Improving Optical Coherence Tomography Capabilities for Imaging the Eye




In vivo retinal imaging of NMDA-induced neurodegeneration in aging mice with combined optical coherence tomography and fluorescence imaging 


Dynamic focusing spiral scan optical coherence tomography for high resolution, wide field, corneal and anterior chamber imaging


Grand Challenge: Engineering the Tools of Scientific Discovery

Date Submitted: April 25, 2018

Michelle Wei

Anthony Kuo M.D.

Department of Biomedical Engineering

Duke University Pratt School of Engineering





Optical coherence tomography (OCT) is a non-invasive optical imaging technique that performs high-resolution, cross-sectional, 3D imaging of biological tissue by measuring backscattered light. This technology is powerful because it provides real-time, in vivo visualization of tissue microstructure; however, its imaging depth is small (~2 mm) because it is limited by light scattering and attenuation in tissue. These characteristics are optimal for imaging structures in the eye because the transparency of the eye provides easy optical access to the anterior segment and retina. As such, it has become a standard in clinical ophthalmology.


In the first part of my project, I augmented OCT function in the retina by improving a specially designed system for simultaneous OCT and fluorescence imaging to concurrently visualize retinal layers and fluorescently labeled neurons in mouse retina. This is the first system built for OCT and fluorescence using a single illumination source and has significant advantages compared to acquiring these images separately. These advantages include decreased light exposure, less imaging time, more efficient post-processing, and one field of view for both images. Subsequently, this system was used to image neurodegeneration in N-Methyl-D-Aspartate (NMDA)-treated, fluorescently labeled mouse models. NMDA injections have been shown to lead to decreased neuron population and retinal thinning. I tested the hypothesis that retinal changes present differentially in younger and older mouse models. Results from qualitative analysis of retinal cross sections and fluorescence intensity demonstrated disruptions in retinal morphology in N-Methyl-D-Aspartate (NMDA)-treated eyes of mouse models aged 8 and 10 months, but not in those aged 3 months. We conclude that simultaneous OCT and fluorescence imaging revealed neurodegenerative retinal changes that present differently as a function of age.


For part two of the project, I focused on improving OCT imaging of the ocular anterior segment by dynamically adjusting the focus during one volume acquisition. Imaging of the ocular anterior segment with OCT typically requires trade-offs between resolution and depth-of-focus due to corneal curvature and the deep anterior chamber. The cornea is a curved structure of approximately 11 mm x 11 mm x 4 mm. This large imaging range and curved anatomy contribute to decreases in signal away from the corneal apex, limits visualization of anterior chamber anatomy, and negatively impacts the performance of automatic segmentation programs needed for large scale analysis. Dynamically adjusting the focus of the beam as it traverses the cornea circumvents the lateral/depth-of-focus tradeoff. Additionally, a constant linear velocity (CLV) spiral scan, which images continuously from the corneal apex to the periphery and back, can improve the speed and efficiency of OCT acquisitions. Together, spiral scanning and dynamic refocusing allow for increased SNR in full range corneal and anterior segment OCT imaging. A subject was imaged with this system demonstrating increased intensity away from the corneal apex compared to a conventional, non-refocusing corneal OCT scan of the same cornea.



Table of Contents

Part 1: In vivo retinal imaging of NMDA-induced neurodegeneration in aging mice with combined optical coherence tomography and fluorescence imaging.



Imaging technologies such as optical coherence tomography (OCT) and fluorescence imaging reveal different types of information. OCT measures the magnitude and echo time delay of backreflected light from internal microstructure in tissues to generate high resolution cross sectional images in vivo. It can obtain axial resolutions of 1-15 μm and is able to detect changes in retinal thickness in label free tissues (Drexler and Fujimoto, 2008). OCT primarily yields anatomical structures in the retina in vivo, but is limited in visualizing individual retinal neurons.  In contrast, fluorescence imaging is typically used in vitro to analyze histological changes. A limitation of this modality is that it requires fluorescent protein markers or immunostaining, and is difficult to perform in vivo. By utilizing both techniques, a dual-modality system is well-suited to examine retinal thickness and neuronal population in the process of neurodegeneration in the retina of young and aging mice.


The importance of this investigation lies in the growing aging population and the lack of accurate animal models for age-related neurodegenerative diseases. The population of individuals over 60 is projected to rise from 900 million in 2015 to 2 billion in 2050 (WHO, 2015). This increasing population is expected to lead to higher rates of age-related neurodegenerative diseases such as Parkinson’s Disease (PD), Alzheimer’s Disease (AD), Amyotrophic Lateral Sclerosis (ALS), and glaucoma (Reitz et al., 2011; Reeve et al., 2014). Increased research in these fields has seen success in the laboratory, but many fail in clinical trials (Glaser, 1997; Evans and Barker, 2008; Burns and Verfaillie, 2015). Perhaps one reason for this breakdown in translation is the continuing use of immature nervous systems and mutant animal models that develop exaggerated symptoms early on and die young (Gordon, 2013; Blesa and Przedborski, 2014; Neha et al., 2014). In contrast, individuals with age-related neurodegenerative diseases develop symptoms and die at old ages. Previous studies have shown that age has an effect on the survival of facial motoneurons (Aperghis et al., 2003), which suggests that age could be an important factor in the translation of neurodegenerative treatments from animal models to the clinic. To extend this evidence, we examine the effects of age on N-Methyl-D-Aspartate (NMDA)-induced retinal neurodegeneration in mouse models.


NMDA has been commonly used to induce neurodegeneration in young mice, with significant effects on retinal thickness and neuron population following intravitreal injections (Nakano et al., 2011; Ohno et al., 2013, Li et al., 2002). To our knowledge, however, no study has demonstrated its effects on mice aged over 6 months. As such, we will use a custom combined optical coherence tomography (OCT) and fluorescence imaging system to investigate the differential effects of NMDA on young and old mice.




Mouse protocol

The retinas of C57 and Thy1-YFP mice aged 3, 7, and 10 months were imaged using the custom simultaneous fluorescence and OCT system to qualitatively visualize the baseline amount of retinal ganglion cells (RGC) and establish retinal thickness. These mice were anesthetized by means of intraperitoneal injection of ketamine (100mg/mL, 120mg/kg) and xylazine (100 mg/mL, 16mg/kg) mixture. One drop of dilating agent (Tropicamide 1%) was applied to each eye prior to imaging. Once the mice were deemed fully anesthetized, one drop of hydrating gel was added to each eye to prevent corneal desiccation, and they were placed onto a 37°C thermostatic heating pad and into an animal holder consisting of a plastic cassette and bite bar. The mice were positioned depending on which eye was to be scanned, the platform was adjusted vertically, and the cassette was rotated to allow the selected eye to face upward, toward the scanning laser. Images were taken of one quadrant (superior in the left eye and inferior in the right eye) of each eye in each mouse. Depending on the quadrant to be scanned, the cassette was rotated and the bite bar was adjusted laterally to angle the eye optimally

Following baseline imaging, the experimental group (n = 5) underwent intravitreal injections with NMDA (20 nanomoles/2 mL), and the control group with PBS (n = 2). All eyes were imaged again 5 days following injections, and then flat-mounted for future confocal microscopy. All data were analyzed quantitatively.

Imaging protocol

The single illumination source dual OCT/fluorescence imaging system used wavelengths centered at 482 nm for both detection of YFP in the transgenic mice and for simultaneous OCT imaging. A rectangular scanning protocol was used and consisted of 1.8 x 1.8 x 0.9 mm lateral and axial dimensions, sampled at 670 mm depth with 500 A-scans by 500 B-scans performed at 10,000 A-scans/sec. 


System realignment of both OCT and fluorescent arms yielded images of comparable resolution to those taken for previous testing (Fig 1). Neurons can be counted, retinal layers are distinguishable, and retinal thickness can be measured.

Initial imaging of a Thy1-YFP mouse aged 8 months showed retinal changes not seen in younger mice (Fig 2). While the fluorescent images show a decreased number of fluorescent neurons at 5 days post-NMDA injection as expected (Fig 2A, B), the OCT image at day 5 demonstrates separation of the retinal pigment epithelium (RPE) and disruptions in the junction of inner and outer retinal segments (IS/OS) (Fig. 2C, D).


Figure 1. Comparable OCT and fluorescent images from before and after realignment. A) Fluorescent image from May 2017 B) Fluorescent image taken after system realignment with comparable resolution to A. C) OCT image from May 2017. D) OCT image taken after system realignment with comparable resolution to C. In both cross-sectional images, the retina is oriented such that the top layer is the outer retina.


Subsequent imaging of the remaining mice showed separation of the RPE in NMDA-treated eyes in mice models aged 8 and 10 months, but not in those aged 3 months (Figure 3A). These morphologic changes were seen at 1.5 mm to 1.7 mm from the optic nerve and were distinct in OCT images but not in the fluorescent images. Both groups saw decreases in RGC population, observed qualitatively in fluorescence microscopy (Figure 3B), and both groups saw slight retinal thinning and ganglion cell layer (GCL) thinning. PBS controls did not experience retinal thinning or neurodegeneration.

Figure 2. Representative OCT and fluorescent images of a Thy-1 YFP retina in a mouse of 8 months. A) Baseline fluorescent image of the inferior quadrant of the right eye. B) Fewer fluorescent neurons are present in the same quadrant 5 days after intravitreal NMDA injection. C) Baseline OCT image of retinal cross section, location indicated by the yellow line in A. D) Separation of the RPE and disruptions in the junction of inner and outer retinal segments (IS/OS) are visible 5 days after intravitreal NMDA injection. The location of this cross section is indicated by the yellow line in C.

Figure 3. OCT and fluorescent images of young and old mice before and after NMDA treatment. A) Day 0 and PBS control retinas have normal morphology, NMDA treated 3-month old mice show slight retinal thinning, while NMDA treated 8- and 10- month old mice show extreme GCL thinning, RPE separation and, and disruptions in the inner and outer retinal segments. B) NMDA treated mice aged 8 months shows decreased fluorescent neuron population compared to Day 0 and PBS controls.


The results show that neurodegenerative retinal changes present differently as a function of age, particularly between the ages of 3 and 8 months. Interestingly, these changes can be seen in OCT images but not in fluorescence microscopy at 5 days post-treatment. This suggests that NMDA-induced neurodegeneration leads to retinal morphology transformations prior to detectable molecular changes in the neurons themselves, particularly in older mice. Longitudinal studies are required to follow the progression of these morphologic perturbations and how they can affect the eye or the whole model organism. Further quantitative analysis of neuronal counts and retinal layer thickness can also add to the validity of this study.

The mechanisms underlying the differential vulnerabilities of retinal neurons to neurodegenerative drugs at different ages are still unclear. However, the findings in this study may help us understand more about neurons in general. It has been shown that neuron excitotoxicity resulting from excessive activation of NMDA receptors enhance the vulnerability of local neurons, which leads to neurodegeneration (Hynd et al., 2004). Animal studies have revealed that NMDA antagonists can prolong survival in mouse models of ALS, but clinical studies fail to demonstrate any treatment effects in ALS patients (Dmitry et al., 2017). Perhaps one reason for this failure to translate is the use of young animal models for testing. There are inherent challenges to researching the aging nervous system: aged animals are not readily available, cost more to maintain or purchase, and many die of health problems or old age. However, if evidence continues to accumulate that immature nervous systems cannot sufficiently model age-related neurodegenerative disease, we must invest in research on aging animal models.




  1. Blesa J., Przedborski S. (2014). Parkinson’s disease: animal models and dopaminergic cell vulnerability.  Neuroanat.8:155. 10.3389/fnana.2014.00155
  2. Burns T. C., Verfaillie C. M. (2015). From mice to mind: Strategies and progress in translating neuroregeneration.  J. Pharmacol. 759, 90–100. 10.1016/j.ejphar.2015.03.041
  3. Drexler W. and Fujimoto J. G. “State-of-the-art retinal optical coherence tomography”. Retinal Eye Res., 2008. Vol. 27, pp. 6-20.
  4. Evans J. R., Barker R. A. (2008). Neurotrophic factors as a therapeutic target for Parkinson’s disease. Expert Opin. Ther. Targets12, 437–447. 10.1517/14728222.12.4.437
  5. Glaser V. (1997). Cephalon adds to clinical disappointment in 1997.  Biotechnol.15, 495–496. 10.1038/nbt0697-495
  6. Gordon P. H. (2013). Amyotrophic lateral sclerosis: an update for 2013 clinical features, pathophysiology, management and therapeutic trials. Aging Dis.4, 295–310. 10.14336/AD.2013.0400295
  7. Li, Yan, et al. “p53 regulates apoptotic retinal ganglion cell death induced by N-methyl-D-aspartate.” Mol Vis9 (2002): 341-350
  8. McNabb R., Blanco T., Bomze H., Tseng H., Saban D., Izatt J., Kuo A. Method for single illumination source combined optical coherence tomography and fluorescence imaging to fluorescently labeled ocular structures in transgenic mice. Experimental Eye Research. 2016, vol. 151, pp. 68-74.
  9. Nakano, Noriko, et al. “Longitudinal and simultaneous imaging of retinal ganglion cells and inner retinal layers in a mouse model of glaucoma induced by N-methyl-D-aspartate.” Investigative ophthalmology & visual science12 (2011): 8754-8762.
  10. Neha Sodhi, R. K., Jaggi A. S., Singh N. (2014). Animal models of dementia and cognitive dysfunction. Life Sci.109, 73–86. 10.1016/j.lfs.2014.05.017
  11. Ohno, Yuta, et al. “Thickness mapping of the inner retina by spectral-domain optical coherence tomography in an N-methyl-D-aspartate-induced retinal damage model.” Experimental eye research113 (2013): 19-25.
  12. Petrov, Dmitry, et al. “ALS Clinical Trials Review: 20 Years of Failure. Are We Any Closer to Registering a New Treatment?.” Frontiers in aging neuroscience9 (2017).
  13. Reeve A., Simcox E., Turnbull D. (2014). Ageing and Parkinson’s disease: why is advancing age the biggest risk factor?Ageing Res. Rev. 14, 19–30. 10.1016/j.arr.2014.01.004
  14. Reitz C., Brayne C., Mayeux R. (2011). Epidemiology of Alzheimer disease.  Rev. Neurol.7, 137–152. 10.1038/nrneurol.2011.2
  15. World Health Organization (2015). “Ageing and health.” World Health Organization,




Part 2: Dynamic Focusing Spiral Scan OCT for High Resolution, Wide Field, Corneal and Anterior Chamber Imaging


Details of part 2 are embargoed  until publication of this research.