective mask emulating the photoreceptor mosaic An SLM is used for amplitude modulation in the pupil and differential imaging allows for capturing off-axis images with an appearance like that of phase contrast images. Further refinement of the model eye may be a valuable way to test new imaging modalities prior to their implementation in the real eye.

Contributed Session III: The limits of resolution in the S-cone pathway.

The resolution of the S-cone pathway is first constrained by the density and arrangement of S-cones in the photoreceptor mosaic. Prior work comparing S-cone isolating visual acuity (sVA) to histological estimates of S-cone density has led to mixed conclusions, likely due to inter-individual differences in the S-cone sub-mosaic. We examined sVA in subjects with spectrally classified cone mosaics to test how the grain of the S-cone sub-mosaic limits resolution. Three observers whose cones were previously classified via adaptive optics (AO)-OCT based optoretinography participated in an sVA task at eccentricities ranging from 1.3° to 12.9° spread along all four cardinal meridians. Observers adapted to yellow light [CIE (0.45, 0.51); 200 cd/m2] for two minutes.Then, acuity was measured using a Tumbling 'E' task that showed blue [CIE (0.16, 0.044); 0.66 cd/m2) letters on the same yellow background, with an S-cone contrast of 0.93 (L/M-cone contrasts <0.01). Simultaneously recorded high-resolution AOSLO videos helped guide stimulus delivery to the spectrally classified retinal area. Measured sVAs were worse than predicted by the calculated S-cone Nyquist limit at all eccentricities, suggesting pooling of information from S-cones. Moreover, this pooling increased with eccentricity. While sVA does not follow the S-cone Nyquist limit, it is in good concordance with the Nyquist limit corresponding to known estimates of the sampling density of small bistratified retinal ganglion cells.

Cone Photoreceptor Morphology in Choroideremia Assessed Using Non-Confocal Split-Detection Adaptive Optics Scanning Light Ophthalmoscopy.

Choroideremia (CHM) is an X-linked inherited retinal degeneration causing loss of the photoreceptors, retinal pigment epithelium, and choriocapillaris, although patients typically retain a central island of relatively preserved, functioning retina until late-stage disease. Here, we investigate cone photoreceptor morphology within the retained retinal island by examining cone inner segment area, density, circularity, and intercone space. Using a custom-built, multimodal adaptive optics scanning light ophthalmoscope, nonconfocal split-detection images of the photoreceptor mosaic were collected at 1°, 2°, and 4° temporal to the fovea from 13 CHM and 12 control subjects. Cone centers were manually identified, and cone borders were segmented. A custom MATLAB script was used to extract area and circularity for each cone and calculate the percentage of intercone space in each region of interest. Bound cone density was also calculated. An unbalanced two-way ANOVA and Bonferroni post hoc tests were used to assess statistical differences between the CHM and control groups and along retinal eccentricity. Cone density was lower in the CHM group than in the control group (P < 0.001) and decreased with eccentricity from the fovea (P < 0.001). CHM cone inner segments were larger in area (P < 0.001) and more circular (P = 0.042) than those of the controls. Intercone space in CHM was also higher than in the controls (P < 0.001). Cone morphology is altered in CHM compared to control, even within the centrally retained, functioning retinal area. Further studies are required to determine whether such morphology is a precursor to cone degeneration.

Multimodal high-resolution retinal imaging using a camera-based DMD-integrated adaptive optics flood-illumination ophthalmoscope.

We demonstrate the feasibility of a multimodal adaptive optics flood-illumination ophthalmoscope, able to provide both bright-field and dark-field images (such as phase contrast). The multimodality was made possible by integrating a digital micromirror device (DMD) at the illumination path to project a sequence of complementary high-resolution patterns into the retina. Through a versatile post-processing method that digitally selects backscattered or multiply scattered photons, we were able: (1) to achieve up to four-fold contrast increase of bright-field images when imaging the photoreceptor mosaic and nerve fibers; and (2) to visualize translucent retinal features such as capillaries, red blood cells, vessel walls, ganglion cells, and photoreceptor inner segments through phase contrast.

Non-Rigid Registration for High-Resolution Retinal Imaging.

Adaptive optics provides improved resolution in ophthalmic imaging when retinal microstructures need to be identified, counted, and mapped. In general, multiple images are averaged to improve the signal-to-noise ratio or analyzed for temporal dynamics. Image registration by cross-correlation is straightforward for small patches; however, larger images require more sophisticated registration techniques. Strip-based registration has been used successfully for photoreceptor mosaic alignment in small patches; however, if the deformations along strips are not simple displacements, averaging can degrade the final image. We have applied a non-rigid registration technique that improves the quality of processed images for mapping cones over large image patches. In this approach, correction of local deformations compensates for local image stretching, compressing, bending, and twisting due to a number of causes. The main result of this procedure is improved definition of retinal microstructures that can be better identified and segmented. Derived metrics such as cone density, wall-to-lumen ratio, and quantification of structural modification of blood vessel walls have diagnostic value in many retinal diseases, including diabetic retinopathy and age-related macular degeneration, and their improved evaluations may facilitate early diagnostics of retinal diseases.

Universality of Form: The Case of Retinal Cone Photoreceptor Mosaics.

Cone photoreceptor cells are wavelength-sensitive neurons in the retinas of vertebrate eyes and are responsible for color vision. The spatial distribution of these nerve cells is commonly referred to as the cone photoreceptor mosaic. By applying the principle of maximum entropy, we demonstrate the universality of retinal cone mosaics in vertebrate eyes by examining various species, namely, rodent, dog, monkey, human, fish, and bird. We introduce a parameter called retinal temperature, which is conserved across the retinas of vertebrates. The virial equation of state for two-dimensional cellular networks, known as Lemaître's law, is also obtained as a special case of our formalism. We investigate the behavior of several artificially generated networks and the natural one of the retina concerning this universal, topological law.

Photoreceptor mosaic before and after macular hole surgery with RTX-1 adaptive optics retinal camera

Photoreceptor mosaic before and after macular hole surgery with RTX-1 adaptive optics retinal camera

Extracting spacing-derived estimates of rod density in healthy retinae.

Quantification of the rod photoreceptor mosaic using adaptive optics scanning light ophthalmoscopy (AOSLO) remains challenging. Here we demonstrate a method for deriving estimates of rod density and rod:cone ratio based on measures of rod spacing, cone numerosity, and cone inner segment area. Twenty-two AOSLO images with complete rod visualization were used to validate this spacing-derived method for estimating density. The method was then used to estimate rod metrics in an additional 105 images without complete rod visualization. The spacing-derived rod mosaic metrics were comparable to published data from histology. This method could be leveraged to develop large normative databases of rod mosaic metrics, though limitations persist with intergrader variability in assessing cone area and numerosity.

Adaptive optics scanning laser ophthalmoscopy characterization of cone photoreceptor reflectivity perturbation by drusen and subretinal drusenoid deposits

AbstractPurpose: To examine the structure of photoreceptors around drusen and subretinal drusenoid deposits (SDD) in patients with age‐related macular degeneration (AMD) using multimodal imaging, including adaptive optics scanning laser ophthalmoscopy (AOSLO).Methods: Thirty‐four (n = 34) patients with early to intermediate AMD were enrolled and classified into 3 groups based on predominant lesion type. Group 1 included 20 eyes of 15 subjects with medium‐large soft drusen (63–1000 μm diameter), group 2 had 15 eyes of 10 subjects with cuticular drusen (&lt;63 μm), and group 3 had 14 eyes of 9 subjects with SDD. Lesion presence was ascertained by colour fundus photography, near‐infrared reflectance, blue reflectance, autofluorescence, and spectral‐domain optical coherence tomography (SD‐OCT). SDD were classified with a 3‐stage OCT‐based grading system. In addition, the reflectivity and packing mosaic of cone photoreceptors were assessed with AOSLO.Results: AOSLO revealed characteristic photoreceptor reflectivity perturbation by different types of AMD's extracellular lesions. Around large soft drusen, the photoreceptor mosaic was invisible at the edge and visible on the top. The photoreceptors in the retinae with cuticular drusen were largely maintained as a contiguous mosaic but with reduced reflectivity over the lesions. Photoreceptors surrounding SDD showed reduced reflectivity associated with the SDD stage. The en face appearance of the photoreceptor mosaic was consistent with its cross‐sectional structure as rendered by SD‐OCT.Conclusions: The dramatic photoreceptor reflectivity perturbation associated with SDD indicates that SDD may impose a more direct and severe impact on overlying photoreceptors than drusen, suggesting that the retinal function in eyes with SDD may be more severely impaired.@@@@

Role of Müller cells in cone mosaic rearrangement and retinal remodelling in photoreceptor degenerations

AbstractPhotoreceptor degenerative diseases (e.g., retinitis pigmentosa or age‐related macular degeneration) are currently one of the leading causes of irreversible blindness in the world. Although their pathophysiological mechanisms may vary, they all share some features such as initial photoreceptor loss associated with glial activation and subsequent retinal remodelling. It is well known that glial activation is an early feature of photoreceptor degenerations that is involved in retinal remodelling, which, in turn, is thought to be the cause of the secondary death of the retinal ganglion cells. The mammalian retina contains three types of glial cells. Two types of macroglial cells: astrocytes and Müller cells, and microglial cells. Of them, Müller cells appear to be involved in the reorganization of the photoreceptor mosaic in empty rings of photoreceptor degeneration in the early stages of photoreceptor loss. Müller cells are the principal glial cells in the retina with a wide range of physiological and retinal structural support functions. The basic morphological scheme of Müller cells consists of a soma located in the inner nuclear layer from which an inner and an outer trunk emerge. The inner trunk contributes to the formation of the inner limiting membrane, and the outer trunk divides into complex rootlets spanning photoreceptor nuclei and finally forming the outer limiting membrane of the retina. In the early stages of photoreceptor degeneration, macroglial cell reactivity, namely increased expression of glial fibrillary acidic protein (GFAP) and Müller cells hypertrophy, leads to the formation of a glial scar or glial seal by hypertrophic Müller cell processes in the outer retina that may be related to the formation of empty rings of photoreceptor degeneration. Interestingly, Müller cells extend their external processes horizontally and radially within the rings of photoreceptor degeneration, filling the empty space left by dead photoreceptors. These rings become larger as photoreceptor loss increases, which suggests that they are the result of photoreceptor loss. As rings increase in size and, eventually, merge, the increased expression in the processes of Müller cells continues in the periphery of the rings. Perhaps in an attempt to complete the glial seal to isolate the remnant of the neural retina from the retinal pigment epithelium and choroid which, in turn, may represent a physical barrier to axonal regrowth and migration and therefore may impede the success of therapies aimed at replacing or substituting photoreceptors. Indeed, as photoreceptor loss progresses, the glial sea is more evident, covering large areas of the retina. Although the exact function of the glial seal is not known at this stage, it is thought that it may have a mechanical function in conferring some stability to the retina during the course of degeneration, but also a protective function, since in the later stages of retinal remodelling, focal gaps in the glial seal may allow the migration of retinal pigment epithelium cells to the inner retina, triggering a series of progressive events that eventually cause retinal ganglion cell death.

Retinal imaging using adaptive optics optical coherence tomography with fast and accurate real-time tracking.

One of the main obstacles in high-resolution 3-D retinal imaging is eye motion, which causes blur and distortion artifacts that require extensive post-processing to be corrected. Here, an adaptive optics optical coherence tomography (AOOCT) system with real-time active eye motion correction is presented. Correction of ocular aberrations and of retinal motion is provided by an adaptive optics scanning laser ophthalmoscope (AOSLO) that is optically and electronically combined with the AOOCT system. We describe the system design and quantify its performance. The AOOCT system features an independent focus adjustment that allows focusing on different retinal layers while maintaining the AOSLO focus on the photoreceptor mosaic for high fidelity active motion correction. The use of a high-quality reference frame for eye tracking increases revisitation accuracy between successive imaging sessions, allowing to collect several volumes from the same area. This system enables spatially targeted retinal imaging as well as volume averaging over multiple imaging sessions with minimal correction of motion in post processing.

Assessment of photoreceptor function with ultrafast retinal densitometry.

The optical density of visual pigment can be measured by imaging the dark-adapted eye while bleaching with visible light. This measurement can be made for individual photoreceptor cells using adaptive optics; however, activation of the phototransduction cascade imparts rapid changes in phase that modulate the signal via optical interference. This limits utility because data must be averaged over many experimental runs. Here we used a "flood" illuminated adaptive optics system at 4000 fps, bright light to achieve rapid bleaching, and broad illumination bandwidth to mitigate interference effects. Data were super-resolved using the natural motion of the eye to overcome the reduced pixel resolution of the ultrafast camera. This approach was applied to classify the trichromatic cone photoreceptor mosaic at a single fixation locus within the foveal region of 3 healthy subjects. Subjects were dark adapted for 6 minutes to replenish cone photopigment. This was followed either directly by imaging at 555 ± 50 nm, or by first pre-adapting the retina to 700 nm light to preferentially deplete "L" cone pigment. A total of 3,252 cones were classified as either "S", "M", or "L" type based on clustering of the intensity data observed under these two conditions. Mean classification probability ranged from 99.3 to 99.8%, with individual cell probabilities exceeding 95% in 97.0 to 99.2% of cones. Accuracy of cone classification peaked when using the first 10-30 ms of data, with significant reductions in accuracy noted with the inclusion of data from later times. Our results show that rapid bleaching and data acquisition significantly improve the robustness of cell-resolved densitometry.

Automatic quantification of cone photoreceptors in adaptive optics scanning light ophthalmoscope images using multi-task learning.

Adaptive optics scanning light ophthalmoscope (AO-SLO) can directly image the cone photoreceptor mosaic in the living human retina, which offers a potentially great tool to detect cone-related ocular pathologies by quantifying the changes in the cone mosaic. However, manual quantification is very time-consuming and automation is highly desirable. In this paper, we developed a fully automatic method based on multi-task learning to identify and quantify cone photoreceptors. By including cone edges in the labels as the third dimension of the classification, our method provided more accurate and reliable results than the two previously reported methods. We trained and validated our network in an open data set consisting of over 200,000 cones, and achieved a 99.20% true positive rate, 0.71% false positive rate, and 99.24% Dice's coefficient on the test set consisting of 44,634 cones. All are better than the reported methods. In addition, the reproducibility of all three methods was also tested and compared, and the result showed the performance of our method was generally closer to the gold standard. Bland-Altman plots show that our method was more stable and accurate than the other two methods. Then ablation experiment was further done, and the result shows that multi-task learning is essential to achieving accurate quantifications. Finally, our method was also extended to segment the cones to extract the size information. Overall, the method proposed here demonstrated great performance in terms of accuracy and reliability, which can be used to efficiently quantify the subtle changes associated with the progression of many diseases affecting cones.