Adaptive optics optical coherence tomography (AO-OCT) provides state-of-the-art volumetric cellular-level imaging of the human retina in vivo. However, when coupled to AO, the OCT system's high transverse resolution reduces the depth of focus. A proven approach to this problem is acquiring AO-OCT volumes with the system focus set on different retinal layers and stitching the resulting volumes together. We demonstrate that this approach can be simplified using computational aberration correction (CAC). CAC enables us to correct AO-OCT volumes computationally in post-processing. So, we achieved an extended depth of focus without acquiring multiple AO-OCT volumes under the variable system focus settings.
The ability to measure retinal blood flow (RBF) accurately and reproducibly is crucial for diagnosing and monitoring ocular diseases such as glaucoma and hypertensive retinopathy. Impaired autoregulation of blood flow plays a key role in both the development and progression of glaucoma. Multimodal adaptive optics (mAO) using scanning laser ophthalmoscopy and optical coherence tomography offer superior spatial and temporal resolution and the ability to measure blood flow in retinal microvasculature. Here we evaluate RBF measurement reproducibility and repeatability using a mAO technique.
Adaptive optics optical coherence tomography (AOOCT) requires a dense sampling of the retina to visualize individual cones in the living human eye. This in turn increases the acquisition time and introduces susceptibility to eye motion artifacts. Here, we present hybrid transformer generative adversarial network (HT-GAN), an artificial intelligence technique that can improve the pixel resolution of images to better reveal cones from sparsely sampled AOOCT volumes. The method can potentially increase the speed of acquisition by four-fold while maintaining the visibility of individual cones despite a lower than ideal pixel sampling.
Although elevated intraocular pressure (IOP) is considered to be a major precursor for glaucoma, up to 45% of the patients with early glaucoma show signs of disease progression despite IOP reduction therapy. Studies have shown strong clinical evidence for abnormal ocular vessel function and impaired autoregulation of blood flow in early glaucoma subjects and its role in disease development and progression. Here we present direct measure of vascular dysfunction and autoregulation in three healthy human subjects using the erythrocyte mediate angiography and adaptive optics scanning laser ophthalmoscopy line-scan techniques. These novel quantitative blood flow metrics can potentially serve as a sensitive biomarker for early diagnosis and monitoring of ocular disease.
The photoreceptor (PR) – retinal pigment epithelium (RPE) – choriocapillaris (CC) complex is an extremely important group of layers in the outer retina. We demonstrate resolution of the CC vascular network across the macula, as well as the methodology to extract and quantify structural metrics from all three layers from averaged AO-OCT volumes. In diseased eyes, small changes in CC structure may portend the initiation of disease and therefore the investigation of CC structural changes may aid early disease diagnosis for many diseases, both prevalent and rare, that begin in the outer retina.
Retinal pigment epithelial (RPE) cells play an integral role in maintaining visual function and retinal health. Adaptive optics-optical coherence tomography (AO-OCT) has enabled the in vivo visualization of the hexagonal structure of RPE at cellular scale resolution. However, it is difficult to clearly visualize the RPE mosaic in single AO-OCT volumes due to inherent speckle noise, which can be overcome by averaging a large number of AO-OCT volumes acquired at sufficiently spaced time intervals to allow speckle decorrelation for improved cell contrast. Here, we present a deep learning based siamese discriminator (DS) generative adversarial network (GAN) to recover the hexagonal RPE mosaic from only a single AO-OCT volume. The DS provides additional feedback to the generator that resulted in improved visualization of RPE morphology compared to traditional GAN. Experimental results from five healthy subjects suggest that the RPE images generated using DS-GAN were comparable to ground truth images obtained by averaging multiple AO-OCT volumes. Quantitative comparison of cell-to-cell spacing, density, and image quality assessment metrics further confirmed the accuracy of recovered RPE mosaics relative to ground truth. These results establish a potential strategy in which deep learning can be leveraged to eliminate the need for volume averaging and speckle decorrelation for more efficient RPE imaging.
Recent clinical studies have shown that abnormal retinal blood is associated with many ocular diseases such as age-related macular degeneration, glaucoma, and diabetic retinopathy. Several ocular imaging techniques have been developed to measure retinal blood flow both invasively and non-invasively, including optical coherence tomography angiography (OCTA), erythrocyte mediated angiography (EMA), laser speckle imaging (LSI) and adaptive optics - scanning laser ophthalmoscopy (AO-SLO). Here we present a simple, compact, well-controlled clinical flow phantom model which allows flow evaluation across several techniques to aid in the clinical diagnosis of ocular diseases with abnormal blood flow.
Multiple sclerosis (MS) is a debilitating autoimmune disease characterized by lesions found in different regions of the central nervous system caused by overactive immune cells. MS also manifests in the retina, in which optic nerve pathology (such as optic neuritis) and neurodegenerative processes can affect inner retinal cells and structures. We observed the inner retina to be profoundly affected by MS, including nerve fiber bundle thinning, enlarged and lower density retinal ganglion cells, and the presence of microcysts. Longitudinal quantification of inner retinal changes enabled by AO-OCT may help accelerate the development of novel therapies for MS patients.
Adaptive optics (AO) has enabled microscopic views of retinal neurons and assessment of their function in living eyes when combined with different imaging modalities, such as scanning laser ophthalmoscopy (SLO) and optical coherence tomography (OCT). Here, we present a novel design of a multimodal AO imaging system, based on Fourier domain mode-locked (FDML) laser technology. The design allows simultaneous and registered collection of AO-SLO images and AO-OCT volumes at near video rates (13.4 Hz). In addition to the multimodal optical design, the system also features an additional stimulus port and software algorithms to provide multiple functional modes with which to investigate living human retinal cells. The optical system was designed in Zemax with spherical mirrors placed in an out-of-plane configuration to reduce system astigmatism. The system was found to achieve diffraction limited image quality across a 4.5° × 4.5° scanning field. The measured AO-OCT system axial resolution is 8.7 μm in the eye, sensitivity was measured at 88 dB with ~7 dB roll-off over the first ~2 mm. The multimodal system performance was demonstrated by imaging various retinal cells and vessels with co-registered AO-OCT and AO-SLO images. The multifunctional feature was demonstrated by measuring the light-induced phase change of the cone outer segment. The methods will enable development of more sensitive AO-based cellular biomarkers for improved retinal disease diagnosis and treatment.
KEYWORDS: Optical coherence tomography, In vivo imaging, Adaptive optics, Adaptive optics optical coherence tomography, Retinal scanning, Image quality, Human vision and color perception, Clinical trials
The development and application of adaptive optics (AO) in retinal imaging have enabled visualization of a plethora of retinal cells and structures. However, major hurdles exist for translating these achievements to the widely-available clinical devices for broad clinical applications. Here, by configuring a research grade AO – optical coherence tomography (AO-OCT) system to simulate a clinical OCT device, we provide evidence that clinical OCT systems have the potential to resolve individual ganglion cell layer somas and determine that a lateral sampling of ~1.5 µm/pixel is required to accurately quantify soma density and size.
Glaucoma is an optic neuropathy characterized by loss of retinal ganglion cells and their axons. Glaucoma has a strong vascular component and decreased macular vessel density is known to be associated with glaucomatous damage. Adaptive optics – optical coherence tomography allows for the simultaneous quantification of vascular and ganglion cell densities. We observed a moderately strong correlation between ganglion cell and vessel densities across the macula, as well as some correlation at individual locations. Vascular density may prove to be a useful surrogate biomarker of glaucoma progression and with further study reveal new information on impairment of neurovascular coupling in glaucoma.
Glaucoma causes progressive loss and degeneration of retinal ganglion cell (RGC) axons and their somas. Until recently, cellular-level RGC imaging in live human subjects was not possible. Our initial cross-sectional study revealed differences in morphological parameters in ganglion cell layer (GCL) soma with glaucoma. The manner in which cellular-level morphology is altered during glaucoma progression has yet to be discovered. In this study, we used AO-OCT to track RGC morphology in one glaucoma patient with a hemifield defect. We found dynamic tissue remodeling in the GCL, resulting in a decrease in soma density and an increase in soma size.
Microglia are central nervous system macrophages and the first responders to neural injury. Herein we characterize their distribution and motility in human eyes using a multimodal AO system. In healthy eyes, microglia are absent in the central macula up to ~5º eccentricity but their density increases monotonically at higher eccentricities. Microglia density decreases linearly with age. ILM microglia are relatively immobile for durations up to two weeks but their processes re-orient over timescales as short as minutes. The density, motility, and reactive state of microglia may serve as an ocular disease biomarker for early detection and progression monitoring.
Quantitative features of individual ganglion cells (GCs) are potential paradigm changing biomarkers for improved diagnosis and treatment monitoring of GC loss in neurodegenerative diseases like glaucoma and Alzheimer’s disease. The recent incorporation of adaptive optics (AO) with extremely fast and high-resolution optical coherence tomography (OCT) allows visualization of GC layer (GCL) somas in volumetric scans of the living human eye. The current standard approach for quantification – manual marking of AO-OCT volumes – is subjective, time consuming, and not practical for large scale studies. Thus, there is a need to develop an automatic technique for rapid, high throughput, and objective quantification of GC morphological properties. In this work, we present the first fully automatic method for counting and measuring GCL soma diameter in AO-OCT volumes. Aside from novelty in application, our proposed deep learningbased algorithm is novel with respect to network architecture. Also, previous deep learning OCT segmentation algorithms used pixel-level annotation masks for supervised learning. Instead in this work, we use weakly supervised training, which requires significantly less human input in curating the training set for the deep learning algorithm, as our training data is only associated with coarse-grained labels. Our automatic method achieved a high level of accuracy in counting GCL somas, which was on par with human performance yet orders of magnitude faster. Moreover, our automatic method’s measure of soma diameters was in line with previous histological and in vivo semi-automatic measurement studies. These results suggest that our algorithm may eventually replace the costly and time-consuming manual marking process in future studies.
High-resolution adaptive optics (AO) imaging of retinal neurons in living eyes holds promise for improved diagnosis and better assessment of treatment outcomes for retinal diseases. By integrating different imaging modalities, such as scanning laser ophthalmoscopy (SLO) and optical coherence tomography (OCT), AO has enabled microscopic views of different retinal neurons, including recently reported retinal ganglion cells. In this study, we present a novel design of a multimodal adaptive optics imaging system to investigate the microscopic structure of living human retina. The optical system was designed using Zemax ray tracing software. The system performance was evaluated in terms of image quality and beam displacement. Optical performance was predicted to achieve diffraction limited image quality at the retinal plane and beam displacement was predicted to be >7× smaller than the pitch of Shack-Hartmann lenslet at the pupil planes for scan angles over 3.6°×3.6° field of view. The initial human subject images are presented. High quality photoreceptor images were acquired in both AO-SLO channel and AO-OCT channel simultaneously at 3° temporal from the fovea. Individual cones are delineated in AO-SLO image, the corresponding AO-OCT image showed four main reflections from outer retina, namely external limiting membrane, cone inner segment/outer segment junction, cone outer segment tip, and retinal pigment epithelium. The system allows flexibly alternating between AO-SLO and AO-OCT modes, which provides complementary views of retinal cells, and the potential to improve disease diagnosis and treatment.
Adaptive optics-enabled optical coherence tomography (AO-OCT) and scanning laser ophthalmoscopy (AO-SLO) devices can resolve retinal cones and rods in three dimensions. To evaluate the improved resolution of AO-OCT and AO-SLO, a phantom that mimics retinal anatomy at the cellular level is required. We used a two-photon polymerization approach to fabricate three-dimensional (3D) photoreceptor phantoms modeled on the central foveal cones. By using a femtosecond laser to selectively photocure precise locations within a liquid-based photoresist via two-photon absorption, we produced high-resolution phantoms with μm-level dimensions similar to true anatomy. In this work, we present two phantoms to evaluate the resolution limits of an AO imaging system: one that models only the outer segments of the photoreceptor cells at varying retinal eccentricities and another that contains anatomically relevant features of the full-length photoreceptor. With these phantoms we are able to quantitatively estimate transverse resolution of an AO system and produce images that are comparable to those found in the human retina.
The inner retina is critical for visual processing, but much remains unknown about its neural circuitry and vulnerability to disease. A major bottleneck has been our inability to observe the structure and function of the cells composing these retinal layers in the living human eye. Here, we present a noninvasive method to observe both structural and functional information. Adaptive optics optical coherence tomography (AO-OCT) is used to resolve the inner retinal cells in all three dimensions and novel post processing algorithms are applied to extract structure and physiology down to the cellular level. AO-OCT captured the 3D mosaic of individual ganglion cell somas, retinal nerve fiber bundles of micron caliber, and microglial cells, all in exquisite detail. Time correlation analysis of the AO-OCT videos revealed notable temporal differences between the principal layers of the inner retina. The GC layer was more dynamic than the nerve fiber and inner plexiform layers. At the cellular level, we applied a customized correlation method to individual GCL somas, and found a mean time constant of activity of 0.57 s and spread of ±0.1 s suggesting a range of physiological dynamics even in the same cell type. Extending our method to slower dynamics (from minutes to one year), time-lapse imaging and temporal speckle contrast revealed appendage and soma motion of resting microglial cells at the retinal surface.
Absorption of light by photoreceptors initiates vision, but also leads to accumulation of toxic photo-oxidative
compounds in the photoreceptor outer segment (OS). To prevent this buildup, small packets of OS discs are
periodically pruned from the distal end of the OS, a process called disc shedding. Unfortunately dysfunction in
any part of the shedding event can lead to photoreceptor and RPE dystrophy, and has been implicated in
numerous retinal diseases, including age related macular degeneration and retinitis pigmentosa. While much is
known about the complex molecular and signaling pathways that underpin shedding, all of these advancements
have occurred in animal models using postmortem eyes. How these translate to the living retina and to humans
remain major obstacles. To that end, we have recently discovered the optical signature of cone OS disc shedding
in the living human retina, measured noninvasively using optical coherence tomography equipped with adaptive
optics in conjunction with post processing methods to track and monitor individual cones in 4D. In this study,
we improve on this method in several key areas: increasing image acquisition up to MHz A-scan rates,
improving reliability to detect disc shedding events, establishing system precision, and developing cone
tracking for use across the entire awake cycle. Thousands of cones were successfully imaged and tracked over
the 17 hour period in two healthy subjects. Shedding events were detected in 79.5% and 77.4% of the tracked
cones. Similar to previous animal studies, shedding prevalence exhibited a diurnal rhythm. But we were
surprised to find that for these two subjects shedding occurred across the entire day with broad, elevated
frequency in the morning and decreasing frequency as the day progressed. Consistent with this, traces of the
average cone OS length revealed shedding dominated in the morning and afternoon and renewal in the evening.
Retinal pigment epithelium (RPE) cells are vital to health of the outer retina, however, are often
compromised in ageing and ocular diseases that lead to blindness. Early manifestation of RPE disruption
occurs at the cellular level, but while in vivo biomarkers at this scale hold considerable promise, RPE
cells have proven extremely challenging to image in the living human eye. Recently we addressed this
problem by using organelle motility as a novel contrast agent to enhance the RPE cell in conjunction with
3D resolution of adaptive optics-optical coherence tomography (AO-OCT) to section the RPE layer. In
this study, we expand on the central novelty of our method – organelle motility – by characterizing the
dynamics of the motility in individual RPE cells, important because of its direct link to RPE physiology.
To do this, AO-OCT videos of the same retinal patch were acquired at approximately 1 min intervals or
less, time stamped, and registered in 3D with sub-cellular accuracy. Motility was quantified by an
exponential decay time constant, the time for motility to decorrelate the speckle field across an RPE cell.
In two normal subjects, we found the decay time constant to be just 3 seconds, thus indicating rapid
motility in normal RPE cells.
Retinal pigment epithelium (RPE) cells are vital to health of the outer retina, but are often compromised in ageing and
major ocular diseases that lead to blindness. Early manifestation of RPE disruption occurs at the cellular level, and while
biomarkers at this scale hold considerable promise, RPE cells have proven extremely challenging to image in the living
human eye. We present a novel method based on optical coherence tomography (OCT) equipped with adaptive optics (AO)
that overcomes the associated technical obstacles. The method takes advantage of the 3D resolution of AO-OCT, but more
critically sub-cellular segmentation and registration that permit organelle motility to be used as a novel contrast mechanism.
With this method, we successfully visualized RPE cells and characterized their 3D reflectance profile in every subject and
retinal location (3° and 7° temporal to the fovea) imaged to date. We have quantified RPE packing geometry in terms of
cell density, cone-to-RPE ratio, and number of nearest neighbors using Voronoi and power spectra analyses. RPE cell
density (cells/mm2) showed no significant difference between 3° (4,892±691) and 7° (4,780±354). In contrast, cone-to-
RPE ratio was significantly higher at 3° (3.88±0.52:1) than 7° (2.31± 0.23:1). Voronoi analysis also showed most RPE
cells have six nearest neighbors, which was significantly larger than the next two most prevalent associations: five and
seven. Averaged across the five subjects, prevalence of cells with six neighbors was 51.4±3.58% at 3°, and 54.58±3.01%
at 7°. These results are consistent with histology and in vivo studies using other imaging modalities.
It has been long established that photoreceptors capture light based on the principles of optical waveguiding. Yet after
decades of experimental and theoretical investigations 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, we took advantage of adaptive-optics optical coherence tomography (AO-OCT, λc=785
nm) to noninvasively image in three dimensions the reflectance profiles generated in the inner and outer segments (IS,
OS) of cones. Mode content was examined over a range of cone diameters by imaging cones from 0.6° to 10° retinal
eccentricity (n = 1802). Fundamental to the method was extraction of reflections at the cone IS/OS junction and cone
outer segment tip (COST). Modal content properties of size, circularity and orientation were quantified using second
moment analysis. Analysis of the cone reflections indicates waveguide properties of cone IS and OS depend on
segment diameter. 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, and show no orientation preference.
Optical coherence tomography with adaptive optics (AO-OCT) is a noninvasive method for imaging the living retina
at the microscopic level. We used AO-OCT technology to follow changes in cone photoreceptor outer segment (OS)
length and reflectance. To substantially increase sensitivity of the length measurements, a novel phase retrieval
technique was demonstrated, capable of detecting changes on a nanometer scale. We acquired volume videos of
0.65°x0.65° retinal patches at 1.5° temporal to the fovea over 75 and 105 minutes in two subjects. Volumes were
dewarped and registered, after which the cone intensity, OS length, and referenced phase difference were tracked over
time. The reflections from inner segment/OS junction (IS/OS) and posterior tips of OS (PT) showed significant
intensity variations over time. In contrast, the OS length as measured from the intensity images did not change,
indicative of a highly stable OS length at least down to the level of the system's axial resolution (3μm). Smaller axial
changes, however, were detected with our phase retrieval technique. Specifically, the PT-IS/OS phase difference for
the same cones showed significant variation, suggesting real sub-wavelength changes in OS length of 125±46 nm/hr
for the 22 cones followed. We believe these length changes are due to the normal renewal process of the cone OS that
elongate the OS at a rate of about 100 nm/hr. The phase difference measurements were strongly correlated among Alines
within the same cone (0.65 radians standard deviation) corresponding to a length sensitivity of 31 nm, or ~100
times smaller than the axial resolution of our system.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.