Embed Size (px)
Citation preview
Aberration correction with sensorless adaptive
optics for imaging the mouse retina
by
Daniel John Wahl
B.S., University of Northern British Columbia, 2014
Thesis Submitted in Partial Fulfillment of the
Requirements for the Degree of
Doctor of Philosophy
in the
School of Engineering Science
Faculty of Applied Sciences
© Daniel John Wahl 2019
SIMON FRASER UNIVERSITY
Summer 2019
Copyright in this work rests with the author. Please ensure that any reproduction or re-use is done in accordance with the relevant national copyright legislation.
ii
Approval
Name:
Degree:
Title:
Examining Committee:
Date Defended/Approved:
Daniel John Wahl
Doctor of Philosophy
Aberration correction with sensorless adaptive optics for imaging the mouse retina
Chair: Bonnie Gray Professor
Marinko V. Sarunic Senior Supervisor Professor
Yifan Jian Supervisor Assistant Professor
Mirza Faisal Beg Supervisor Professor
Robert J. Zawadzki Supervisor Associate Research Professor
Pierre Lane Internal Examiner Associate Professor of Professional Practice
Jennifer Hunter External Examiner Associate Professor Flaum Eye Institute Department of Ophthalmology Department of Biomedical Engineering Center for Visual Science The Institute for Optics University of Rochester
May 2, 2019
iii
Ethics Statement
iv
Abstract
Small animals, such as mice, are commonly used in biomedical research as models for
studying human diseases. Imaging the retina in a living animal can provide valuable
insights into the causes and mechanisms of vision loss. However, often imaging in vivo
results in low resolution due to optical aberrations that can be caused by the biological
tissue in front of the retina. Imaging systems that could non-invasively image the mouse
retina with cellular-level resolution would be beneficial to many vision scientists.
Adaptive optics (AO) is a technology that was originally developed for astronomers to
image through the turbulent atmosphere. AO technology has been extended for
microscopy and ophthalmoscopy to restore imaging performance lost due to optical
aberrations from biological samples. Often, AO systems employ a wavefront sensor for
direct measurement of the aberrations. Alternatively, Sensorless AO (SAO) has been
implemented for imaging into tissue with multiple scattering layers, which can confound
the optical wavefront measurements from a single imaging plane.
In this thesis, I present several imaging systems for imaging the mouse retina with cellular-
level resolution by using custom and novel SAO methods. The imaging modalities include
Scanning Laser Ophthalmoscopy with fluorescence detection, Optical Coherence
Tomography, and Two-Photon Excited Fluorescence imaging. The simple and robust
optical designs in this thesis feature wide imaging field of views for navigation and a
compactable system layout. Using SAO enables depth-resolved aberration correction in
the different layers of the mouse retina. My results demonstrate detailed non-invasive
cellular imaging capabilities in the living mouse eye of GFP labelled cells, nerve fibers
bundles, volumetric imaging of vasculature, as well as the RPE mosaic of the outer retina.
Keywords: Sensorless Adaptive Optics; Scanning Laser Ophthalmoscopy; Optical
Coherence Tomography; Two-Photon Excited Fluorescence; Mouse
Retina
v
Dedication
To my grandfather, Edward Hark.
vi
Acknowledgements
I would like to acknowledge the contributions of my supervisor, Dr. Marinko
Sarunic. His dedication to research, teaching, and his students is above and beyond,
which has enabled the work presented in this thesis. I would like to thank all of the
supervisors including, Drs. Marinko Sarunic, Yifan Jian, Mirza Faisal Beg, and Robert
Zawadzki for sharing their knowledge through mentorship and guidance during my
program at Simon Fraser University. It has been a privilege to work with this group of
advisors who have inspired me to keep striving.
I am grateful to all of the past and present members of the Biomedical Optics
Research Group for making this a great place to be. I am thankful for the collaborative and
friendly environment created by the people in our group.
There have been many people who have helped me along the way. However, in
particular I would like to thank Dr. Pengfei Zhang for his contributions to the work
presented in Chapter 5. Also, I would like to thank Ms. Christine Huang for her
contributions in Chapter 4 during her undergraduate thesis. And, a special thanks to Mr.
Ringo Ng for supporting everyone in the lab with great designs and fabrication work,
presented in Chapter 6.
Finally, I would like to thank my all of my family and friends for supporting me
during many years of post-secondary school and into the future.
vii
Table of Contents
Approval .......................................................................................................................... ii
Ethics Statement ............................................................................................................ iii
Abstract .......................................................................................................................... iv
Dedication ....................................................................................................................... v
Acknowledgements ........................................................................................................ vi
Table of Contents .......................................................................................................... vii
List of Tables ................................................................................................................... x
List of Figures................................................................................................................. xi
List of Acronyms .......................................................................................................... xviii
Chapter 1. Introduction .............................................................................................. 1
1.1. Overview ............................................................................................................... 1
1.2. The mouse eye ...................................................................................................... 2
1.3. Imaging the mouse retina ...................................................................................... 6
1.4. Outline ................................................................................................................... 8
1.5. Contributions ......................................................................................................... 9
Chapter 2. Background on retinal imaging systems and adaptive optics ............ 10
2.1. Scanning Laser Ophthalmoscopy ........................................................................ 10
2.2. Optical Coherence Tomography .......................................................................... 11
2.3. Fluorescence imaging ......................................................................................... 13
2.4. Adaptive optics for ophthalmic imaging ................................................................ 15
2.5. Summary ............................................................................................................. 19
Chapter 3. Wavefront sensorless adaptive optics fluorescence biomicroscope for in vivo retinal imaging in mice .................................................................... 20
3.1. Introduction .......................................................................................................... 20
3.2. Methods .............................................................................................................. 22
3.2.1. Mouse handling ........................................................................................... 23
3.2.2. Biomicroscope optical setup ........................................................................ 23
3.2.3. Image acquisition and optimization .............................................................. 25
3.3. Results ................................................................................................................ 27
3.3.1. WSAO f/c biomicroscope resolution ............................................................. 27
3.3.2. In vivo WSAO confocal fluorescence imaging of retinal ganglion cells ......... 28
3.3.3. In vivo WSAO confocal fluorescence imaging of retinal microglia cells ........ 30
3.4. Discussion ........................................................................................................... 31
3.5. Summary ............................................................................................................. 34
Chapter 4. Pupil segmentation adaptive optics for in vivo mouse retinal fluorescence imaging ........................................................................................ 35
4.1. Introduction .......................................................................................................... 35
4.2. Methods .............................................................................................................. 37
4.3. Discussion ........................................................................................................... 42
viii
4.4. Summary ............................................................................................................. 44
Chapter 5. Adaptive optics in the mouse eye: Wavefront sensing based vs. image-guided aberration correction ................................................................. 45
5.1. Introduction .......................................................................................................... 45
5.2. Methods .............................................................................................................. 47
5.2.1. AO SLO system description ......................................................................... 47
5.2.2. WFS AO description .................................................................................... 49
5.2.3. WFS-less AO algorithm. .............................................................................. 50
5.2.4. WFS and WFS-less AO system calibration .................................................. 51
5.2.5. Animal handling and image processing ....................................................... 53
5.3. Results ................................................................................................................ 54
5.3.1. WFS and WFS-less AO for phantom imaging, comparison of performance . 54
5.3.2. WFS and WFS-less AO comparison on mouse photoreceptor mosaic ........ 57
5.3.3. AO SLO reflectance imaging of an albino mouse strain ............................... 60
5.3.4. AO SLO fluorescence imaging of EGFP microglia cells ............................... 61
5.4. Discussion ........................................................................................................... 64
5.5. Summary ............................................................................................................. 68
Chapter 6. Multi-modal imaging .............................................................................. 69
6.1. Introduction .......................................................................................................... 69
6.2. Methods .............................................................................................................. 70
6.2.1. Optical design .............................................................................................. 70
6.2.2. Sensorless adaptive optics .......................................................................... 75
6.2.3. Animal handling ........................................................................................... 76
6.2.4. Image processing ........................................................................................ 76
6.3. Results ................................................................................................................ 77
6.3.1. Imaging without adaptive optics ................................................................... 77
6.3.2. Structural imaging with sensorless adaptive optics OCT and SLO ............... 79
6.3.3. Fluorescence imaging with sensorless adaptive optics ................................ 81
6.4. Discussion ........................................................................................................... 86
6.5. Summary ............................................................................................................. 88
Chapter 7. Non-invasive cellular-resolution imaging of the retina with two-photon excited fluorescence ............................................................................ 89
7.1. Introduction .......................................................................................................... 89
7.2. Methods .............................................................................................................. 90
7.2.1. System setup ............................................................................................... 90
7.2.2. Animal handling and image processing ....................................................... 93
7.3. Results ................................................................................................................ 95
7.3.1. Fluorescein angiography ............................................................................. 95
7.3.2. GFP and YFP labelled cells ......................................................................... 98
7.3.3. RPE imaging ............................................................................................. 101
7.4. Discussion ......................................................................................................... 105
7.5. Summary ........................................................................................................... 107
ix
Chapter 8. Future work and conclusion................................................................ 108
8.1. Technology refinement ...................................................................................... 108
8.2. Non-confocal Scanning Laser Ophthalmoscopy ................................................ 109
8.3. Extensions of two-photon excited florescence technology ................................. 110
8.4. Conclusion......................................................................................................... 111
References ................................................................................................................. 113
x
List of Tables
Table 2.1. Zernike polynomials, names and index up to the 5th radial order. ........... 16
Table 5.1. Key optical parameters of the AO-SLO system components .................. 48
Table 7.1. Laser specifications used for each fluorescent sample and the calculated resolution. .............................................................................................. 92
Table 7.2. Summary of mice that were used in this report. ...................................... 93
xi
List of Figures
Figure 1.1. Simplified schematic of the mouse eye compared to a human eye [6]. ..... 3
Figure 1.2. Organization of the retinal layers. The image is from Webvision [8] and used under the Creative Commons Licenses. .......................................... 4
Figure 1.3. The visual cycle and the location of each step. Image is from Wikimedia.org and used here under the Creative Commons License. ...... 5
Figure 1.4. The options for focusing light onto the mouse retina: collimated light or focused light from an objective lens.......................................................... 7
Figure 2.1. Simplified Jablonski diagram of single photon excited fluorescence and two-photon excited fluorescence. The excitation light, λex, and emission light, λem. ................................................................................................ 14
Figure 3.1. Schematic of the WSAO f/c biomicroscope using 488 nm excitation from an Ar/Kr laser. Relay lenses are achromatic doublets. Other optical elements: 80/20 beam splitter (BS), dichroic mirror (DC), deformable mirror (DM), zero-order quarter wave plate (QWP), objective lens (OBJ), linear polarizer (LP), pinhole (PH), variable lens (VL), galvanometer scanning mirrors (GM). Electronic elements: avalanche photo diode (APD), photo multiplier tube (PMT). The images on the computer icon are representative images of the structural and fluorescence imaging channels................................................................................................. 25
Figure 3.2. WSAO modal hill-climbing algorithm flowchart for the fluorescence image optimization process; deformable mirror (DM), variable lens (VL)........... 27
Figure 3.3. US Air Force resolution target with line width 2.19 µm highlighted by the red rectangle to demonstrate the reflectance resolution. Scale bar: 50 µm. ............................................................................................................... 28
Figure 3.4. Images of 2.1 µm diameter fluorescent beads acquired (a) before WSAO optimization and (b) after optimization. (c) The line plots for a bead before and after optimization. Scale bars: 10 µm. ............................................. 28
Figure 3.5. (a,b) Ganglion cells labelled by EGFP comparing the images acquired before and after the WSAO optimization. These images are an average of 50 frames of an off-axis ganglion cell. Scale bars: 20 µm. ...................... 29
Figure 3.6. (a) The Zernike coefficients applied to the DM (deformable mirror) after the optimization. (b) The impact of the optimization on the intensity-based merit function are plotted for each mode. The intensity is normalized from zero when the DM is flat. The Zernike coefficients are reported by the OSA standard for optical aberrations of eyes [68]. (c) The intensity plot of a dendrite on the EGFP-labelled ganglion cell at the location and in the direction indicated by the arrows. ........................................................... 30
Figure 3.7. Images of EGFP-labelled retinal microglia cells acquired in vivo before and after WSAO correction with different field of views: (a), (b), and (c). Images (b) and (c) were taken at the same location with different field of views as indicated by the red dashed box. Each image is an average of 50 frames. Scale bars: 10 µm. ............................................................... 31
Figure 4.1. Schematic of the Scanning Laser Ophthalmoscope: 488 nm laser; dichroic mirror (DC); deformable mirror (DM); variable lens (VL);
xii
galvanometers (GM); {L1,L2, L3,L4,L5,L6} = {200,200,150,100,50,19} mm. (a),(b) PS-AO on 6 µm fluorescent beads with aberration correction (AO On) and without (AO Off). These images are an average of 30 frames. Scale bar: 8 µm. (c) Wavefront aberration map. (d) Normalized intensity plotted at the location indicated by the dashed lines with a ~30% increase in the peak intensity after correction. (e) The Zernike coefficients for the corrected wavefront. ............................................................................... 38
Figure 4.2. Aberration correction performed with both hill-climbing and PS-AO. (a) Image without aberration correction. (b) Correction performed with hill-climbing. (c) Correction performed with PS-AO. (d) The Zernike coefficients for the corrected wavefronts. ............................................... 40
Figure 4.3. PS-AO aberration correction on (a) static and (b, c) moving samples. The aberration correction was performed with (b) using the multiple intra-frame reference images and (c) a single reference image. ..................... 41
Figure 4.4. (a), (b) PS-AO for retinal fluorescein angiography with aberration correction (AO On) and without (AO Off) for two mice. In each panel, the top row of images (angular FOV 5.2°) is an optically zoomed in section of the bottom row of images (angular FOV 10.4°). Scale bars: 20 µm. (c) Zernike coefficients for the corrected wavefront. (d) On the top panel, the normalized intensity plot at the location indicated by the dashed lines had a ~30% increase in the peak intensity after correction, and (d) on the bottom panel, the wavefront aberration map. ......................................... 42
Figure 5.1. Adaptive Optics Scanning Laser Ophthalmoscopy (AO-SLO) system schematic. The layout is presented in a scale drawing. Abbreviations: L#, lens; F#, filter; BS#, beamsplitter; M, mirror; SM, spherical mirror; DM, deformable mirror; D#, dichroic mirror; Hsc, horizontal resonant scanner; Vsc, vertical scanner; PMT, photomultiplier tube; P (circled in blue) optical planes conjugate with the pupil; SLD, superluminescent diode. Collimated beams are marked as dashed lines and focusing beams are marked as solid lines. The on-axis beams are represented by red lines and scanned beams by green and blue. Image credit: Pengfei Zhang. ........................ 48
Figure 5.2. Phantom imaging of fluorescent beads and wavefront measurements during Wavefront Sensor Adaptive Optics (WFS AO) and Wavefront Sensorless Adaptive Optics (WFS-less AO). (a) Fluorescence images of 30 µm beads on white paper with a 100 mm focal length model eye before AO, after WFS AO, and after WFS-less AO. For the inset image before AO, the pixel intensity values were multiplied by 8, so the beads could be visualized. (b) The increase in the fluorescence image quality during the WFS-less AO optimization. (c) The wavefront RMS excluding defocus, tip and tilt during WFS AO correction. (d) The wavefront RMS excluding defocus, tip and tilt during WFS-less AO optimization. (e) The Zernike decomposition of the wavefront measured before and after each method of AO correction. ....................................................................... 56
Figure 5.3. Imaging the mouse photoreceptor mosaic with Wavefront Sensor based Adaptive Optics (WFS AO) and Wavefront Sensorless Adaptive Optics (WFS-less AO). (a,b) Images after WFS AO and WFS-less AO. Scale bar: 10 µm. (c) The image quality improvement during WFS-less AO optimization. (d) The wavefront RMS during WFS-less AO optimization.
xiii
(e) The Zernike decomposition of the wavefront measured before and after each method of AO. ....................................................................... 58
Figure 5.4. (a, b) Further mouse photoreceptor imaging with Wavefront Sensor Adaptive Optics (WFS AO) and Wavefront Sensorless Adaptive Optics (WFS-less AO). Images mouse photoreceptor mosaic after WFS AO and WFS-less AO. Scale bar 10 µm. The Zernike decomposition of the wavefront measured before and after each method of AO. The wavefront RMS during WFS-less AO optimization. The image quality improvement during WFS-less AO optimization. .......................................................... 59
Figure 5.5. SH-WFS measurements from an Albino mouse strain (BALB/cJ) retina. (a) The SH-WFS centroids of an albino mouse compared to a pigmented mouse. (b) The RMS of the wavefront measurement without defocus. (c) The image quality metric during WFS-less AO optimization. .................. 60
Figure 5.6. Imaging the inner retinal of an Albino mouse (BALB/cJ) retina with Wavefront Sensorless Adaptive Optics (WFS-less AO). Images of the retina vasculature before and after WFS-less AO in the Nerve Fiber Layer (NFL), and after WFS-less AO in the Plexiform Layer (IPL), and Outer Plexiform Layer (OPL). Scale bar: 10 µm. .............................................. 61
Figure 5.7. Imaging EGFP labelled microglia with Wavefront Sensor Adaptive Optics (WFS AO). (a) Reflectance imaging in the inner retinal blood vessels. (b) Fluorescence imaging of EGFP labelled microglia. (c) The fluorescence image superimposed in green on the reflectance image in magenta. Scale bar: 20 µm. (d) The measured wavefront RMS during WFS AO without defocus. (e) The wavefront measurements in Zernike decomposition before and after the WFS AO aberration correction. ............................... 62
Figure 5.8. (a) Imaging EGFP labeled microglia within the inner retina of a mouse with Wavefront Sensor based Adaptive Optics (WFS AO) and Wavefront Sensorless Adaptive Optics (WFS-less AO). Fluorescence image with WFS AO aberration correction (left). Fluorescence image with WFS-less AO aberration correction (middle). Fluorescence images before and after WFS-less AO with a ~40 µm FOV (right). Scale bar: 20 µm. (b) The intensity line plot between the red arrows on the WFS AO image and between the blue arrows on the WFS-less AO image............................. 63
Figure 5.9. (a) Imaging EGFP labeled microglia within the inner retina of a mouse with Wavefront Sensor based Adaptive Optics (WFS AO) and Wavefront Sensorless Adaptive Optics (WFS-less AO). Fluorescence image after WFS AO (left). Fluorescence image after WFS AO and WFS-less AO aberration correction of residual aberration (middle). Fluorescence image with a smaller FOV of microglia dendrites superimposed in green on the reflectance image of the retinal blood vessels in magenta (right). (b) The Zernike decomposition of the wavefront measured before WFS AO and after both methods of AO. Scale bar: 20 µm........................................... 64
Figure 6.1. (a) Schematic of Optical Coherence Tomography (OCT) and confocal Scanning Laser Ophthalmoscopy (SLO) system. The cyan represents the beam path of only 488 nm light, the green represents the beam path of only the fluorescence emission and the red represents the beam path of only the SLD light. The pink represents the co-aligned beam path of the 488 nm light, fluorescence emission, and SLD light. System components: Superluminescent diode (SLD), fiber coupler (FC), polarization controller
xiv
(PC), polarization beam splitter (PBS), dichroic mirror (DC), emission filter (EF), cold mirror (CM), variable focus lens (VFL), deformable mirror (DM), galvanometer-scanning mirrors (GM), quarter wave plate (QWP), photomultiplier tube (PMT), dispersion compensation block (DCB), mirror (M). Achromatic doublet lenses: L1=50mm, L2=150mm, L3=300mm, L4=75mm, L5=2x125mm, L6=2x50mm. (b) Computer simulation of optical layout on custom optical mounts using OpticStudio and SolidWorks. ..... 72
Figure 6.2. (a) Spot diagrams of the OCT light at 820 nm (red), 840 nm (pink) and 860 nm (purple) across a 15-degree FOV, where the black circle represents the Airy disk with a 2.1 µm radius. Spot diagrams of the 488 nm (blue) SLO light scanned across a 15-degree FOV with 0 D of vergence at the sample pupil plane and 7-degress with 20 D of vergence at the sample pupil plane where the black circle represents the Airy disk with a 1.2 µm radius. (b) The boundary of the imaging beam at the final pupil plane of the system. The black circle represents a 2 mm aperture. Each color represents a different scan position across a 15-degree and 7-degree FOV to simulate the pupil wander due to the space between the scanning mirrors in the optical design. ................................................... 74
Figure 6.3. (a) OCT B-scan across 50 degrees in the mouse retina and en face projection of the outer plexiform layer (OPL) across 44 degrees. The B-scan is an average of 200 consecutively acquired cross-sectional frames and the en face OCT image is an average of 5 frames. (b,c) Average of 5 adjacent OCT B-scans and an average of 5 en face OCT frames of the OPL. The B-scans are located at the position of the red dashed lines. Vertical scale bar: 50 µm. Horizontal scale bars: 100 µm. ...................... 78
Figure 6.4. Confocal SLO images of a mouse retina with 488 nm light. (a) Structural image of the nerve fiber layer from back-scattering. (b) Fluorescein angiography composited with a MIP from images of three different vascular layers. Scale bar: 100 µm. ....................................................... 79
Figure 6.5. (a) En face OCT-A images of the OPL in a mouse retina. (b) En face OCT intensity image from the same image data. (c) En face OCT-A images that were generated from the OPL (red), IPL (green), and NFL (blue). Scale bar: 50 µm. ................................................................................... 79
Figure 6.6. (a) En face images of the outer plexiform layer (OPL, top row, ~250 µm FOV) and nerve fiber layer (NFL, bottom row, ~280 µm FOV) retinal layers before and after Sensorless Adaptive Optics (SAO). SAO-OCT B-scans with the imaging focal plane on the OPL (red arrows) and NFL (blue arrows). (b) The normalized image quality for each step in the SAO optimization over two iterations and the Zernike coefficients selected for each iteration. Vertical scale bars: 50 µm. Horizontal scale bars: 20 µm.80
Figure 6.7. (a) Confocal SLO images before and after Sensorless Adaptive Optics (SAO) of the nerve fiber layer (NFL) with a FOV ~250 µm. Images of the outer plexiform layer (OPL) after SAO. (b) The normalized image quality metric values for each step used for the SAO optimization for each iteration. The Zernike coefficients selected for each iteration. Scale bar: 20 µm. .................................................................................................... 81
Figure 6.8. Confocal SLO images of a mouse retina with labelled retinal ganglion cells (Tg(Thy1-EGFP)MJrs/J). (a) Fluorescence images before and after Sensorless Adaptive Optics (SAO) and an intensity line plot between the
xv
blue arrows (before SAO) and red arrows (after SAO). (b) The left column presents structural images focused on the nerve fiber layer at a ~750 µm FOV (top) and ~230 µm FOV (bottom). The right column presents the structural image in magenta overlaid by the fluorescence image in green. The fluorescence image was composited from two different focal planes for the axon and the dendrites of the RGC. Scale bars: 50 µm. .............. 82
Figure 6.9. Confocal SLO fluorescein angiography of a mouse retinal vasculature after Sensorless Adaptive Optics. Images (left to right) of the nerve fiber layer (NFL), inner plexiform layer (IPL), outer plexiform layer (OPL), and the MIP with the NFL in red, IPL in green, and NFL in blue. Scale bar: 50 µm. ......................................................................................................... 83
Figure 6.10. Confocal SLO images with Sensorless Adaptive Optics of EGFP labelled microglia in the mouse retina (B6.129P-Cx3cr1{tm1Litt}/J) acquired at different focal position between the outer plexiform layer (OPL) and the nerve fiber layer (NFL) selected from Visualization 1 of reference [16]. The microglia images were color-coded in depth between the OPL and the NFL of the retina and rendered in 3D for Visualization 2 reference [16]. Scale bar: 20 µm. ........................................................................... 84
Figure 6.11. (a) Confocal SLO fluorescence images with Sensorless Adaptive Optics of EGFP labelled microglia in the mouse retina (B6.129P-Cx3cr1{tm1Litt}/J) from three time points in the time-lapse video from Visualization 3 reference [16]. (b) The microglia images color-coded with time. The white arrows 1-4 note areas of significant growth and retraction. Scale bar: 20 µm. ................................................................................... 85
Figure 6.12. (a) Confocal SLO fluorescence images with Sensorless Adaptive Optics of EGFP labelled microglia in the mouse retina (B6.129P-Cx3cr1{tm1Litt}/J) from three time points in the time-lapse video from Visualization 4 of reference [16] with an increase in laser power at 39 minutes. (b) The microglia images color-coded with time. The white arrows 1-2 note areas of significant growth and retraction. Scale bar: 20 µm. ......................................................................................................... 86
Figure 7.1. Schematic of the Sensorless Adaptive Optics (SAO) Optical Coherence Tomography (OCT) and Two-Photon Excitation Fluorescence (TPEF) imaging system. The imaging system was constructed with a pellicle beam splitter (PeBS), a variable focus lens (VFL), a deformable mirror (DM), a dichroic mirror (DcM), galvanometer-scanning mirrors (GM), emission filters (EF), a photo-multiplier tube (PMT), dispersion compensation (DC), and the following lenses: L1=100 mm, L2=300 mm, L3=400 mm, L4=100 mm, L5=2×125 mm, L6=2×50 mm. The reference arm denoted as a dashed line. ............................................................... 91
Figure 7.2. Optical Coherence Tomography (OCT) and Two-Photon Excited Fluorescence (TPEF) images of the mouse retina before (top row) and after (bottom row) OCT-guided Sensorless Adaptive Optics (SAO). The improvement in the OCT B-scan is shown in the left column, the improvement in the en face OCT is shown in the middle column, and the improvement in the TPEF is shown in the right column. The yellow arrows represent the imaging focal position and the line between the blue arrows represents the cross-sectional location of the OCT B-scans. Scale bars: 50 µm. .................................................................................................... 96
xvi
Figure 7.3. (a) OCT B-scans (top row), OCTA en face (middle row), and TPEF (bottom row) with the focal plane at the Outer Plexiform Layer (OPL), Inner Plexiform Layer (IPL), and Nerve Fiber Layer (NFL). In the right column, the images of the vascular layers were composited with a MIP. The red arrows point out connecting vessels in the TPEF. (b) Cross-sectional TPEF images (left) of the inner retinal vasculature before and after Adaptive Optics (SAO) acquired with a 25-step z-stack that was interpolated to 75 image pixels. The axial intensity profile plot between the red and blue arrows of the TPEF cross-sectional images. Scale bars: 50 µm. .................................................................................................... 98
Figure 7.4. TPEF imaging of GFP labelled microglia (B6.129P2(Cg)-Cx3cr1{tm1Litt}/J) in the mouse retina. (a) Single TPEF frame (left) and an average of 100 frames (right) at a ~0.8 mm FOV. The red square represents a 100 µm FOV to represent the scale of the microglia. Scale bar: 100 µm. (b) TPEF images of a GFP labelled microglia cells before (left) and after (right) Sensorless Adaptive Optics (SAO). (c) TPEF image after SAO. Scale bars: 20 µm. ................................................................ 99
Figure 7.5. Comparison of a GFP labelled retinal ganglion cell that was imaged using SAO TPEF (left) and using SAO SPEF with the same 200 µm FOV (middle). A SPEF image is also shown at a ~1.3 mm FOV (right), where the red square represents the 200 µm FOV that was used for the other images. Left scale bar: 20 µm. Right scale bar: 100 µm. ...................... 100
Figure 7.6. OCT B-scans (top row) and TPEF (middle row) imaging with the focal plane at the Nerve Fiber Layer (NFL), Inner Plexiform Layer (IPL), and Outer Plexiform Layer (OPL) of a Thy-1 YFP-16 Line (B6.Cg-Tg(Thy1-YFP)16Jrs/J) transgenic mouse. The blue arrow and yellow arrow point at fluorescently labelled cell bodies. The red arrow points at fluorescently labelled axons. In the bottom row, the OCTA en face image (magenta) was composited with the TPEF image (green). Vertical scale bar: 50 µm. Horizontal scale bars: 20 µm. ............................................................... 101
Figure 7.7. (a) The SAO-OCT B-scans in linear scale (top row) and the en face OCT (bottom row) with the focal plane at the Nerve Fiber Layer (NFL), Outer Plexiform Layer (OPL), and Retinal Pigment Epithelium (RPE) in the mouse retina. The en face OCT images were extracted between the cyan arrows (NFL), yellow arrows (OPL), and green arrows (RPE). The OCT B-scans were located between the red arrows on the en face OCT image. (b) TPEF images of the RPE of the mouse retina before and after SAO. (c) An intensity line plot between the blue arrows and the red arrows on the TPEF images of the RPE mosaic. Scale bars 50 µm. ..................... 102
Figure 7.8. (a) TPEF images of the RPE (left), en face OCT (middle), and OCT B-scans (right). (b) TPEF images of the RPE (left), en face OCT (middle), and OCT B-scans (right) from the same mouse four days later. (c) The digital enlargement of the TPEF images on day 1 (green) and day 4 (magenta), which were combined with a MIP. Scale bars 50 µm. ......... 103
Figure 7.9. TPEF from the RPE layer of the mouse retina in three different mouse strains, including a pigmented B6 mouse (C57BL/6J), an albino B6 mouse (B6(Cg)-Tyr{c-2J}/J), and a pigmented rpe65 mouse (B6(A)-Rpe65{rd12}/J). Scale bar 100 µm. ...................................................... 104
xvii
Figure 7.10. TPEF image of a pigmented rpe65 mouse (B6(A)-Rpe65{rd12}/J) with different central wavelengths, including 760 nm, 780 nm, 800 nm, and 820 nm. The red arrow highlights an RPE cell where the fluorescence near the cell membrane is reduced with longer wavelengths. Scale bar 50 µm. ....................................................................................................... 105
Figure 8.1. Volumetric averaging of 150 OCT volumes. Scale bar: 50 µm. ............ 109
xviii
List of Acronyms
ADD Airy Disk Diameters
AMD Age-Related Macular Degeneration
ANSI American National Standards Institute
AO Adaptive Optics
APD Avalanch Photodiode
Ar/Kr Argon/Krypton
BS Beam Splitter
CMOS Complementary Metal-Oxide-Semiconductor
CNIB Canadian National Institute for the Blind
CS Coordinate Search
DC Dichroic Mirror
DM Deformable Mirror
DMD Digital-Micro-Device
DONE Data-based Online Nonlinear Extremum-seeker
EGFP Enhanced Green Fluorescent Protein
f/c Fluorescence/Confocal
FA Fluorescein Angiography
FAD Flavin Adenine Dinucleotide
FLIM Fluorescence Lifetime Imaging Microscopy
FOV Field of View
FWHM Full Width at Half Maximum
GCL Ganglion Cell Layer
GFP Green Fluorescent Protein
GM Galvanometer Mirror(s)
GPU Graphics Processing Unit
IPL Inner Plexiform Layer
IS Inner Segment
LP Linear polarizer
MEMS Micro-Electro-Mechanical Systems
MIP Maximum Intensity Projection
MIRT Medical Image Registration Toolbox
MPE Maximum Permissible Exposure
xix
MSPS Mega-Samples Per Second
NA Numerical Aperture
NADH Nicotinamide Adenine Dinucleotide
NFL Nerve Fiber Layer
NI National Instruments
NIR Near Infrared
OBJ Objective Lens
OCT Optical Coherence Tomography
OCT-A Optical Coherence Tomography Angiography
ONH Optic Nerve Head
OPL Outer Plexiform Layer
OS Outer Segment
OSA Optical Society of America
PBS Polarization Beam Splitter
PH Pinhole
PMT Photomulitpler Tube
PS-AO Pupil Segmentation Adaptive Optics
PSF Point Spread Function
QWP Quarter Wave Plate
RGC Retinal Ganglion Cell
RMS Root-Mean-Squared
ROI Region of Interest
RPE Retina Pigment Epithelium
SAO Sensorless Adaptive Optics
SD Spectral Domain
SH Shack-Hartmann
SH-WFS Shack-Hartmann Wavefront Sensor
SLD Superluminescent Diode
SLM Spatial Light Modulator
SLO Scanning Laser/Light Ophthalmoscopy
SNR Signal-to-Noise Ratio
SPEF Single Photon Excited Fluorescence
SS Swept Source
SVD Singular-Value Decomposition
xx
TPEF Two-Photon Excited Fluorescence
TWCoG Thresholded Weighted Center of Gravity
USAF United States Air Force
UV Ultra Violet
VFL Variable Focus Lens
VIS Visible
VL Variable Lens
WFS Wavefront Sensor
WFS-AO Wavefront Sensor Adaptive Optics
WFS-less Wavefront Sensorless
WSAO Wavefront Sensorless Adaptive Optics
YFP Yellow Fluorescent Protein
1
Chapter 1. Introduction
1.1. Overview
Vision is an invaluable sense and losing visual function is life changing for many
people. Currently, it is estimated by the Canadian Nation Institute for the Blind (CNIB)
that ~0.5 million Canadians are blind or partially sighted and ~5.6 million Canadians have
an eye disease that could lead to irreversible vision loss [1]. Many of us are or will be
directly affected by the current state of diagnostics and therapies available. Retinal
diseases that disrupt the ability of the eye detect light, including age-related macular
degeneration and diabetic retinopathy, are among the leading causes of vision loss.
Along with most technology, the ability to non-invasively image the retina has
experienced rapid development in the past couple decades. Several imaging modalities
are commercially available and they are used routinely by clinicians for disease diagnosis.
Fundus photography and confocal scanning laser ophthalmoscopy (SLO) provide en face
imaging of the large features of the retina, such as the blood vessels, the optic nerve head,
and the fovea. SLO systems can also include fluorescence detection for alternative
sources of contrast including, autofluorescence imaging from fluorophores that are
intrinsic to the retina, or fluorescein angiography (FA) through the intravenous injection of
fluorescent dye. Optical Coherence Tomography (OCT) provides volumetric and cross-
sectional imaging to visualize the different layers of the retina and OCT-Angiography
provides a non-invasive method for visualizing blood flow. The inclusion of adaptive optics
(AO) with these modalities has provided unprecedented spatial resolution of cellular
structures of the human retina [2]. These technologies provided ophthalmologist and
vision scientists with the ability to track degeneration and evaluate treatments. However,
future imaging tools will allow for non-invasive functional imaging of cellular processes in
the human retina, which can be used to develop novel therapies and to discover earlier
indicators of vision loss [3].
The study of animal models of human diseases is used extensively throughout the
therapy development process. In particular, the mouse has become a very important
2
animal model for many of reasons, including the availability, easy handling, institutional
management, and short reproductive cycles. Furthermore, one of the primary benefits to
using mice is the unparalleled availability of genetically engineered strains for a wide range
of purposes including disease modelling and fluorescence labelling. Conventional ex vivo
immunohistochemistry is often used to study the eye and it provides exquisite cellular
contrast and high cellular resolution of the retina, but only at a single point in time. This
results in studies with large cohorts of animals multiplied by the number of time points that
are needed.
Non-invasive imaging is highly desirable for longitudinal studies for many reasons
including, reducing the effects of inter-animal variation, reducing the number of animals
required for a study, and thereby reducing the development time and the cost of new
therapies [4]. Furthermore, in vivo imaging allows for the study of physiological processes
that are not possible to study ex vivo. With technological advancements, high-resolution
systems have been able to study anatomy and physiology in vivo at the cellular level [2,3].
There would be profound benefits and advancements if more researchers had access to
high-resolution in vivo imaging systems with the functional and structural detection
capabilities that previously were only attainable through histology.
The topic of this thesis is focused on technological development, in order to expand
the ability of in vivo imaging techniques for the mouse retina. The small size of the mouse
eye has consequences for optical imaging. However, it is useful to study because the
mouse retina has similarities to the human retina.
1.2. The mouse eye
The eye is the organ responsible for gathering light and image formation. Much
like a camera, most vertebrate eyes have focusing elements in the front and detecting
elements in the back, as shown in Figure 1.1. The amount of light that enters the eye is
determined by the size of the pupil, which is modulated by the contraction and dilation of
the iris. The light entering the eye is refracted by the cornea and lens to be focused onto
the retina. The human eye has an axial length of 22-25 mm from cornea to retina and a
maximum pupil size of ~9 mm. The mouse eye is ~8 times smaller with an axial length of
~3.3 mm and a maximum pupil size of ~2 mm, which also creates a larger Numerical
Aperture (NA) for light entering the eye. However, the structure and function of the mouse
3
retina has many similarities to the human retina. The thickness of the mouse retina is ~220
µm on average, which is comparable to the human retina at ~250 µm on average [5].
Figure 1.1. Simplified schematic of the mouse eye compared to a human eye [6].
The retina is the tissue at the back of the eye that is responsible for detecting the
light and forwarding the stimuli to the brain for visual processing into images for perception.
The mouse retina is organized in the same way as the human retina with various
differences, for example the absence of a macula for high-acuity vision and the greater
ratio of rod photoreceptors to cones photoreceptors for low light vision [6].
The retina consists of several layers with different types of cells, as shown in Figure
1.2. Starting with the outermost layer in the retina and moving inwards, first there is a
single layer of cells containing melanin called the retina pigment epithelium (RPE), which
is responsible for supporting the light sensing cells. The next layer inwards contain the
photoreceptors, which include rods for high sensitivity to visible light, and cones for colour
vision. The photoreceptors extend inwards with the outer segment (OS), the inner
segment (IS), cell body, and then the synaptic terminals in the outer plexiform layer (OPL).
Bipolar cells extend axially and they can transfer signal directly to the ganglion cell
dendrites in the inner plexiform layer (IPL). Within the inner retina there are also lateral
connections to horizontal cells and amacrine cells, which regulate signals and combine
signals from multiple photoreceptors before transferring signals onwards. The ganglion
cell bodies and the axons form the innermost layers of the retina, which includes ganglion
cell layer (GCL), and then the nerve fiber layer (NFL). The optic nerve, as shown in Figure
4
1.1, exits the retina at the optic nerve head (ONH), which is where the ganglion cell axons
leave of the eye carrying the neural signals to the brain. Vasculature also enters the eye
from the ONH to supply blood to the inner retina and distinct vascular layers are found in
the NFL, IPL, and OPL.
Other types of cells in the retina support the retina, including Müller cells, and
microglia. Müller cells structurally and functionally support the other cells in the retina.
They stretch axial from the outer retina to the inner retina. Microglia can be found in many
of the retinal layers and they perform surveillance tasks, help to maintain homeostasis,
and perform phagocytosis of degenerating retinal neurons. They are highly sensitive to
changes to their micro-environment and can be stimulated into activation by disturbances
such as optic nerve damage, light injury, and disease. Activated microglia change their
morphology from having extended branches to a rounder amoeboid shape, and attempt
to repair the damage [7].
Figure 1.2. Organization of the retinal layers. The image is from Webvision [8] and used under the Creative Commons Licenses.
Light must travel through many retinal layers before detection by the
photoreceptors. The photoreceptors enable the conversion of light into electrochemical
signals through a process called phototransduction. In phototransduction, a photon is
5
absorbed by a visual pigment molecule (photopigment) in the outer segment of the
photoreceptor, which causes a cascade of chemical reactions that results in an
electrochemical potential that can be transmitted through the neural retina to the brain.
After phototransduction, the photopigment must be regenerated before it can absorb
another photon. The regeneration process is carried out by the visual cycle within the outer
segment of the photoreceptor and RPE. In the photoreceptor, the absorption of a photon
causes 11-cis-retinal to be converted to all-trans-retinal and dissociate from the
photopigment. A series of reactions convert all-trans-retinal back to 11-cis-retinal, which
occurs primary in the RPE [9]. The location of each step is summarized in Figure 1.3.
Figure 1.3. The visual cycle and the location of each step. Image is from Wikimedia.org and used here under the Creative Commons License.
6
In the outer retina, the visual cycle and normal cellular metabolism generate
compounds that are also fluorescent. This provides an opportunity to detect these
compounds that are critical to visual function, which enables measurements that could be
used to asses visual function [10–12].
1.3. Imaging the mouse retina
The efficient transmission of light through the eye enables non-invasive optical
imaging of the retina unlike any other internal tissue, but the refractive properties of the
eye must be considered by the imaging system. The anatomical differences in the mouse
eye to the human eye are significant enough that specialized equipment should be
developed for imaging the mouse retina with optimal performance.
Similar to human eye imaging systems, the mouse retina can be imaged with
modalities such as fundus photography, scanning laser ophthalmoscopy (SLO), and
optical coherence tomography (OCT). Fundus photography detects scattered light from
the retina on a two-dimensional pixel array. Alternatively, techniques such as SLO and
OCT, require light to be scanned across the retina and interrogate each point individually
for each image pixel. The diffraction-limited lateral resolution in the sample is related to
the smallest spot size of the focused light on the sample, which is dependent on the
numerical aperture (NA) into the eye. The NA is the half angle of the cone of light being
focused to a point, given by the Equation 1.1:
𝑁𝐴 = 𝑛 sin 𝜃, (1.1)
where 𝑛 is the index of refraction and 𝜃 is the angle from the optical axis. An advantage
of the point scanning system is that a confocal pinhole can be placed in the detection path
to remove light scattered from outside the focal plane, providing an axial sectioning ability
within biological specimen. For OCT, the lateral resolution is determined in a similar way
as SLO, but the axial resolution is not dependent on the focal spot size. Instead, the axial
resolution for OCT is determined by the spectral bandwidth. Therefore, OCT typically
provides superior axial resolution for cross-sectional imaging. These modalities are further
explained in the next chapter.
Non-invasive optical imaging of the mouse eye is enticing due to the large
Numerical Aperture (NA) available that would permit a small focal spot on the retina.
7
Theoretically, the geometry of the mouse eye allows for sub-micrometer resolution
imaging. There are two options for focusing light onto the mouse retina, as shown in Figure
1.4. Focusing light onto the retina can be achieved with an objective lens, like in the
traditional confocal microscope. In this configuration, the refraction caused by the cornea
must be canceled out with plano-concave lens or ‘fundus lens’ that matches the curvature
of the cornea. Alternatively, the focusing ability of the eye can be used by the imaging
system if collimated light is directed into the mouse eye. The benefits of each method are
discussed in Chapter 3. For either method, the NA is related to the diameter of the beam
across the eye upon entry. However, for imaging with a large NA, optical aberrations are
introduced by the refractive elements, including the biological tissues in the optical path to
the retina. Optical aberrations from the tear film, cornea, and lens of the mouse enlarge
the size of the spot on the retina, reducing the actual imaging resolution.
Figure 1.4. The options for focusing light onto the mouse retina: collimated light or focused light from an objective lens.
Aberrations from the mouse eye can be corrected with adaptive optics (AO), in
order to obtain diffraction-limited imaging. The goal of the AO within an imaging system is
to restore the smallest possible spot on the sample, which has been enlarged by
aberrations. Since every mouse will have different errors in the focusing system, this will
result a variety of aberrations for a given area on the sample, known as the isoplanatic
patch. Therefore, AO systems need a method to correct the wavefront, as well as a
method of determining the aberrations to be corrected. To enable cellular resolved
imaging in the mouse retina, there are high-orders of aberrations that must be corrected
for optimal performance.
8
Conventional AO employs a method to directly measure the aberrations from the
sample, and a corrective element to restore optical performance. Alternatively, the images
from the system can be used to indirectly determine the optimal aberration correction
required. These image-based methods are usually referred to as Sensorless Adaptive
Optics (SAO).
Performing accurate wavefront measurements for AO imaging in a small animal
retina typically requires a high level of system complexity due to the short length of the
eye creating an optically thick sample with multiple scattering surfaces that can confound
the wavefront measurement [5]. Although despite the difficulty in performing good
measurements in the mouse eye, retinal imaging system that use SH-WFS AO have been
reported in the Literature with high quality state-of-the-art performance, such as in
references [4,13,14]. Sensorless AO (SAO) imaging methods developed in this thesis can
avoid the complexities of WFS measurements at the cost of execution time. SAO allows
for AO retinal imaging with systems that are compact, easily operated, and robust.
1.4. Outline
The remaining chapters of this thesis are organized as follows. Chapter 2 presents
background information on the imaging modalities and methods for adaptive optics used
in this thesis. Chapter 3 details a preliminary optical design for imaging the mouse retina,
where we used focused light into the eye with a fundus lens to cancel the refraction from
the cornea. This chapter demonstrates the ability of image-based adaptive optics for
fluorescence imaging of fluorescently labelled cells. Chapter 4 demonstrates a novel
method for image-based adaptive optics that was developed for imaging the retina, which
was based on a technique called ‘pupil segmentation’. In this chapter, we also transition
the optical design to use collimated light into the retina. In Chapter 5, wavefront
measurements from the mouse eye show that the image-based approaches to AO are
indeed providing aberration correction in the mouse eye. Chapter 6 demonstrates
significant improvements to the image quality with a novel optical design in a compact
form factor suitable for translational research. Chapter 7 shows that the optical techniques
developed in the previous chapters can be applied to a Two-Photon Excited Fluorescence
(TPEF) system for high resolution imaging of both endogenous and exogenous
fluorophores in the retina. Finally, Chapter 8 discusses future research directions to
advance imaging technology for the mouse retina.
9
1.5. Contributions
The over-arching goal of my work has been to develop advanced imaging
technology that has the potential to be translated to scientists that are not specialized in
Adaptive Optics (AO), which has required that AO imaging systems be easy-to-use,
robust, and compact. My first project demonstrated the performance of a lens-based
optical design with SAO to provide aberration correction for fluorescence imaging, which
is described Chapter 3 and published by Biomedical Optics Express in reference [15]. This
work was extending into a compact design, which is described in Chapter 6 and published
by Biomedical Optics Express in reference [16]. The compact form factor imaging system
also included SAO-OCT for improved multi-modal functionality. The results in this work
demonstrated that SAO could provide imaging resolution that are comparable to traditional
wavefront sensing methods in the mouse eye. Also, my work with vision science
collaborators has resulted in measurements that helped explain underlying mechanisms
of a retinal disease in a mouse model, described in reference [17].
While working on these optical systems, I also improved the image-based SAO
techniques. The ability of SAO to correct for aberrations in the mouse eye was investigated
using a wavefront sensor and the final image quality to evaluate the performance. This
investigation is described in Chapter 5, which was published in Biomedical Optics Express
[18]. I also contributed to further investigations on AO for the mouse eye using a contact
lens, which is described in reference [19]. I demonstrated the proof-of-principle of a novel
image-based AO technique for retinal imaging using pupil segmentation, which is
described in Chapter 4 and published by Optics Letters in reference [20]. My contributions
to the SAO methods were used in human imaging systems as well, which are described
in references [21,22].
Using the methods developed in previous projects, I contributed to the success of
the TPEF and the visible-light OCT imaging systems with my optical designs and imaging
experiments, which are described in references [23,24]. Finally, I also incorporated these
methods into an improved TPEF imaging system capable of imaging the RPE mosaic of
the mouse retina. The TPEF imaging system is described in Chapter 7 and published in
Biomedical Optics Express [25].
10
Chapter 2. Background on retinal imaging systems and adaptive optics
2.1. Scanning Laser Ophthalmoscopy
Scanning Laser Ophthalmoscopy (SLO) is an imaging tool for an en face view of
a living eye, which is based on confocal scanning laser microscopy. However, the eye is
used as the objective lens to focus the light onto the retina. For SLO, the illumination light
is focused onto a single point on the sample and the back-scattered light returns to a
detector to be measured. The focused point is scanned across the retina to perform a
measurement for each pixel in the image. Light can be scanned across the sample with
different types of scanners, including galvanometer mirrors, resonant scanners, and
MEMS-based scanners, and usually video frame rates are achievable.
An advantage of sampling each point in time is that a pinhole can be optically
conjugated to the sample plane, which will remove light that is scattering from out of the
focal plane and improve the axial sectioning ability of the imaging system. A pinhole in this
configuration can be called a confocal aperture. Also sampling each pixel in time allows
for SLO systems to use highly sensitive detectors such as photo-multiplier tubes and
avalanche photodiodes that have a single detector. Similar to conventional microscopy,
optical filters can be used to isolate the back-scattered light from fluorescence emission
from the sample, further described in Section 2.3.
In order to define the diffraction-limited resolution of SLO, a point illumination
imaging system can be characterized by the point spread function (PSF). Consider an
ideal system where the objective lens is illuminated with a circular aperture of uniform
intensity. The light will focus into the Airy pattern, which will be the PSF or impulse
response (h) of the system. At a given NA and wavelength (λ), the radius to the first
minimum of the Airy pattern in the lateral direction at the focal plane is given by Equation
2.1:
𝑟𝐴𝑖𝑟𝑦 =
0.61 𝜆
𝑁𝐴 . (2.1)
11
In the axial direction, the distance from the center of the diffraction pattern will have a first
minimum (𝑧𝑚𝑖𝑛) at a distance given by Equation 2.2:
𝑧𝑚𝑖𝑛 =
2 𝑛 𝜆
𝑁𝐴2 , (2.2)
where 𝑛 is the index of refraction of the medium [26,27] . The FWHM of the axial PSF can
be calculated by a multiplying the half width (𝑧𝑚𝑖𝑛 ) by a factor of ~.84 [26]. Using a
confocal aperture that is close to the size of the Airy radius can provide an additional
improvement to the lateral imaging resolution. However, in this thesis, a confocal aperture
several times the size of the Airy disk was used in order to balance signal with depth
sectioning. Therefore, the non-confocal PSF calculations were used to approximate the
theoretical diffraction-limited resolutions.
The final intensity distribution, 𝐼, that is measured by the detector of the SLO
system will be a convolution of the PSF of the system, ℎ, with the pinhole (circle function)
[26,28], described by Equation 2.3:
𝐼 = |ℎ|2 (|ℎ|2 ⊗ 𝐶𝑖𝑟𝑐) (2.3)
In a real imaging situation, the sample or system will introduce aberrations that
must be corrected to restore diffraction-limited resolution, which is further discussed in
Section 2.4.
2.2. Optical Coherence Tomography
Optical Coherence Tomography (OCT) provides high-resolution cross-sectional
and volumetric imaging of the retina. OCT was initially developed as a tool for
ophthalmology to better visualize the layers of the retina but also has been adopted by
other areas of biomedical imaging. Other advancements in OCT systems have enabled
polarization sensitivity [29], as well as functionally imaging of blood flow often called OCT-
Angiography (OCT-A). OCT-A is generated by analysing changes to the cross-sectional
images that are caused by the blood moving through the vessels [30–33].
OCT volumes are generated by acquiring adjacent depth intensity profiles (A-
scans) to generate a cross-section (B-Scan) into the sample. Then, B-scans can be
acquired adjacently in the remaining dimension for volumetric imaging. The OCT volumes
12
typically have comparable resolution in both the axial and lateral directions, therefore en
face views can be generated at any depth within the sample.
Fourier domain OCT can be separated into two types: Spectral Domain (SD) OCT
and Swept Source (SS) OCT. Both types are used in modern systems and both types
share the same principle based on the Michelson interferometer using a low-coherence
light source. The difference between each type is that SD-OCT relies on a spectrometer
to measure the interference pattern on an array of detectors, and SS-OCT uses a single
detector and a laser that sweeps through a spectral bandwidth. The typical bandwidth of
an OCT system for imaging the retina is 50 nm to 100 nm, commonly with a near infrared
center wavelength.
For an OCT system, the imaging light is separated into a reference path and the
sample path, and then recombined at a fiber coupler. The interference pattern modulations
correspond to the path mismatch between the reference arm and the sample arm where,
higher frequency fringes correspond to a larger mismatch (∆𝑧). Therefore, by calculating
the Fourier transform of the fringes, the axial location of light that is scattered from the
sample can be determined. For example, consider a single reflector at position ∆𝑧. The
measured intensity on the detector (𝐼𝐷) will be a function of wavenumber (𝑘) from the
interference pattern, as described by Equation 2.5:
𝐼𝐷(𝑘) = 𝑆(𝑘){[𝐼𝑅 + 𝐼𝑆] + 2√𝐼𝑅𝐼 cos(2∆𝑧𝑘)}, (2.4)
where 𝐼𝑅 is the reference light, 𝐼𝑆 is the sample light, and 𝑆(𝑘) is the intensity of the source
spectrum. After the Fourier transform, the location of the reflector, ∆𝑧, is revealed as delta
functions convolved with the Fourier transform of the source spectrum, �̂�(𝑘), in Equation
2.6.
𝐼(𝑧) ∝ �̂�(𝑘) ⊗ {[𝐼𝑅 + 𝐼𝑆]𝛿(𝑧) + 2√𝐼𝑅𝐼𝑠(𝛿(𝑧 − ∆𝑧) + 𝛿(𝑧 + ∆𝑧))}. (2.5)
The axial resolution of an OCT system is determined by the coherence length (𝑙𝑐), given
by Equation 2.7:
𝑙𝑐 =
2 ln 2
𝜋∙
𝜆𝑜2
∆𝜆, (2.6)
13
which is dependent on the center wavelength of the light source, 𝜆𝑜, and the spectral
bandwidth, ∆𝜆. OCT typically provides a 1 µm - 10 µm axial resolution, with a depth
penetration on the order of millimeters. The lateral resolution is dependent on the focal
spot size similar to traditional microscopy, as described in the SLO section.
OCT systems have been increasing in speed with advancement in the detectors
and lasers available. SD OCT systems are often limited by the speed the camera can
acquire A-scans. Similarly, SS OCT systems are often limited by the speed that the laser
can sweep the spectral bandwidth. For real-time processing and display, OCT images
often require GPU accelerated processing programs to keep up with the A-scan
acquisition rate.
2.3. Fluorescence imaging
Fluorescence imaging is an essential tool that has been used in microscopy to
provide contrast that may not be available by other methods. Fluorophores are often
added to the biological specimen to identify cells, blood flow or specific molecules. Green
Fluorescent Protein (GFP) and its derivatives have been used extensively to label selected
cell classes in a variety of organisms, including mice. Labeling otherwise transparent cells
with fluorescent compounds allows them to be imaged with fluorescence detection. Also,
many biological samples have intrinsic fluorescent properties or autofluorescence that can
be imaged to provide insights into the health of the tissue [34–36].
The process, known as fluorescence, starts when a fluorophore is excited by the
absorption of a photon. Then, some of the energy gained by the initial photon is lost
through non-radiative processes, and finally a photon with a longer wavelength is released
from the fluorophore. This chromatic difference, known as the Stokes shift, can be used
to isolate the fluorescence emission from the fluorescence excitation. Various
fluorophores have different excitation and emission spectra.
Alternatively, fluorescence emission can occur from a process called Two-Photon
Excited Fluorescence (TPEF). The principle of TPEF is that two photons can provide
sufficient energy as a single photon to excite the same fluorophore, as shown in Figure
2.1. This requires that the pair of lower energy photons arrive at the fluorophore at
14
practically the same time. Therefore, TPEF imaging uses ultrashort pulses of light to
improve the probability that two-photon absorption occurs.
Figure 2.1. Simplified Jablonski diagram of single photon excited fluorescence and two-photon excited fluorescence. The excitation light, λex, and emission light, λem.
Two-Photon Excited Fluorescence (TPEF) can be used as an imaging technique
that enables fluorescence imaging deeper into tissue than the equivalent single photon
excitation [37,38]. Often TPEF uses near infrared light to excited the same fluorophores
that require UV or visible light for excitation, which has advantages for imaging the retina
that are further discussed in Chapter 7. For TPEF imaging, signal can be improved by
optimizing the probability of two-photon absorption. The number of photons absorbed (𝑛)
is related to the intensity of the fluorescence emission, which is described by Equation 2.8
[38]:
𝑛 ≈𝑃2𝜎
𝜏𝑝𝑓𝑝2 (
𝑁𝐴2
ℎ𝑐 𝜆)
2
. (2.7)
Factors that affect the absorption include, the average incident power (P), the cross-
section of the fluorophore (𝜎), the laser pulse width (𝜏𝑝), the laser repetition rate (𝑓𝑝), and
the NA of the incident light (𝜆).
The axial and lateral resolution of fluorescence imaging in point scanning systems
are related to the size of the spot on the sample. For this thesis, we will use the
conventional spot size given by the PSF in order to determine the theoretical resolution of
the systems, as described in Section 2.1.
15
2.4. Adaptive optics for ophthalmic imaging
AO technology is used to improve the performance of many types of optical
systems. Originally, AO was developed for astronomy to dynamically correct for
aberrations caused by the transmission of light through atmospheric turbulence [39]. Later,
AO techniques started to be used to correct for aberrations caused by imperfections
through biological specimens, such as imaging of the brain. Naturally, AO was also
developed for retinal imaging to correct focusing errors from many different eyes, including
humans, primates, and rodents. AO has been used for imaging the eye with many
modalities, including fundus photography, SLO, and OCT [2]. An AO system consists of a
device to correct the optical performance, and a method for determining the correction
required for the sample.
There are a wide variety of wavefront correctors commercially available, such as
deformable mirrors. The deformable mirror can change shape to modify the incident
wavefront. In this thesis, a segmented DM and a continuous membrane DM are both used.
The segmented DM offers quick MEMS based actuators, and low flatness of the reflection
surface, <20 nm. The continuous membrane DM provides large stroke with the magnetic
actuators, ~80 µm (peak to valley) and low settle time, ~0.5 ms. For the AO systems in
this thesis, it is ideal for the deformable mirrors to operate quickly and accurately with little
hysteresis and drift in the actuators position over time.
AO systems for imaging the retina were developed using a Shack-Hartmann (SH)
wavefront sensor (WFS) to directly measure the aberrations [40–42]. A SH-WFS is
constructed by an array of identical micro lenses, often called lenslets, that are mounted
a focal length away from a 2D detector. For measuring a wavefront, the lenslets should
be positioned at a pupil plane and the wavefront will form an array of spots on the detector.
Each lenslet of the WFS will gather a sample of the wavefront. If a measured wavefront
has no aberrations, the spots on the detector will form an even grid, corresponding to
centers of the lenslets in the array. However, a local gradient in the wavefront will cause
a displacement in the spot on the sensor, where magnitude and direction can be used to
reconstruct the aberrations in the pupil plane.
16
A pupil plane of a system is related to the PSF by the Fourier transform, which
means that modulations due to aberrations in the pupil plane will enlarge the size of the
focal spot on the sample [43]. However, if there are no aberrations in the pupil plane, then
the system is only limited by diffraction. Wavefront aberrations are commonly described
by a decomposition into the Zernike polynomials [39]. The aberrations in the pupil
plane, 𝑊(𝜌, 𝜃), can be described by Equation 2.9:
𝑊(𝜌, 𝜃) = ∑ 𝑎𝑗𝑍𝑗(𝜌, 𝜃)
𝑗=0
, (2.8)
where 𝑍𝑗 is a Zernike polynomial with a given coefficient, 𝑎𝑗. Table 1 lists the Zernike
polynomials with the corresponding index (j) that is used for this thesis.
Table 2.1. Zernike polynomials, names and index up to the 5th radial order.
Index (j) Radial order (n) Aberration term Zernike Polynomial
𝒁𝒋(𝝆, 𝜽)
0 0 Piston 1 1 1 Tilt 2𝜌 sin 𝜃 2 1 Tip 2𝜌 cos 𝜃 3 2 Oblique astigmatism √6𝜌2 sin 2𝜃 4 2 Defocus √3(2𝜌2 − 1) 5 2 Vertical astigmatism √6𝜌2 cos 2𝜃 6 3 Vertical trefoil √8𝜌3 sin 3𝜃 7 3 Vertical coma √8(3𝜌3 − 2𝜌) sin 𝜃 8 3 Horizontal coma √8(3𝜌3 − 2𝜌) cos 𝜃 9 3 Oblique trefoil √8𝜌3 cos 3𝜃 10 4 Oblique quadrafoil √10𝜌4 sin 4𝜃 11 4 Oblique secondary
astigmatism √10(4𝜌4 − 3𝜌2) sin 2𝜃
12 4 Primary spherical √5(6𝜌4 − 6𝜌2 + 1) 13 4 Vertical secondary
astigmatism √10(4𝜌4 − 3𝜌2) cos 2𝜃
14 4 Vertical quadrafoil √10𝜌4 cos 4𝜃 15 5 Higher orders √12𝜌5 sin 5𝜃 16 5 √12(5𝜌5 − 4𝜌3) sin 3𝜃 17 5 √12(10𝜌5 − 12𝜌3 + 3𝜌) sin 𝜃 18 5 √12(10𝜌5 − 12𝜌3 + 3𝜌) cos 𝜃 19 5 √12(5𝜌5 − 4𝜌3) cos 3𝜃 20 5 √12𝜌5 cos 5𝜃
A basic SH-WFS AO system will illuminate a spot on the retina and measure the
optical wavefront that returns to the SH-WFS. The wavefront corrector will attempt to
17
remove the aberrations with the complex conjugate of the measured wavefront. This
process will repeat in what is referred to as closed-feedback loop, until the measured
wavefront aberrations are reduced to an acceptable flatness. The WFS AO will require
optical elements to re-direct the scattered light from the sample to the WFS in a different
optical path from the illumination and detection components.
In confocal microscopy, AO can be used for imaging into volumetric biological
samples. The ideal configuration is to measure the wavefront from the focal plane that is
being imaged by the system. Confocal imaging systems rely on a pinhole to reject light
from out of the focal planes in order to image a specific depth in the sample. However,
most wavefront sensors do not have the ability to reject out of focus light. Therefore, back-
scattered light from multiple planes in the sample can corrupt the measurement from the
imaging plane. However, an alternative to using a direct measurement from the sample is
to use the images from the system to indirectly infer the aberrations, thereby using the
confocal pinhole to provide depth discrimination.
As described in Section 1.3, the short focal length of the mouse eye and a large
NA illumination create multiple scattering surfaces from the volumetric layering of retina.
In this thesis, SAO methods were developed to determine optimal aberration correction
for the chosen focal plane in the retina.
A common image-based AO method operates by applying aberrations to the
wavefront corrector and recording the effect on the image, which is then used to determine
the best correction. This is called open-loop control. The image quality due the aberration
on the corrective element can be quantified with a sharpness or brightness metric. Then,
the image metric values can be used as the merit function for an optimization problem that
uses a chosen number of degrees of freedom (Zernike modes) to find the optimal image.
A few versions of the hill-climbing coordinate search optimization algorithm are presented
in each chapter. Section 5.2.3 and Section 5.2.2 provide be best descriptions for each
case.
Another type of image-based AO investigated in this thesis uses a computational
algorithm based on pupil segmentation [44–46]. A pupil plane can be divided into sub-
regions by an active element, such as a segmented deformable mirror or spatial light
modulator. If only one sub-region illuminates the sample, a smaller diameter beam or
18
‘beamlet’ will illuminate a point in the sample. An image can be acquired using the beamlet
in the center of the pupil, which will be defined as the reference. A sample with aberrations
will cause the beamlets from the other pupil regions will be deflected from the reference
focal point and the images formed will be shifted from the reference image. The amount
of translation between the images from each sub-region of the pupil and the reference can
be related to the aberrations in the entire pupil.
The translated distance, ∆𝑥, ∆𝑦, of the images can be used as wavefront slopes,
which can be reconstructed into Zernike coefficients that approximate the wavefront in a
similar way to how a wavefront sensor would perform the calculation. The local slope of
the wavefront in the x and y direction is related the Zernike polynomials, Z, by Equation
2.10 and 2.11.
∆𝑥
𝑓=
𝜕𝑊
𝜕𝑥= ∑ 𝑎𝑗
𝑗
𝜕𝑍
𝜕𝑥 , (2.9)
∆𝑦
𝑓=
𝜕𝑊
𝜕𝑦= ∑ 𝑎𝑗
𝑗
𝜕𝑍
𝜕𝑦 , (2.10)
where the partial derivatives of the wavefront, W, can be calculated on the Zernike
polynomials, Z, to the jth term. Using this relationship, a conversions matrix, Z, can be
constructed to calculate the slope of the wavefront at n positions of the pupil, 𝐬 =
[∆𝑥1, ∆𝑦1 … , ∆𝑥𝑛, ∆𝑦𝑛 ]𝑇, for a given vector of Zernike coefficients, a = [𝑎0 , … , 𝑎𝑗]𝑇 ,
described by Equation 2.12:
𝐬 = 𝐙𝐚 , (2.11)
Then, any wavefront can be approximated by a vector of Zernike coefficients, 𝐚𝒍𝒔, for the
measured slope values, s, by a least-squared fit in Equation 2.13:
𝐚𝒍𝒔 = 𝐙ϯ𝐬 . (2.12)
The active element that divides the imaging pupil can be a separate device than
the wavefront corrector. As further described in Chapter 4, we use the segmented
deformable mirror for both aberration correction and pupil segmenting.
19
2.5. Summary
This chapter presented a background to the imaging modalities SLO, OCT, and
TPEF that are used in this thesis. Also, the background information for AO used in
ophthalmoscopy was presented, including an introduction to SAO with optimization
algorithms and with pupil segmentation. The remaining chapters utilize these topics and
provide additional depth.
As a note, in the remainder of this thesis, various acronyms were used for the
abbreviation of “Sensorless Adaptive Optics” and “Wavefront Sensorless Adaptive
Optics”. All of the acronyms are listed in the directory and SAO, WSAO, and WFS-less
can be used interchangeably.
20
Chapter 3. Wavefront sensorless adaptive optics fluorescence biomicroscope for in vivo retinal imaging in mice
3.1. Introduction
Small animal models of diseases are a vital component in vision research because
they facilitate the understanding of underlying biological processes, the identification of
potential causative genes for human disorders, and the development of therapies against
vision-robbing diseases. Mice are commonly used for preclinical vision research due to
the significant anatomical and functional similarity of their eyes to human eyes and to the
availability of transgenic strains that model human diseases. Non-invasive in vivo retinal
imaging has the potential to reduce the number of animals required for a study, which in
turn reduces the development time and the cost of new therapies [4]. Transgenic mice
expressing endogenous fluorescent markers, such as Enhanced Green Fluorescent
Protein (EGFP), are particularly important for vision research. The ability to image
molecular markers has the potential to accelerate vision research by allowing retinal
function to be observed in vivo and by permitting longitudinal studies of the same animal
[13]. Research animals expressing EGFP in neuronal cells, including retinal ganglion cells
and axons, are useful for studying retinal neurodegenerative diseases such as glaucoma
[47]. Similarly, mice with EGFP-labelled microglia enable the in vivo study of the retinal
response to diabetic retinopathy, glaucoma, and age-related macular degeneration [48–
50]. Additional examples of vision research based on fluorescence imaging are described
in Zhang et al. [49], and Alt et al. [50], just to name a few. Non-invasive fluorescence
imaging of the mouse retina with even higher resolution is desirable, but requires
correction of optical aberrations in the mouse eye [5].
Adaptive optics (AO) for ophthalmoscopy is an important tool for ophthalmologists
and vision scientists, permitting cellular-resolution imaging of the retina. For non-invasive
in vivo imaging in humans, AO has been demonstrated to improve the resolution for fundus
photography, Scanning Laser Ophthalmoscopy (SLO), and Optical Coherence
Tomography (OCT); there are several reviews on this topic, including [2,39,51–53]. The
geometry of the mouse eye allows for focusing light on the retina with a higher numerical
21
aperture (NA) than in humans, which allows for even higher resolution imaging in vivo.
Due to the higher NA, the mouse eye is even more sensitive to aberrations induced when
imaging with a large diameter beam, with aberrations introduced from the ocular tissues
(i.e. cornea, lens and vitreous humour) [5]. AO facilitates retinal imaging with diffraction-
limited performance at cellular resolution in mice.
Conventional AO compensates for aberrations in the wavefront with an adaptive
element such as a deformable mirror (DM) controlled by a Shack-Hartmann (SH)
wavefront sensor (WFS) in a closed feedback loop. The common approach of wavefront
sensing is to use an extra light source (beacon) with low NA to measure the wavefront
aberrations with the SH-WFS [4,14,53]. However, the beacon adds to the limited light
power that is allowed into the eye and contributes to non-common path errors [54]. The
ability of the WFS-based AO system to correct wavefront aberrations can be limited by the
WFS design (its accuracy and dynamic range) and a geometry mismatch between the
WFS and the adaptable element, leading to wavefront correction errors [55]. The SH-WFS
performance is also susceptible to specular reflection from lenses and optical elements
within the system [56]. Many of these issues have been addressed in different ways with
AO system modifications. For example, efficient spherical mirror-based telescopes can be
implemented to maintain signal and minimize back-reflections into the SH-WFS [57]. More
compact AO systems have been developed with lens-based optics and polarizing
elements to reduce these back-reflections [56]. For applications in vision science and
small animal retinal imaging, a SH-WFS is further hampered by the ‘small eye artifact’, in
which multiple reflecting/backscattering surfaces in the retina affect the wavefront
measurement [5].
Some of the challenges associated with the SH-WFS can be better managed
through Wavefront Sensorless Adaptive Optics (WSAO), which uses information from the
quality of the image to guide aberration correction. Wavefront sensorless techniques have
been applied in microscopy [58] and ophthalmic imaging for both human and mouse eyes
[59–61].
In this chapter, we expand on previous work combining WSAO with Fourier-
domain OCT for mouse retinal imaging, and present fluorescence images of mouse retina.
AO fluorescence retinal imaging in mice has been presented with different modalities and
methods throughout the Literature [4,13,14,62–65], including both with and without a
22
WFS. Notably, WFS-AO for in vivo subcellular-resolution imaging has been presented
using an annular beam for the beacon, a high-resolution SH-WFS, and a mirror-based
optical system [4,13,14]. We have implemented a lens-based WSAO system with a modal
hill-climbing optimization algorithm using fluorescent image intensity as a metric. We used
the combination of a variable focus lens and a small-stroke MEMS
(microelectromechanical system) deformable mirror to perform defocus and aberration
correction in the mouse eyes. The details of the experimental methods are described in
the next section. The system performance was evaluated on phantoms, and
representative images acquired from mouse retina in vivo are presented. Our results
demonstrate that AO can be simple and compact with cellular resolution.
3.2. Methods
We have implemented a compact and low-cost confocal biomicroscope in order to
acquire both reflectance (structural) and fluorescence (functional) images from mouse
retina simultaneously. The system used off-the-shelf lenses with a smaller footprint and
simpler design compared to the AO configurations that are based on SH-WFS and
constructed from curved mirrors. A benefit of the lens-based system is that it permits a
relatively wide field of view (FOV) to be imaged on the mouse retina [56,66], albeit without
diffraction-limited imaging performance outside of the isoplanatic patch. The wide field
structural image was used in real-time to navigate on the retina using features such as
blood vessels and the optic nerve head. Once centered at the desired location using a
FOV of about 0.8 mm, the FOV was reduced to 0.2 mm or smaller for acquisition of the
high-resolution fluorescence images that are presented below.
During imaging, the mouse was aligned to the imaging system using a plano-
concave ‘fundus lens,’ with a 2 mm contact diameter and no magnification (Volk Optical
Inc, Mentor, OH) that canceled out most of the refractive power from the mouse cornea;
we utilized an external objective to focus the light on the mouse retina. For this reason,
we refer to the imaging system as a fluorescence confocal (f/c) biomicroscope rather than
an ophthalmoscope. The fundus lens approach to retinal imaging has several benefits: it
facilitates alignment of the mouse, provides mechanical stability, and retains the moisture
of the mouse cornea during imaging.
23
3.2.1. Mouse handling
The two strains of EGFP-labelled mice, B6.Cg-Tg(Thy1-EGFP)MJrs/J (ganglion
cells) and B6.129P-Cx3cr1tm1Litt/J (microglia cells), that were imaged in this report were
obtained from Jackson Laboratories (Bar Harbor, ME). The mouse imaging sessions were
performed under protocols compliant to the Canadian Council on Animal Care, and with
the approval of the University Animal Care Committee at Simon Fraser University. Prior
to the imaging experiment, the mice were anesthetized with a subcutaneous injection of
ketamine (100 mg/kg of body weight) and dexmedetomidine (0.1 mg/kg of body weight).
Next, the eyes were dilated with a drop of topical solution (Tropicamide, 1%) and a couple
minutes later, a drop of topical anesthetic (Alcaine, 0.5%) was applied. Artificial tear gel
(Alcon, Fort Worth, TX) was applied liberally to protect the cornea from dehydration. The
anesthetized mouse was placed on a translation stage and the eye was gently aligned
with direct contact to the fundus lens. The laser power at the fundus lens was ~150 µW.
The mice were recovered after the experiment using atipamezole injected at 1.8 mg/kg of
body weight. The local anesthetic was applied to reduce potential irritation to the mouse
after recovery due to the contact with the fundus lens.
3.2.2. Biomicroscope optical setup
The optical system schematic is presented in Figure 3.1. We used a fiber-coupled
Ar/Kr ion laser and a diffraction grating to select the 488 nm spectral line as the excitation
source. At the fiber output, the beam was collimated to be approximately 3.5 mm in
diameter. The first element was the segmented MEMS deformable mirror (DM) with modal
control and a 5 µm stroke (PTT111, Iris AO, Inc, Berkeley, CA), which also defined the first
pupil plane. Two lenses, f1 = 200 mm and f2 = 200 mm, relayed the conjugate plane to the
tunable liquid lens (Varioptics, Arctic 39N0) for focus control and for optical sectioning of
the different retinal layers. Relay lenses with focal lengths of f3 = 150 mm, and f4 = 100
mm were used to decrease the beam’s diameter and to relay the pupil to the scanning
mirrors. The beam position was scanned over the sample with two 6210H galvanometer
mounted mirrors (Cambridge Technology Inc.) for each direction. Next, two lenses, f5 = 25
mm, and f6 = 75 mm expanded the beam’s diameter and relayed the pupil to the final
objective lens. The beam was focused with a NA = 0.17 to a final spot (Airy disk radius) of
1.8 µm and the FWHM of the axial point spread function was 38 µm with a Mitutoyo infinity-
corrected long working distance 10x objective lens. The backscattered light (reflectance)
24
was coupled into a 100 µm multimode fiber (3.9 Airy disks in diameter) and detected by
an Avalanche Photo Diode (APD, Hamamatsu S5343). We also used a quarter-wave plate
(QWP) to act on the linearly polarized light from the laser and a linear analyzer at the
detector to reduce the specular reflection from the optical elements [56]. The backward-
directed fluorescence signal was isolated by a dichroic mirror and a long-pass filter
(Semrock 496 nm blocking edge BrightLine, FF01-496/LP-25). The fluorescence light was
focused by an 80 mm lens through a 50 µm confocal pinhole with diameter 1.8 times the
size of the Airy disk, and a Photo-Multiplier Tube (PMT, Hamamatsu H7827-002) was used
to detect the weak fluorescence signal. An analog-to-digital converter digitized the signal
from the APD and PMT simultaneously, and the imaging system speed was limited to 1.00
Mega-Samples Per Second (MSPS) per channel by the NI PCIe-6361. The galvanometers
were driven by a 1 kHz sinusoidal waveform scanning pattern, and acquired data from
both forward and backward sweeps of the scan. Image distortion caused by the non-linear
scanning pattern was corrected with de-warping in real-time using custom software
developed in C/C++ for acquisition and display. This allowed for 400x400 reflectance and
fluorescence samples per frame at 5 fps, which was used for aligning the mouse eye and
for data streaming. The AO optimization was performed at 400x100 samples which
corresponded to an acquisition and display rate of 20 fps. The icon of a computer in Fig.
3.1. has representative images for each channel displayed during imaging.
25
Figure 3.1. Schematic of the WSAO f/c biomicroscope using 488 nm excitation from an Ar/Kr laser. Relay lenses are achromatic doublets. Other optical elements: 80/20 beam splitter (BS), dichroic mirror (DC), deformable mirror (DM), zero-order quarter wave plate (QWP), objective lens (OBJ), linear polarizer (LP), pinhole (PH), variable lens (VL), galvanometer scanning mirrors (GM). Electronic elements: avalanche photo diode (APD), photo multiplier tube (PMT). The images on the computer icon are representative images of the structural and fluorescence imaging channels.
3.2.3. Image acquisition and optimization
The anesthetized mouse was placed in front of the fundus lens to initiate imaging.
The retinal imaging location was determined based on landmarks such as the vascular
pattern and the optic nerve head. The maximum FOV on the mouse retina was ~0.8 mm
with the mouse cornea approximately perpendicular to optical axis of the system. The
position of the fundus lens was fixed in the center of the optical path, and different
eccentricities on the mouse retina were imaged by rotating the orientation of the mouse
with respect to the fundus lens. The focus was adjusted using the Varioptic lens in order
to get the best qualitative image appearance [67].
The WSAO optimization algorithm that we used is a modified version from our
previous report for WSAO mouse imaging with OCT [61]. During optimization, the frame
rate of the WSAO f/c biomicroscope presented in this report was 20 fps with a frame size
26
of 400x100 samples. Following optimization, higher quality images were acquired with
400x400 samples per frame at 5 fps. The optimization algorithm used a modal control to
build up an optimal shape of the DM that corrected for the wavefront aberrations. For each
Zernike mode, the algorithm searched through 21 coefficient values; the search range for
a particular Zernike mode was selected from typical aberration amplitudes for mouse eyes.
A larger range was searched for lower order aberrations and a smaller range for higher
order aberrations. We were able to correct for up to the 20th Zernike mode (OSA
convention [68]) in ~30 seconds.
The WSAO used a hill-climbing search algorithm to find the best set of Zernike
coefficients that corrected for aberrations based on the image quality metric. The overall
intensity of an image was calculated as the sum of each pixel, defining the merit function
J(k) as in Equation 3.1:
𝐽(𝐤) = ∑ Iw(𝐤)(𝑥, 𝑦),
𝑥,𝑦
(3.1)
where k is a vector of Zernike coefficients and Iw(k) is the acquired intensity if the pixel at
the image coordinates x, y. The wavefront shape, w(k), applied to the DM is given by
Equation. 3.2:
𝑤(𝐤) = ∑ 𝑘𝑛𝑍𝑛
20
𝑛=3
. (3.2)
The flowchart in Fig. 3.2 is a summary of the optimization process. The algorithm
applied a linearly spaced range of coefficients (kn) for each Zernike mode (Zn) to the DM
and recorded an image for each pre-set coefficient value. The optimal image, and thus the
optimal coefficient value, was determined as the one that corresponded to the highest
value of the merit function. The optimization began with defocus (n = 4) [68] and first three
Zernike modes (piston, tip and tilt) were assigned a coefficient value of zero and not
included as part of the optimization. The Zernike modes were optimized in the order as
presented on the abscissa of Fig. 3.6(b)
27
Figure 3.2. WSAO modal hill-climbing algorithm flowchart for the fluorescence image optimization process; deformable mirror (DM), variable lens (VL).
3.3. Results
3.3.1. WSAO f/c biomicroscope resolution
The biomicroscope system design was computer simulated with Zemax (ZEMAX
Development Corporation, Bellevue, WA) to model the spot size on the mouse retina off
of the optical axis. The simulation results predicted diffraction-limited performance within
a 0.2 mm FOV. The FOV on the retina was limited to ~0.8 mm by lens L6 in Fig. 3.1, note
that this imaging mode was not diffraction-limited, but useful for navigation to landmarks
on the retina.
The performance of the optical imaging system was evaluated for both reflectance
and fluorescence channels on resolution targets. The performance of the reflectance
imaging system was measured by setting the DM to the flat position and placing a US Air
Force (USAF) resolution target (Fig. 3.3) in the retinal plane. We imaged groups 6 and 7
of the USAF target and we were able to resolve the smallest element of group 7 that has
a line width of 2.19 µm.
28
Figure 3.3. US Air Force resolution target with line width 2.19 µm highlighted by the red rectangle to demonstrate the reflectance resolution. Scale bar: 50 µm.
The resolution of the fluorescence channel was measured by imaging fluorescent
beads (190508, Polyscience, Inc.) that had a diameter of 2.1 µm with a standard deviation
of 0.018 µm. The WSAO optimization was performed on the fluorescent signal from the
beads and the intensity profile across the diameter of a representative bead was plotted
to demonstrate the system resolution and AO performance (Fig. 3.4). A line plot of a bead
after optimization shows a 5% intensity increase and a narrower full width at half
maximum, suggesting that the optimization corrected small system aberrations. The RMS
of the Zernike polynomial coefficients obtained from this optimization was 0.02 µm, which
is below the Maréchal criterion for diffraction-limited imaging (λ/14).
Figure 3.4. Images of 2.1 µm diameter fluorescent beads acquired (a) before WSAO optimization and (b) after optimization. (c) The line plots for a bead before and after optimization. Scale bars: 10 µm.
3.3.2. In vivo WSAO confocal fluorescence imaging of retinal ganglion cells
We acquired retinal images of anesthetized mice to demonstrate the WSAO f/c
biomicroscope performance in vivo. Images of an EGFP-labelled ganglion cell are
displayed before and after the optimization in Fig. 3.5. A video of the change in the
appearance of the ganglion cell during the optimization process as displayed on the
29
screen is included in Visualization 1 of [15]. The images in Fig. 3.5 were produced using
identical processing steps, and by averaging 50 registered frames. The Medical Image
Registration Toolbox (MIRT) for Matlab was used to register the frames prior to averaging
[69]. Given the small amplitude and slow speed of intra-frame motion in the anesthetized
mice during the 5 frames per second acquisition, we used a non-rigid cubic B-spline
registration algorithm with a sum of squared differences similarity metric. In the presence
of larger amplitude or faster motions, more advanced registration algorithms, for example
strip-based registration [70–72], may achieve better performance. The intensity-based
similarity measurement performed well for the frames with a strong signal from Fig. 3.5(b),
although the registration was less effective for the dendrites in Fig 3.5(a) due to the lower
signal intensity. The dendrites in Fig. 3.6(b) appear to be less resolved on the right side of
the image; this is due to the dendrites being out of the focal plane.
Figure 3.5. (a,b) Ganglion cells labelled by EGFP comparing the images acquired before and after the WSAO optimization. These images are an average of 50 frames of an off-axis ganglion cell. Scale bars: 20 µm.
The optimization results for these images are plotted in two ways: the optimum
Zernike coefficient value determined by the search algorithm (Fig. 3.6(a)) and the
corresponding merit function for each optimized mode in the order of the optimization (Fig.
3.6(b)). The intensity profile of a dendrite is displayed in Fig. 3.6(c) to demonstrate the AO
performance. Ganglion cells are classified into types based on the cell structure with
properties including soma size and dendrite patterns; a short review of this topic was
presented in Geng et al. [4]. The image of the ganglion cell in Fig. 3.5 appears to have a
round soma that is approximately 20 µm in diameter with thick and straight dendrites.
Based on this description, this ganglion cell may belong to the RGA2 category as described
30
by Sun et al. [73]. This classification is further supported with the description by Coombs
et al. [74] and images by Geng et al. [4].
Figure 3.6. (a) The Zernike coefficients applied to the DM (deformable mirror) after the optimization. (b) The impact of the optimization on the intensity-based merit function are plotted for each mode. The intensity is normalized from zero when the DM is flat. The Zernike coefficients are reported by the OSA standard for optical aberrations of eyes [68]. (c) The intensity plot of a dendrite on the EGFP-labelled ganglion cell at the location and in the direction indicated by the arrows.
3.3.3. In vivo WSAO confocal fluorescence imaging of retinal microglia cells
We also imaged EGFP-labelled microglia cells with similar results. The images of
microglia presented in Fig. 3.7 before and after the optimization were processed by the
same method as the ganglion cell images. In Fig. 3.7(b) the fluorescence signal on the left
side of the image has lower intensity after the optimization due to the microglia being at a
different depths within the retina. This demonstrates the optical axial sectioning effects of
the confocal pinhole and the ability of WSAO to reject out-of-focus features.
31
Figure 3.7. Images of EGFP-labelled retinal microglia cells acquired in vivo before and after WSAO correction with different field of views: (a), (b), and (c). Images (b) and (c) were taken at the same location with different field of views as indicated by the red dashed box. Each image is an average of 50 frames. Scale bars: 10 µm.
3.4. Discussion
The mouse eye is commonly used as a model of the human eye for vision
research. High quality images of the retina can be acquired without AO in some mice, in
particular for animals with healthy eyes [49,50]. Time course studies in mice and research
on transgenic mouse models of degenerative retinal diseases stand to benefit from the
incorporation of AO with the imaging system for increased resolution across a wider range
of animals, improved consistency in image resolution between time points, and for
locations that are off the optical axis of the eye. Additionally, AO provides the ability to
control the focus plane of the imaging system, which allows for easy and controlled
transition between layers of interest within the retina. Developing accessible AO that is
low-cost and small in size has the potential to be more widely used across multiple
research specialties.
We demonstrated WSAO for non-invasive in vivo fluorescence imaging of the
mouse retina. Our results showed cellular-resolution images acquired using a lens-based
32
AO system without the difficulties associated with implementing a WFS for mouse retinal
imaging. Our system included two electronically controllable elements on optically
conjugated pupil planes: a tunable lens controlling defocus for layer selection (which can
loosely be referred to as a ‘woofer’) and a deformable mirror for higher order corrections
(‘tweeter’) [75–77]. We used an intensity-based image quality metric to search for the
Zernike coefficients that would produce the strongest fluorescent signal and hence better
resolution. Recent research results suggest that using an intensity-based metric
exclusively is not sufficient for confocal scanning ophthalmoscopes, due to differences in
the illumination and the collection paths. Sulai et al. have demonstrated an approach using
a sharpness metric to optimize the point spread function of the illumination path in order
to increase the resolution [54]. In future work, we will investigate the use of a sharpness
metric, for example as defined in [54], in place of the intensity metric used in this chapter.
We anticipate that in cases where the fluorescent signal arises from non-planar structures
that a sharpness metric will result in an improved optimization performance.
The WSAO f/cSLO method used the same source of light (the fluorescence)
confined to a single retinal layer for both imaging and for guiding the aberration correction
(since the merit function was derived from the fluorescence images). In common AO
imaging configurations using a SH-WFS, the sensing (‘beacon’) and imaging wavelengths
are different. According to Zhou et al. [78], the chromatic defocus between the sensing
light and the imaging light are significant in the small animal eye, and can introduce higher
order aberrations in addition to defocus. This further accents a strength of the WSAO
technique, in that the aberration correction is performed at the same retinal depth section
that is being imaged, for example with the ganglion cells in the inner retina.
Our optimization algorithm required ~30 seconds to perform an exhaustive search
up to the 20th Zernike mode with 21 steps per mode. In addition to the modal hill-climbing
algorithm demonstrated in this chapter, the performance of other approaches such as the
simulated annealing algorithm or the stochastic parallel gradient descent algorithm could
also be explored [79,80]. For human imaging, the WSAO optimization speed is essential
due to the motion of the patient [61]. However, for an anesthetized mouse the amount of
motion is low and the optimization speed is not as crucial; this is supported by a recent
report by Palczewska et al. [81], where they used WSAO for two-photon microscopy and
required 4-6 minutes for an optimization.
33
The approach to WSAO retinal imaging in mice described in this chapter utilized a
fixed microscope objective lens and a plano-concave lens with radius of curvature
matched approximately to the curvature of the eye, reducing the refraction at the cornea.
This configuration constrained the FOV by the size of the mouse pupil and the mismatch
between the field curvature at the image plane of the objective lens and the curvature of
the mouse retina. Delivering a collimated beam into the mouse eye and allowing the
cornea and lens to focus it on the retina would provide a larger FOV. However, with a
collimated beam approach, a significantly larger amount of focus correction at the final
pupil plane would be required to optically section through the retina, on the order of ~ 30
– 40 diopters [5]. Another possible solution for our system would be to design a custom
multi-element objective lens with a scan pivot closer to the pupil in order to permit a wider
field of view. An advantage of this approach is in the potential ease of reconfiguring the
system for AO retinal imaging in different animal species, such as rats (see recent work
by Geng et al. [82]). This would require only a change in the final objective lens to
accommodate a different NA and the final concave surface to accommodate a different
corneal curvature.
A significant benefit of the WSAO algorithm is that the aberration correction is less
sensitive to multiple reflections from the sample and optical elements. Even in the
presence of undesired reflections that would affect SH-WFS wavefront measurement,
WSAO could still perform the optimization if these reflections do not change significantly
with the DM shape or if they are removed from the region of interest used for the merit
function with image processing techniques. As reported by Biss et al. [62] and Geng et al.
[5], challenges in wavefront sensing can arise due to reflections from different layers within
the relatively thick mouse retina as well as due to increased scattering. This effect is even
more pronounced in albino animals, in which the choroid and sclera layers also generate
a large backscattered signal. We employed a confocal detection which allowed the WSAO
to optimize the image signal within the depth of focus and minimize the out of focus signal.
The mice used for imaging in this chapter were pigmented; however, since the WSAO
algorithm does not rely on wavefront sensing, similar performance is anticipated in non-
pigmented animals. It has recently demonstrated WSAO OCT in albino mice with no
difference in performance with respect to pigmented mice [83].
34
3.5. Summary
In this chapter, we demonstrated WSAO for non-invasive in vivo fluorescence
imaging of the mouse retina. We imaged transgenic mice with EGFP-labelled ganglion
and microglia cells and used WSAO to increase the image resolution by correcting for
wavefront aberrations introduced by the eye. The AO system demonstrated cellular-
resolution imaging with a low-cost, simple and robust lens-based system.
In the next chapter, we investigate an alternative method for image-based AO
using pupil segmentation and provide a comparison to the hill-climbing optimization. We
transition to using collimated light into the eye with a 0-diopter contact lens placed on the
eye to prevent dehydration instead of focused light with a fundus lens.
35
Chapter 4. Pupil segmentation adaptive optics for in vivo mouse retinal fluorescence imaging
4.1. Introduction
Adaptive Optics (AO) is a technology used for human retinal imaging [2] that is
also of high interest for pre-clinical research. Vision research commonly uses small animal
models of vision robbing diseases, particularly mice, because they are inexpensive, and
are versatile to genetic manipulations. Non-invasive optical imaging of the mouse retina
permits retinal diseases to be characterized and the effects of potential therapies to be
studied in vivo and longitudinally. The mouse eye is well suited for high resolution non-
invasive optical imaging due to its large Numerical Aperture (NA). The large NA
exacerbates aberrations and necessitates AO in the case of a non-ideal eye to restore
diffraction-limited performance, and to allow for in vivo high resolution analysis [62].
Conventional AO employs a Shack-Hartmann (SH) Wavefront Sensor (WFS) to
measure the aberrations in combination with a deformable element to perform the
correction [2]. SH-WFS AO has been reported in the Literature for mouse retinal imaging
achieving high quality performance, for example [4,13,14]. While this method has
demonstrated excellent aberration correction ability for rodent imaging, SH-WFS based
approaches can be challenging as they are sensitive to wavefront reconstruction errors
produced by non-common path errors, multiple reflective retinal planes (the ‘small eye
artifact’) [5], and specular reflections [78].
Wavefront sensorless adaptive optics (WSAO) is an alternative method that uses
images acquired with the optical system to determine the optimal shape of a deformable
element to correct the wavefront aberrations. WSAO has demonstrated promising results
in microscopy as well as retinal imaging in humans and mice [15,62,63,80,84,85] A recent
review of WSAO algorithms can be found in [86]. A method that is common to many WSAO
reports in the Literature is iteratively changing the shape of the deformable mirror while
optimizing an image quality metric [86]. The quality of the aberration correction obtained
with merit function based WSAO is often sensitive to the number of images used to
36
optimize the shape of the deformable element, which comes at the cost of an increased
optimization time. Reducing the time required for aberration correction becomes a critical
goal to translate WSAO for in vivo mouse retinal imaging applications in vision science. In
this chapter, we demonstrate an alternative method of WSAO that uses the acquired
images to indirectly measure the wavefront aberrations in the sample; this approach is
known as Pupil Segmentation Adaptive Optics (PS-AO). This approach to ocular
aberration measurement is also conceptually similar to that reported in [87].
PS-AO indirectly measures a wavefront using images acquired with different
segments of the imaging pupil to determine the gradient of the wavefront at each pupil
region. In the case where no aberrations are present, all of the rays across the pupil of the
imaging system will converge at the sample to a focal spot size limited by diffraction.
However, in the presence of aberrations, the rays at different segments across the pupil
are deflected to different lateral positions at the focal plane due to heterogeneity in the
index of refraction and imperfections in the shape of the ocular structures. PS-AO
measures the deflection of the beam at each pupil segment through comparison of the
image acquired at that position with a reference image, commonly selected as the central
region of the pupil, in order to determine the local wavefront slope at that pupil segment.
A set of images acquired at each segment across the pupil. The wavefront gradient at
each segment is determined by measuring the lateral shift in the corresponding image
with respect to the reference image. These indirect measurements of the wavefront slope
using PS-AO are conceptually similar to the operation of a SH-WFS. The aberrations are
corrected by shaping the deformable mirror into the phase conjugate of the measured
wavefront. The PS-AO method has been demonstrated with great success by Ji et al. for
in vivo mouse brain imaging [44–46,88], which encourages the extension of this AO
method to in vivo retinal imaging modalities.
In this chapter, we present PS-AO for fluorescence Scanning Laser
Ophthalmoscopy (SLO) in mice. PS-AO for in vivo retinal imaging requires considerations
for the motion from the living sample as this has potential to corrupt the image-based
wavefront measurement. Our implementation used a Micro-Electro-Mechanical (MEMS)
segmented deformable mirror to segment the imaging pupil and also to correct the
aberrations. Our results demonstrate the improvement in the resolution of images
acquired of mouse retinal capillaries for fluorescein angiography.
37
4.2. Methods
We have implemented a compact, low cost, lens-based AO system for
fluorescence SLO to acquire mouse retinal images. The imaging system schematic in Fig.
4.1 represents the optical layout that is based upon the system described in Chapter 3
[15], with modifications for pupil segmentation. The imaging system had dual adaptive
elements, with a tunable liquid lens (Varioptic, Arctic 39N0) and a MEMS segmented
deformable mirror (PTT111, Iris AO, Inc.). The liquid lens was used to select the depth
focal position within the retina. The deformable mirror (DM) consisted of 37 individually
controllable mirror segments. The ability to tilt light from the segments, which were
mapped to the pupil, made the DM well suited for pupil segmentation. Single mirror-
segment beamlets were initially created to provide a densely sampled wavefront, however
the laser power at the sample was inadequate and the signal to noise ratio was poor.
Consequently, we created ‘beamlets’ by grouping 7 mirror segments (a central segment
surrounded by six neighbors); for the groups at the edge of the DM, 4-5 mirror segments
were used. To create a beamlet, a group of segments remained flattened on the DM,
directing light to the sample, while the remaining segments were tipped to deflect the beam
toward a spatial filter created with an iris diaphragm.
38
Figure 4.1. Schematic of the Scanning Laser Ophthalmoscope: 488 nm laser; dichroic mirror (DC); deformable mirror (DM); variable lens (VL); galvanometers (GM); {L1,L2, L3,L4,L5,L6} = {200,200,150,100,50,19} mm. (a),(b) PS-AO on 6 µm fluorescent beads with aberration correction (AO On) and without (AO Off). These images are an average of 30 frames. Scale bar: 8 µm. (c) Wavefront aberration map. (d) Normalized intensity plotted at the location indicated by the dashed lines with a ~30% increase in the peak intensity after correction. (e) The Zernike coefficients for the corrected wavefront.
We acquired in vivo fluorescent mouse retinal images using a 488 nm laser
(Coherent Inc. 488 OBIS). The full beam diameter at the cornea was 0.89 mm, focused
by the mouse eye with an NA of ~0.23. In this chapter, the maximum NA was limited by
the stroke of the DM. The beam diameter of the central beamlets at the mouse pupil was
0.38 mm with an NA of ~0.1, while the diameter of the outer beamlets was approximately
0.27 mm. The size of the focal spot of an individual beamlet (~3 µm Airy disk radius at the
retina) constrained the lower limit of the dynamic range for the wavefront sensing, while
the upper limit was restricted by the largest image shift that could be detected. The
39
fluorescence signal was focused through a 100 µm pinhole with a confocal aperture 6.5
times the Airy disk, detected by a photomultiplier tube (Hamamatsu H7827-002). A 1 kHz
bi-directional sinusoidal scan pattern was applied to the galvanometer mirror (6210H
Cambridge Technology Inc.), capturing data in both the forward and backward scan
directions. Phantom imaging was performed with a 3 mm objective lens in place of the
mouse eye, with an NA of 0.15 while acquiring 400 x 400 sampled images (ie. pixels). For
in vivo applications, the image size was 200 x 200 samples at 10 frames per second.
Image acquisition was performed with custom software [89].
In our initial experiments on PS-AO in vivo mouse retinal imaging, the image shift
introduced from sample motion impeded the image-based measurement of the wavefront
slope. In order to mitigate the effect of motion introduced from live animal retinal imaging,
we modified the PS-AO image acquisition process to collect a reference image in rapid
succession with each target image. For the deformable mirror used in this chapter, this
resulted in 36 reference images and 36 target images acquired at each of the pupil sub-
regions. The total of 72 images was acquired in ~7.2 seconds. To further minimize the
effect of motion from in vivo samples, the acquisition alternated between the reference
and target beamlets within a frame. We acquired sections of the frame corresponding to
50 lines at a time, alternating between reference and target regions on the segmented DM
every 25 ms.
The reference images were acquired with the central region of the pupil, selected
as the desired focal point in the image plane for the target beamlets. Aberrations in the
optical path caused a lateral shift between the reference and target images. A rigid
registration algorithm was used to find the shift between each pair of images. The
translation distance was used to calculate the local wavefront slope at each pupil segment.
Using this information, we then applied a modal wavefront reconstruction. The wavefront
was approximated by multiplying the slopes by the pseudoinverse of the Zernike gradients
to determine the Zernike coefficients. The set of Zernike coefficients up to the 15th mode
was applied to the DM for aberration correction. The final images with the aberration
correction were acquired with the full imaging pupil, and the deformable mirror shaped to
the phase conjugate of the overall wavefront. Modes 0, 1, 2 (piston, tip, tilt) were set to
zero [68].
40
We demonstrated PS-AO imaging on fluorescent phantom samples before
performing in vivo experiments. The phantom samples were created using fluorescent
beads or lens tissue fibers labelled with fluorescein. Aberrations were induced by placing
a gel between two non-uniform plastic surfaces in front of the sample. In Fig. 4.1 the
aberration correction was performed on a 400 x 400 sample image with a field of view
(FOV) of ~80 µm. Fig. 4.2 shows a comparison between PS-AO and a hill-climbing
coordinate search algorithm [15] to investigate the amount of wavefront correction
achieved with each method.
Figure 4.2. Aberration correction performed with both hill-climbing and PS-AO. (a) Image without aberration correction. (b) Correction performed with hill-climbing. (c) Correction performed with PS-AO. (d) The Zernike coefficients for the corrected wavefronts.
An experiment was performed on the lens tissue phantom to investigate the
sensitivity of PS-AO to sample motion. Aberration correction was demonstrated three
times on a sample at nominally the same location. In the first scenario, the sample was
static, and the results from the aberration correction are shown in Fig. 4.3(a). Next, lateral
motion was manually introduced in the sample, approximating the motion observed due
to respiration in an anesthetized mouse. The results of PS-AO using only a single
reference frame (i.e. a total of 37 images) in Fig. 4.3(c) indicate that the aberration
correction was not successful. By using the intra-frame reference acquisition method
described above (i.e. a total of 72 images), the PS-AO results shown in Fig. 4.3(b),
demonstrate good aberration correction even in the presence of motion.
41
Figure 4.3. PS-AO aberration correction on (a) static and (b, c) moving samples. The aberration correction was performed with (b) using the multiple intra-frame reference images and (c) a single reference image.
Next, PS-AO was performed to optimize the images of the capillaries for
fluorescein angiography in mouse retina in vivo. Wild type C57BL/6J mice (Jackson
Laboratories, Bar Harbor, ME) were used for imaging in this chapter. All imaging
experiments were conducted with the approval of the University Animal Care Committee
at Simon Fraser University while following the protocols compliant to the Canadian Council
on Animal Care. The mice were anesthetized with a subcutaneous injection of ketamine
(100 mg/kg of body weight) and dexmedetomidine (0.1 mg/kg of body weight) prior to the
imaging session. Following the injection, the eyes were dilated with a drop of topical
solution (Tropicamide, 1%). A contact lens (Cantor & Nissel Ltd, UK) was then applied to
protect the cornea from dehydration. The anesthetized mouse was placed on a translation
stage and aligned to the imaging beam without contact. When switching between imaging
with a fully illuminated pupil versus a pupil segment, the laser power was adjusted to not
exceed ~150 µW incident on the mouse cornea. After the experiment, the recovery of the
mice was induced with an injection of atipamezole (1.8 mg/kg of body weight). The PS-
AO process was performed on a small FOV (~200 µm) sampled with a 200 x 200 points.
The defocus term in the second panel of images in Fig. 4.4 was set to zero in order to
examine the same imaging plane before and after correction. Following aberration
correction, images were acquired at the same FOV, and then optically zoomed out to a
larger FOV (~400 µm). Each image in Fig. 4.4 is an average of 30 frames registered using
the Medical Image Registration Toolbox [69]. The images before and after correction were
selected at the best focus with the Varioptic lens prior to the acquisition. The results show
the improvement in the image brightness and sharpness after PS-AO.
42
Figure 4.4. (a), (b) PS-AO for retinal fluorescein angiography with aberration correction (AO On) and without (AO Off) for two mice. In each panel, the top row of images (angular FOV 5.2°) is an optically zoomed in section of the bottom row of images (angular FOV 10.4°). Scale bars: 20 µm. (c) Zernike coefficients for the corrected wavefront. (d) On the top panel, the normalized intensity plot at the location indicated by the dashed lines had a ~30% increase in the peak intensity after correction, and (d) on the bottom panel, the wavefront aberration map.
4.3. Discussion
Wavefront sensorless AO alleviates some of the challenges with direct wavefront
sensing, facilitating the extension to applications of retinal imaging in small animals.
Strong specular reflections of a laser beacon from refractive elements, such as the cornea
or lens-based optics in the imaging system, can impede conventional SH-WFS
approaches to AO. The ‘small eye artifact’ is another source of potential error confounding
SH-WFS methods, particularly in albino animals. In this chapter, we have demonstrated a
43
wavefront sensorless lens-based AO system that is robust to these issues by using the
fluorescence images for aberration correction.
A segmented deformable mirror was used in this chapter as it was naturally well
adapted for pupil segmentation. Each beamlet contained 4-7 mirror segments that were
used to sample the gradient of the wavefront, defining the reconstruction resolution of the
wavefront. Using one mirror segment per beamlet would improve sampling density as long
as the laser can deliver enough power to maintain good signal-to-noise images. Using
multiple segments to define a beamlet provided a smaller depth of focus (~50 µm) than a
single segment (~500 µm), reducing noise due to fluorescence from other layers in the
sample.
We demonstrated PS-AO as an approach to WSAO using 72 frames, which were
acquired in 7.2 s to measure the wavefront slope at 36 locations across the pupil. For
PS-AO, the limit to the number of correctable modes is dependent on the number of
samples taken across the wavefront. This upper limit can be determined by minimizing
the difference in the root-mean-square error between a measured and reconstructed
wavefront, and approximately corresponds to the sampling density across the wavefront
[39], or the number of beamlets created within the imaging pupil. Had the individual
mirror segments been used to form beamlets with our particular implementation of PS-
AO, up to 36 Zernike modes could be corrected while maintaining the same optimization
time. In this chapter, we performed aberration correction up to the 15th mode as our
previous work demonstrated that higher modes had less impact in correcting mouse
ocular aberrations [15]. Compared to our previous work using a modal hill-climbing
algorithm [15], this is ~5 times quicker for the optimization of the same number of
modes. However, the amount of modes that are required for aberration correction can
vary due to factors such as motion artifact, the quality of the mouse eye, or the
eccentricity of the region of interest. We showed that PS-AO has the potential to perform
well while requiring fewer images.
44
4.4. Summary
In this chapter, we demonstrated aberration correction for fluorescein angiography
of retinal capillaries. PS-AO is susceptible to errors in the wavefront reconstruction due to
motion of the sample. We presented an implementation of PS-AO that minimizes the
effects of motion artifact, and demonstrated a viable technique for WSAO aberration
correction for in vivo retinal imaging in mice.
45
Chapter 5. Adaptive optics in the mouse eye: Wavefront sensing based vs. image-guided aberration correction
5.1. Introduction
Retinal imaging with Adaptive Optics (AO) is necessary to allow reliable
visualization and monitoring of single retinal cell morphology in vivo by correcting for
ocular aberrations of the eye, which acts as the microscope objective. Animal models are
important for studying pathophysiology and treatment of many human diseases. This also
includes common eye diseases such as diabetic retinopathy, glaucoma, and age-related
macular degeneration, as well as rare genetic diseases such as retinitis pigmentosa.
Conventional ex vivo immunohistochemistry often used in these studies provides exquisite
cellular contrast and high cellular resolution of the retina, but only at a single point in time.
This results in studies with large cohorts of animals multiplied by the number of time points
that are needed. Non-invasive imaging of living animals enables the characterization of
the progression of pathology and the evaluation of new therapies for eventual use in
humans within a single mouse during longitudinal studies, greatly reducing numbers of
animals needed for the experiment [4]. Mouse models are widely used in preclinical
research partially due to the availability of transgenic strains, which includes mice with the
relevant cell classes labeled by fluorescent proteins. Imaging modalities such as Adaptive
Optics - Scanning Light Ophthalmoscopy (AO-SLO) employed with fluorescence detection
can be used to examine the structural and functional features [64,65] in the retina at
cellular resolution. A recent review article [3] further describes the significance of AO for
retinal imaging in vision science.
In comparison to primates, mice have eyes with a relatively high numerical
aperture that provides the potential for sub-micrometer diffraction limited resolution. There
are challenges associated with consistent high-resolution AO imaging for small animal
eyes that arise from sensitivity to alignment, and the length of the eye relative to the
thickness of the retina. The short focal length creates a large relative optical thickness of
the retina, requiring a large dioptric power to optically section through the retinal layers
[78], which is often also a source of aberrations.
46
Accurate measurements with a Shack-Hartmann (SH) Wavefront Sensor (WFS)
require a well-defined reference light source or ‘guide star’, with sufficient signal above the
background noise photons from other scattering tissue for layer specific aberration
correction [90]. The so-called ‘small eye artifact’ [5] means that instead of a single
scattering reference for the SH-WFS as in the case in human retinal imaging, the various
strongly scattering retinal layers, such as the heavily pigmented Retinal Pigmented
Epithelium (RPE), Choroid, and Nerve Fiber Layer (NFL), have the potential to confound
wavefront sensor measurements. The difficulty of performing direct wavefront
measurements from the small animal eye further increases for albino strains where the
backscattered signal from the desired imaging target, such as RPE, is not present due to
lack of melanin and the WFS is overwhelmed by background scattering from the sclera.
The conventional method for aberration correction with AO requires closed-loop
feedback between residual aberrations measured by SH-WFS in response to changing
the shape of wavefront corrector. Multiple groups have demonstrated WFS AO to provide
high-resolution retinal images in the mouse eye, including [4,13,14] to list a few. However,
a direct measurement of the wavefront from the retinal volume of interest in small animal
eyes requires a complex system and may not always be possible due to the multi-layered
structure of the sample, in addition to mild cornea or intraocular lens opacities.
Wavefront Sensorless (WFS-less) AO is an alternative approach to the WFS
based AO method that uses an image-based optimization method to correct the
aberrations. Multiple approaches to WFS-less AO have been reported, including amongst
others: stochastic steepest gradient descent, simulated annealing, hill climbing modal
optimizations, and pupil segmentation [45,85,86]. WFS-less AO SLO has been
demonstrated for cellular resolution retinal imaging in mice [15,20,63], and WFS-less AO
has also been demonstrated to correct for non-common path errors in combination with
WFS AO [54]. WFS-less AO has the potential to alleviate the dependency on the WFS
alone and provide an alternative method for aberration correction at multiple depths in the
retina as well as simplifying the imaging system [16]. However, WFS-less AO requires an
optimization execution time several times longer, which can be disrupted by the motion
inherent in a living, breathing animal.
In this chapter, we demonstrate WFS-less AO for aberration correction that was
implemented in a state-of-the-art WFS AO SLO system design for mouse retinal imaging,
47
and present comparisons of the image quality and wavefront measurements during AO
correction performed by each method. Confocal SLO back-scattering images of the mouse
photoreceptor mosaic are presented, representing a case where the WFS and WFS-less
AO are both using the same reference to guide the AO. As an example of a case where
the wavefront sensing and the imaging planes are different, images of mouse retinal
microglia labeled with fluorescent proteins are presented. We demonstrate that WFS-less
AO can correct the same aberrations that are measured by a SH-WFS and that WFS-less
AO can perform depth-resolved aberration correction resulting in reliable imaging focal
plane shift in a mouse eye.
5.2. Methods
Mouse retinal imaging was performed using a custom designed AO SLO system
that has previously been described [14,19]; please see Section 5.2.1 for details. The AO
SLO design was based on SH-WFS AO, using wavefront measurements from the mouse
eye to perform aberration correction. The differences from the previously reported
configuration include: the scanning mirrors were changed to galvanometer mounted
mirrors, and the optical detection path. For WFS AO, the deformable mirror (DM) control
was provided by the WFS measurement software, which allowed for the closed-loop
functionality. The SH-WFS sub-system is described in Section 5.2.2. Only software
changes were required to implement the image-based WFS-less AO approach; the WFS-
less AO software is described in Section 5.2.3. The WFS-less AO required the image
acquisition software to control the DM. The system recorded wavefront data from the
mouse eye as measured by the SH-WFS while performing WFS-less AO.
5.2.1. AO SLO system description
The AO SLO system was custom designed for reflectance and fluorescence
imaging of the mouse retina, further described in Zawadzki et al. [14,19]. Light from a
superluminescent diode (SLD, Superlum, SLD-26-HP) with a 663 nm center wavelength
was used for reflectance imaging, and as a beacon for wavefront sensing. The system
also used co-aligned light from a 488 nm laser (Coherent, OBIS 488 nm LX) for
fluorescence excitation. The laser power at the mouse eye was limited to 100 W for each
light source. The first pupil plane was defined by the continuous membrane DM (ALPAO,
48
DM97-15). The optical plane was relayed to the horizontal galvanometer scanning mirror
and then to the vertical galvanometer scanning mirror (Cambridge, 6215H) using afocal
telescopes made from pairs of spherical mirrors. The pupil plane was further relayed with
a spherical mirror and a lens to the mouse eye with a final beam diameter of 2 mm to be
focused to the retina with the maximum available numerical aperture of ~0.5. The contact
lens was mounted at the last pupil plane for alignment of the mouse eye. The optical layout
of the system is shown in Fig. 5.1.
Figure 5.1. Adaptive Optics Scanning Laser Ophthalmoscopy (AO-SLO) system schematic. The layout is presented in a scale drawing. Abbreviations: L#, lens; F#, filter; BS#, beamsplitter; M, mirror; SM, spherical mirror; DM, deformable mirror; D#, dichroic mirror; Hsc, horizontal resonant scanner; Vsc, vertical scanner; PMT, photomultiplier tube; P (circled in blue) optical planes conjugate with the pupil; SLD, superluminescent diode. Collimated beams are marked as dashed lines and focusing beams are marked as solid lines. The on-axis beams are represented by red lines and scanned beams by green and blue. Image credit: Pengfei Zhang.
Table 5.1 lists the optical parameters of the important system components.
Table 5.1. Key optical parameters of the AO-SLO system components
Item BS1 BS2 BS3 D1 D2 F1 F2 DM
Type R:T= 50:50
R:T= 10:90
R:T= 30:70
T635lpxr-UF, Chroma
ZT488rdc-UF, Chroma
FF01-660/13, Semrock
FF01-525/45, Semrock
DM97-15, Alpao
Item SM1 SM2 SM3 SM4 SM5 SM6 SM7 L0
Focal length
900mm 1350mm 1350mm 375mm 150mm 150mm 762mm 400mm
49
The back-scattered light from the 488nm laser was split and relayed with pairs of
lenses from the DM to the first photo multiplier tube (PMT1, Hamamatsu Photonics,
H7422-20). The fluorescence light was separated with a dichroic mirror and relayed to
PMT2 (Hamamatsu Photonics, H7422-40). The back-scattered light from the 663nm SLD
was split by BS3 (beamsplitter, R:T=30:70), the reflected portion was acquired by PMT3
(H7422-50), and the other portion went to the SH-WFS, which was created by a lenslet
array (Pitch = 150 μm, f = 6.43 mm) and a CMOS camera (UI306xcp-M; IDS Imaging
Development Systems GmbH). The back-scattered light created a 6 mm circular aperture
on the SH-WFS with a total of 1264 wavefront samples. The WFS-AO control is described
in Section 5.2.3.
The image acquisition program was developed using custom C/C++ for real-time
image display and to control the galvanometer mirrors. The current output from each PMT
was converted to voltages with transimpedance amplifiers (Femto, HCA-2M-1M-C) and
digitized with an analog-to-digital converter (NI PCIe-6363) capable of multi-channel
acquisition at 1.00 MSPS (mega samples per second). The frames were sampled at 400
x 200 pixels during acquisition, and the sampling density was reduced to 400 x 100 pixels
to increase the frame rate during WFS-less AO optimizations. The galvanometer mirrors
were scanned using a bi-directional pattern in a 1 kHz sinusoid that acquired data in the
forward and backward scan directions. The images were de-warped for display in real-
time.
5.2.2. WFS AO description
We used a custom control software to perform WFS measurements from the
sample and to control the DM for closed-loop AO aberration correction. This software was
provided by the University of Verona [91]. The WFS centroids were selected with a circular
aperture for wavefront reconstruction and display in Zernike modes. The WFS-AO was
activated by closing the control loop between the WFS and DM. In order to image different
retinal layers within the eye, the user could enter the desired amount of defocus. Under
this condition, the wavefront measurement would be relative to a reference with the
defocus value included. The WFS software could release the connection from the DM so
that the acquisition software could control the DM for WFS-less AO, yet allow the
wavefront measurements to still be recorded for analysis. The exposure time of the WFS
was set to capture light as the beacon was scanned across the sample, which
50
accumulated light in the WFS from across an area within the eye. The camera exposure
time and wavefront reconstruction time limited the WFS AO system to 100Hz.
The AO control began with calibrated actuator settings or ‘system flat’, which
removed the system’s static aberrations measured by the WFS. The calibration procedure
is further described in Section 5.2.4
5.2.3. WFS-less AO algorithm.
We implemented a hill climbing Coordinate Search (CS) algorithm, which was
driven by either the reflectance images or the fluorescence images [15]. The CS algorithm
searched within the range of Zernike coefficients expected for a mouse eye. Modal control
of the DM was calibrated using the procedure further described in Section 5.2.4. The
optimization algorithm used a merit function for the highest image sharpness (Simg), which
was calculated by the sum of the pixel intensity squared on the entire image, Equation
5.1. This metric has been used extensively in implementations of WFS-less AO [92–95],
since it is easy to compute with good performance on both reflectance and fluorescence
SLO imaging of a variety of features in the retina [16].
𝑆𝑖𝑚𝑔 = ∑[𝐼(𝑥, 𝑦)]2
𝑥,𝑦
, (5.1)
where I(x,y) is the pixel intensity at the location x,y in the image.
The CS algorithm began with the system flat and then for a given mode (k), a
range of coefficients (±α) were applied to the DM. The coefficient (a*n) that corresponded
to the best image according to the image metric was applied to the DM and the algorithm
moved onto the next mode. Before recording the metric values used for the optimization
algorithm, extra imaging frames were included each time the DM returned to the best
coefficients to guarantee sufficient settling time between searched modes.
For the first iteration, the CS algorithm began by finding the best initial defocus (k
= 4) value, then the astigmatisms, and then continuing in ascending order to include up to
the 5th radial order of the Zernike polynomials (18 modes total). The Zernike polynomials
were ordered and reported using the mode number according to the OSA/ANSI standard
[68]. Further iterations always began with defocus and then the other 17 modes were
51
optimized in a variety of sequences. Typically, we searched 3 to 4 iterations of n = 18
modes with m = 21 coefficients for each mode, which would require 20 seconds per
iteration. Note that the optimization speed was limited by the imaging system frame rate.
The number of coefficients and the number of iterations could be easily adjusted within
the same imaging session if required. We stopped iterating when the image quality metric
no longer significantly increased from the previous iteration. The following procedure
further explains each step in the CS algorithm.
1) Set the DM actuators values to the calibrated system flat and set the Zernike
coefficients to be 𝑎𝑘 = 0, for 𝑘 = 1, 2, … , 𝑛.
2) If this is not the first iteration, then the selected Zernike coefficients from the previous
optimization are applied to the DM.
3) For each Zernike mode 𝑘 for 𝑘 = 1, 2, … , 𝑛 starting with 𝑘 = 1.
a. Update DM shape using Zernike mode k with varying amplitude over
a range of ± ak,max by incrementing with 𝑚 evenly spaced steps:
𝑎𝑘,𝑚 = − 𝑎𝑘,𝑚𝑎𝑥 +2𝑎𝑘,𝑚𝑎𝑥
𝑚−1(𝑖 − 1), for 𝑖 = 1,2, … , 𝑚.
b. Calculate the merit function on the image, 𝑆𝑖𝑚𝑔, for each coefficient
𝑖 = 1,2, … , 𝑚 and select the coefficient with the highest value from
the search, 𝑎𝑘,𝑚∗ .
c. Apply the selected coefficient 𝑎𝑘,𝑚∗ .
d. Move optimization search to the next Zernike mode, 𝑘 = 𝑘 + 1.
4) After each iteration through the Zernike modes, the algorithm can repeat and search
around the chosen coefficient values from the previous iteration. The Zernike modes
can also be searched in a different sequence in further iterations.
The WFS-less could also be performed after the SH-WFS optimization. In this case,
the best corrected wavefront as determined using the SH-WFS software could be applied
as a starting point for the WFS-less algorithm to perform additional image improvement.
5.2.4. WFS and WFS-less AO system calibration
The system was calibrated for WFS AO by placing a model eye and using the
imaging light source for measurements. The model eye was constructed from a 100 mm
focal length achromatic doublet (AC254-100-A) that focused light onto a scattering sample
52
such paper. Wavefront measurements were extracted from the SH-WFS camera image
feed. The incoming light passed through the Shack-Hartmann lenslet array, where each
lenslet focuses the incoming beam over a portion of the CMOS pixel array. The distance
from the measured position of the focused spot to the position of the ideal, non-aberrated
spot is related to the local wavefront gradient and is called slope [52]. The spot position
relative to each lenslet was determined with a Thresholded Weighted Center of Gravity
(TWCoG) algorithm. Then, the slopes are used to reconstruct the wavefront with Zernike
coefficients.
The influence of each actuator on the wavefront was measured by poking the
actuators and collecting the slope responses. Hadamard pattern was used to reduce
calibration time and maximize SNR [96]: Each poke pattern was a vector of half 0’s and
half 1’s, with an orthogonal column space generated from all the patterns. The slopes are
then measured after the actuators reach steady state. To account for stroke hysteresis,
the actuators are poked with the same patterns but with an inverted sign, and the average
between the two slopes responses is kept. The slopes are used to generate the actuator
control value space, obtained with the SVD of the calibrated influence matrix. The control
values are processed in an integrator array which asymptotically steers to remove the
difference between the aberrated wavefront and the DM shape compensating for it. With
the model eye in place, the actuator control values to remove the system’s static
aberrations were characterized by closing the WFS AO control loop using the calibrated
control matrix. Those values, called system flat, are saved for the WFS-less AO mouse
imaging.
For the WFS-less AO algorithm, the system was calibrated for modal control with
the Zernike polynomials. This was performed by stopping the scanning and illuminating
the DM and WFS in the reverse direction, starting at the sample pupil plane, illuminating
the sample with a ‘single pass’ of the DM. This was necessary in order to include the
Zernike modes tip and tilt in the actuator measurements. The wavefront influence of each
actuator was measured in Zernike modes, and each measurement was used as a column
in an influence matrix (A). Then, the pseudo-inverse was calculated so that the actuator
control values (c) for any Zernike vector (z) could be calculated by Equation 5.2.
𝐜 = 𝐀ϯ𝐳 . (5.2)
53
5.2.5. Animal handling and image processing
The animal handling in this work was performed in accordance with guidelines of
the animal study protocol approved by the University of California Animal Care and Use
Committee (IACUC). The three strains of adult animals, pigmented (C57Bl/6), albino
(BALB/cJ), and Cx3cr1GFP/+ mice from Jackson Labs (2-6 months old, female, 5 for each
strain) were used in the experiments, including Cx3cr1GFP/+ mice strain had retinal
microglia cells labeled with Enhanced Green Fluorescent Protein (EGFP). During imaging,
the mice were anesthetized with isoflurane (2% in O2), and the eye was dilated with
tropicamide (1%) and phenylephrine (2.5%). The anesthetized mouse was aligned to a
zero Diopter contact lens (Unicon Corporation, Osaka, Japan) with a gel (GenTeal, Alcon,
Fort Worth, United States) placed between the lens and the cornea to prevent dehydration
and the development of cataract.
Motion within and between imaging frames was mostly caused by the respiration
of the anesthetized mouse [81]. Registering and aligning frames in post-acquisition
processing was required for averaging of frames to improve the image quality. Typically,
frames were recorded at 10 fps for 5 to 10 seconds for a total of 50 to 100 frames for
averaging.
The registration process began by manually selecting a single frame, ideally free
of motion artifact, as the template to align the other frames. The rigid registration process
included a global frame translation followed by the translation of image slices created in
the horizontal fast-scan direction. The global frame translation was determined by
maximizing the cross-correlation between the target frame and the template frame using
the fast Fourier transform. Frames that had a much lower cross-correlation value than
most other frames due to large amplitude of motion were discarded. The remaining frames
were broken up into horizontal strips of 3 vertical pixels. Each strip was translated
horizontally and then vertically with subpixel resolution to maximize the cross-correlation
with equivalently sized strips on the template. Finally, the translated strips were averaged
together and down-sampled to the original frame size. This method was effective on low
SNR images [97]. However, it would not be capable of correcting motion within the 1.5 ms
required to acquire 3 lines as well as rotational and torsional distortions.
54
5.3. Results
5.3.1. WFS and WFS-less AO for phantom imaging, comparison of performance
We constructed an imaging phantom (a ‘model eye’) prior to in vivo mouse retinal
imaging in order to evaluate the AO performance in an ideal case. The model eye was
composed of a 100 mm focal length achromatic doublet and 30 µm diameter fluorescent
particles (Cat. No. 36-6, Ex/Em Max: 542/612 nm, Dry Fluorescent Particles, Thermo
Scientific, Waltham, US) on white paper. This enabled good WFS measurements for the
WFS-based AO and the fluorescence images provided well-defined features for the WFS-
less optimization using the fluorescent channel. Aberrations were loaded onto the DM
using a prior wavefront measurement from a mouse eye [19], which decreased the
fluorescence image quality as shown in Fig 5.2(a). First, we used WFS AO to correct the
aberrations; then we used WFS-less AO to correct the same aberrations. As described in
the Methods, the WFS-less AO correction proceeded for 4 iterations and increased the
sharpness metric during the optimization, as shown in Fig 5.2(b). The image quality metric
after WFS-less AO was 16% better than the image produced by the WFS-based AO.
The residual ocular wavefront was recorded before, during, and after aberration
correction using the slopes from the 1264 wavefront centroids, which was used to
reconstruct the first 300 Zernike modes and to calculate the RMS of the wavefront. The
WFS AO-corrected aberrations typically within the first 0.25 seconds, as shown in the
recorded wavefront in Fig 5.2(c). The WFS-less AO total execution time depended on the
optimization parameters selected for the CS algorithm, which could be modified in the
software user interface. In this optimization, we densely sampled 18 modes with 21
coefficients. This required about 80 seconds to decrease the WFS measurement to below
the diffraction limit according to the Maréchal criterion (λ/14), as shown Fig 5.2(d). The
optimization proceeded for 4 iterations, but it only took 2 iterations to get to 70% of the
final image metric value. Fig 5.2(e) compares the wavefront measurements in Zernike
modes before and after each method of AO to show that both methods effectively remove
aberrations.
The measured RMS of the residual wavefront after WFS AO was 0.023 ± 0.003
µm and after WFS-less AO was 0.047 ± 0.002 µm over 100 measurements (1 s of
55
wavefront data) including defocus. The RMS after WFS-less AO was 0.029 ± 0.002 µm
when defocus was removed from the calculation. The defocus term could explain the
image quality improvement of the WFS-less image over the WFS image, which would be
a result from the WFS-less AO having an imaging plane at the middle of the fluorescent
beads and the WFS AO having an imaging plane on the paper. Tip/tilt aberrations caused
by scanning across the sample were removed from all RMS calculations.
56
Figure 5.2. Phantom imaging of fluorescent beads and wavefront measurements during Wavefront Sensor Adaptive Optics (WFS AO) and Wavefront Sensorless Adaptive Optics (WFS-less AO). (a) Fluorescence images of 30 µm beads on white paper with a 100 mm focal length model eye before AO, after WFS AO, and after WFS-less AO. For the inset image before AO, the pixel intensity values were multiplied by 8, so the beads could be visualized. (b) The increase in the fluorescence image quality during the WFS-less AO optimization. (c) The wavefront RMS excluding defocus, tip and tilt during WFS AO correction. (d) The wavefront RMS excluding defocus, tip and tilt during WFS-less AO optimization. (e) The Zernike decomposition of the wavefront measured before and after each method of AO correction.
57
5.3.2. WFS and WFS-less AO comparison on mouse photoreceptor mosaic
To perform in vivo imaging experiments, we first we used pigmented mouse strain
(C57BL/6J) and targeted the strongly scattering RPE/choroid layer for good WFS
measurements. The SLD source was used for imaging and wavefront measurements with
a Field of View (FOV) of ~70 µm. First, we corrected and recorded the aberrations from
the mouse eye with WFS AO and acquired reflectance images of the rod photoreceptor
mosaic after the aberration correction, shown in Fig. 5.3(a). After WFS AO correction, the
RMS was calculated to be 0.07 ± 0.02 µm over 100 measurements. The system flat was
applied to the DM and aberration correction was performed with the reflectance images to
drive the WFS-less AO. The images after optimization are shown in Fig. 5.3(b). The
optimization is shown Fig. 5.3(c) used 5 iterations; however, the image quality was 75%
of the final quality after only 3 iterations. Fig. 5.3(d) shows the aberrations removed from
the wavefront as the RMS of the wavefront decreased during the optimization from 0.80 ±
0.01 µm to 0.22 ± 0.01 µm over 100 measurements, which included defocus in the
calculation. Note, we did not separately correct non-common path aberrations although
based on Fig. 5.2 results, correcting them would have a negligible impact on the remaining
results presented this chapter. In this optimization, the Zernike modes were searched in
the same order for each iteration. During iteration 4, the motion of the mouse disrupted
the optimization but the algorithm recovered in the final iteration. Fig. 5.3(a,b) shows that
the image quality is similar after both methods of AO finish correcting residual waterfront
error and the image quality metric after WFS-less AO was 9% better than the image
produced by the WFS-based AO before post-processing. Fig. 5.3(e) shows the wavefront
represented by Zernike modes before and after each correction method.
58
Figure 5.3. Imaging the mouse photoreceptor mosaic with Wavefront Sensor based Adaptive Optics (WFS AO) and Wavefront Sensorless Adaptive Optics (WFS-less AO). (a,b) Images after WFS AO and WFS-less AO. Scale bar: 10 µm. (c) The image quality improvement during WFS-less AO optimization. (d) The wavefront RMS during WFS-less AO optimization. (e) The Zernike decomposition of the wavefront measured before and after each method of AO.
We repeated this imaging experiment and measurements with different mouse
eyes with a variety of optimization parameters with similar imaging performance. Fig. 5.4
represent additional measurements for imaging the mouse retina photoreceptors. In both
of these cases, the WFS AO and the WFS-less AO demonstrated similar image quality
and measured aberrations were removed. In Fig. 5.4(a), after WFS AO correction, the
RMS was calculated to be 0.06 ± 0.02 µm. For the WFS-less optimization after the first
iteration, the search order of the Zernike modes was shuffled for each of the following two
iterations. The wavefront RMS decreased during the optimization from 0.93 ± 0.04 µm to
0.20 ± 0.01 µm. In Fig. 5.4(b), after WFS correction, the RMS was calculated to be 0.08 ±
0.03 µm. For the WFS-less optimization, the number of coefficients searched for each
Zernike mode was decreased from 21 to 11 and further iterations were included instead.
The wavefront RMS decreased during the optimization from 1.16 ± 0.02 µm to 0.24 ± 0.01
µm. (Note: The Zernike decomposition coefficient values of the measured wavefront after
WFS AO were too small to be well visualized on the same plot as the measured wavefront
before AO.)
59
Figure 5.4. (a, b) Further mouse photoreceptor imaging with Wavefront Sensor Adaptive Optics (WFS AO) and Wavefront Sensorless Adaptive Optics (WFS-less AO). Images mouse photoreceptor mosaic after WFS AO and WFS-less AO. Scale bar 10 µm. The Zernike decomposition of the wavefront measured before and after each method of AO. The wavefront RMS during WFS-less AO optimization. The image quality improvement during WFS-less AO optimization.
60
5.3.3. AO SLO reflectance imaging of an albino mouse strain
To demonstrate the potential advantage of WFS-less AO over classical AO, we
performed imaging of an albino mouse strain (BALB/cJ). Since these mice lack melanin,
which highly scatters light in the RPE and choroid, we expect the WFS measurements to
be greatly degraded due to lack of well-defined reference plane. The SLD source was also
used for imaging as well as wavefront measurements. The WFS centroids shown in Fig.
5.5(a) show the scattered light used for the WFS measurements in an albino mouse for
inner retinal imaging compared to the ideal case in a pigmented mouse and the centroids
from the albino mouse are larger and less sharp than the centroids from a pigmented
mouse. Fig. 5.5(b) shows the RMS excluding tip, tilt, and defocus of the measured
wavefront and Fig. 5.5(c) shows the improvement in the image quality metric during the
3-iteration WFS-less AO optimization on the NFL layer in an albino mouse. Defocus was
excluded from the RMS calculation since the imaging light was manually focused on the
inner retina. The optimization that had a 52% increase in the image quality metric.
However, the WFS could not measure the wavefront properly in albino. The attempted
WFS measurements reported an RMS change during the optimization from 0.47 ± 0.02
µm to 0.33 ± 0.01 µm with little response to the changes to the wavefront during the
optimization, especially in the high-order Zernike modes.
Figure 5.5. SH-WFS measurements from an Albino mouse strain (BALB/cJ) retina. (a) The SH-WFS centroids of an albino mouse compared to a pigmented mouse. (b) The RMS of the wavefront measurement without defocus. (c) The image quality metric during WFS-less AO optimization.
Fig. 5.6 shows WFS-less AO images of the blood vessels of the retina. The
optimization was performed on the NFL layer, and then the focus was incremented
through the other vascular layers of the inner retina to image the Nerve Fiber Layer (NFL),
Inner Plexiform Layer (IPL), and Outer Plexiform Layer (OPL). The WFS data was not
61
acquired during this optimization; however, the optimization results demonstrated a 2.2-
fold increase in the image quality after WFS-less AO. The image brightness increased for
NFL, and the image pattern differed among layers, which indicates that the WFS-less AO-
corrected the aberrations to some extent. In contrast, WFS AO was not effective due to
its inability to perform wavefront measurement on the albino fundus.
Figure 5.6. Imaging the inner retinal of an Albino mouse (BALB/cJ) retina with Wavefront Sensorless Adaptive Optics (WFS-less AO). Images of the retina vasculature before and after WFS-less AO in the Nerve Fiber Layer (NFL), and after WFS-less AO in the Plexiform Layer (IPL), and Outer Plexiform Layer (OPL). Scale bar: 10 µm.
5.3.4. AO SLO fluorescence imaging of EGFP microglia cells
Fluorescence SLO imaging was demonstrated with both WFS and WFS-less AO
on EGFP labeled microglia cells (Cx3cr1GFP/+), which were typically found at varying
depths throughout the inner retina. Fig. 5.7(a) shows the WFS AO SLO reflectance image
of the 488 nm excitation light and the simultaneous fluorescence image of a microglia cell
in Fig. 5.7(b). The fluorescence image was superimposed in green on the reflectance
image in magenta in Fig. 5.7(c). Fig. 5.7(d) shows the RMS of the wavefront during the
WFS AO correction calculated with the defocus value excluded since the imaging system
was focused at the inner retina. Fig. 5.7(e) presents the Zernike decomposition of the
wavefront before and after WFS AO to show that measured aberrations are removed.
62
Figure 5.7. Imaging EGFP labelled microglia with Wavefront Sensor Adaptive Optics (WFS AO). (a) Reflectance imaging in the inner retinal blood vessels. (b) Fluorescence imaging of EGFP labelled microglia. (c) The fluorescence image superimposed in green on the reflectance image in magenta. Scale bar: 20 µm. (d) The measured wavefront RMS during WFS AO without defocus. (e) The wavefront measurements in Zernike decomposition before and after the WFS AO aberration correction.
In the next example, the aberration correction was performed at the same region
of the retina twice, first by the SH-WFS AO, followed by WFS-less AO for comparison as
shown in Fig. 5.8. The WFS-less AO was optimized using the fluorescence images of
microglia and started from a system flat. The optimization was performed with a small FOV
(~40 µm across) and had a 3.5-fold improvement in the image quality after two iterations,
as shown in the before and after WFS-less images. Both methods of aberration correction
show a similar AO imaging performance in the over all image quality in Fig. 5.8(a). The
intensity line plot across the WFS and WFS-less AO images in Fig. 5.8(b) show small
differences in sharpness and brightness at the top and the bottom of the images, which
could be due to a small shift in the focal plane due to the WFS-less AO. However, the
overall resolution is similar.
63
Figure 5.8. (a) Imaging EGFP labeled microglia within the inner retina of a mouse with Wavefront Sensor based Adaptive Optics (WFS AO) and Wavefront Sensorless Adaptive Optics (WFS-less AO). Fluorescence image with WFS AO aberration correction (left). Fluorescence image with WFS-less AO aberration correction (middle). Fluorescence images before and after WFS-less AO with a ~40 µm FOV (right). Scale bar: 20 µm. (b) The intensity line plot between the red arrows on the WFS AO image and between the blue arrows on the WFS-less AO image.
In some cases, during mouse retinal imaging, the wavefront aberrations are not
reliably measured for WFS AO as presented in Fig 5.7, so we used WFS-less AO to
provide additional aberration correction. In a representative case, we first used WFS AO
to measure and correct aberrations, which resulted in the left image in Fig. 5.9(a) where
the residual aberrations are apparent by the blurred image. The wavefront measured
before WFS AO as shown in Fig. 5.9(b), then WFS AO decreased the wavefront RMS
from 1.60 ± 0.01 µm to 0.06 ± 0.01 µm before manually shifting the imaging plane to the
microglia in the inner retina. The DM actuator values for the AO correction were loaded
as the starting point for the WFS-less AO to provide further improvements to the image
quality, with the results of the correction shown in the images (middle, right) of Fig.
5.9(a). The measured wavefront RMS (excluding defocus) increased to 0.18 ± 0.01 µm
after shifting the imaging plane and WFS-less AO optimization. However, the aberrations
in the inner retina were not correctly measured, as indicated by improvement in the
image quality despite the measured increase in aberrations. This optimization was
performed with two iterations on a smaller FOV (~80 µm across) and demonstrated that
WFS-less AO can correct for residual aberrations and improve upon WFS AO images
during in vivo retinal imaging sessions.
64
Figure 5.9. (a) Imaging EGFP labeled microglia within the inner retina of a mouse with Wavefront Sensor based Adaptive Optics (WFS AO) and Wavefront Sensorless Adaptive Optics (WFS-less AO). Fluorescence image after WFS AO (left). Fluorescence image after WFS AO and WFS-less AO aberration correction of residual aberration (middle). Fluorescence image with a smaller FOV of microglia dendrites superimposed in green on the reflectance image of the retinal blood vessels in magenta (right). (b) The Zernike decomposition of the wavefront measured before WFS AO and after both methods of AO. Scale bar: 20 µm.
5.4. Discussion
Aberration correction in the living mouse eye presents challenges that include the
relatively thick retina, multiple scattering surfaces, the motion of the sample, and light-
sensitive tissue. This requires a WFS AO system with a fast and densely sampled WFS
and additional system complexity to make wavefront measurements that are suitable for
aberration correction. Alternatively, WFS-less AO can provide image-based aberration
correction at the cost of wavefront optimization time. In this chapter, we presented a
comparison between the imaging performance between the two AO techniques and
demonstrated their trade-offs. Our results include depth-resolved AO-SLO images at
multiple layers of the mouse retina, including reflectance imaging of the photoreceptor
layer of a pigmented mouse and vascular layers of an albino mouse, and fluorescence
imaging of various layers in the inner retina where microglia cells were located. The
wavefront was measured during the WFS-less optimizations to verify that the measured
aberrations from the mouse eye were removed. We also demonstrated that the WFS-less
AO could provide similar image resolution as WFS AO for imaging the rod photoreceptors
in the outer retina and fluorescently labeled microglia cells found in the inner retina of a
living mouse.
65
The WFS AO was operated at 100 Hz and, unlike the WFS-less AO, was not
limited by the imaging frame rate and the aberrations were typically corrected within ~0.25
seconds. The WFS-less approach implemented in this chapter used a coordinate search
(CS) algorithm, requiring ~20 seconds per iteration through 18 Zernike modes. This makes
WFS AO advantageous in the presence of varying aberrations from the sample (in our
case, mostly due to eye movements) as the speed of WFS AO enables continuous
correction. However, WFS-less AO, as a method of correcting static aberrations, is valid
for in vivo mouse retinal imaging since the aberration correction can converge and data
can be acquired before the aberrations change (e.g. 5 minutes). The WFS AO was also
less affected by the amplitudes of the aberrations from the sample, as long as the
aberrations were within the dynamic range of the WFS. In contrast, the sample aberrations
in each case can affect the length and search space required for WFS-less AO. The results
presented in this report showed that WFS-less AO was able to provide similar or slightly
better imaging performance to the WFS-based AO. However, it may be required to search
higher-order Zernike modes to guarantee diffraction-limited performance in all cases [5]
or, alternatively, a more compact search space could be used [98].
The CS approach was used because it is straight forward to understand and
reproduce. Our optimization time could be reduced by calculating the best Zernike mode
amplitude based on fewer measurements [53]. Theoretically, as few as 2n+1
measurements for n Zernike modes could be used [86], but practically the convergence
of our algorithm would be limited by the image noise and motion artifact. Real-time image
registration and tracking on a region of interest could also be used to reduce erroneous
measurements due to the motion of the sample [71]. Other algorithms have also been
demonstrated such as Steepest gradient descent, and simulated annealing [79,80,99].
Alternatively, model-based approaches can theoretically converge even faster such as
‘sphere packing’ described by Booth et al. [100], and the DONE algorithm described by
Verstraete et al. [101]. Also, so-called pupil segmentation approaches have been
demonstrated [20,45]. Several reviews are available on the topic such as [86,102] just to
name a few. However, all of these methods will experience the same challenges to enable
rapid convergence in vivo.
Exclusively using WFS-less AO allows for more flexibility and simplicity when
designing the imaging system, such as our recent report on a compact WFSless AO
system [16]. The lens-based system with WFS-less AO allowed for a larger FOV and could
66
potentially have better peripheral retinal imaging [103]. WFS-less AO also provides
flexibility with the imaging sample since often retinal feature in other layers such as the
large blood vessels can cause strong reflection that confounds the WFS measurement.
Although for both WFS and WFS-less AO, we must also consider the size of the
isoplanatic patch. If our FOV is too large when performing AO, we may only partially
correct for the aberrations.
A comparison can be made between AO for human retinal imaging versus other
applications, such as microscopy of small animal brain imaging in vivo. In the first case,
with a relatively long eye length and thin retina, the focal planes of the WFS beacon and
the imaging plane are nearly the same. In the latter case, backscatter from multiple depths
of the thick tissue layers impedes conventional WFS. The case of mouse retina AO is an
intermediate case between these extremes. Under ideal circumstances for mouse retinal
imaging, WFS works extremely well. When alignment to the mouse eye is near perfect,
and when the beacon and image focal planes are coincident, the convergence is rapid,
and diffraction limited imaging is quickly attained. For the more general case of mouse
retinal imaging, when the region of interest is outside of the mouse eye optical axis, or
when there are features (like blood vessels) that impede WFS measurement, the WFS-
less provides a solution.
The accuracy of WFS-less AO has been investigated by others including
Facomprez et al. [104] for microscopy and demonstrates that WFS-less approaches are
capable of diffraction limited imaging. WFS-less AO has been compared to WFS AO in
microscopy by Bourgenot et al. [105] demonstrating a benefit of an image-based approach
to AO. A comparison of imaging quality between WFS and WFS-less AO has also been
performed for human retinal imaging by Hofer et al. [80]. Although our results demonstrate
a similar image quality between each method in Fig 5.3 (photoreceptor imaging), the
residual wavefront RMS values are different. This may be due to small differences
between the focal planes of the beacon used for the WFS and the imaging plane, or
imaging an area larger than the isoplanatic patch where the aberrations are not uniform
across the entire FOV. Non-common path errors could also be a suspected source of
measurement errors; however, in this system, the phantom imaging in Fig 5.2 indicates
the amount of non-common path aberrations in the system. The phantom images revealed
a 16% improvement in the image quality in favor of the WFS-less AO but, the difference
in wavefront measurements between the WFS AO and WFS-less AO had an RMS of only
67
~0.029 µm (excluding defocus). In the mouse eye, the difference between the wavefront
measurements after each method of AO was typically much larger. For example, the case
presented in Fig. 5.3 had a measurement difference of ~0.15 µm imaging the
photoreceptor layer and the case presented in Fig 5.9 had a measurement difference of
~0.17 µm imaging the inner retina. So, it is likely that the non-common path aberrations
as a source of discrepancy could be neglectable here.
WFS-less AO is advantageous in the optically thick small animal eyes where
aberration correction can be performed at the specific depths where the features of
interest are located, such as the microglia cells that are found at many layers throughout
the inner retina. The outer retinal layers typically provide the strongest scattering plane for
the WFS, and a large defocus is required to image the inner retinal layers. When the WFS
beacon and the imaging light share the same focal position at the inner retina, and the
scattering from the outer retina dominates and thus reduces the WFS spot quality. This is
the case represented in Fig. 5.9, where the WFS-less further improved the imaging quality
after the WFS AO. Another solution is to defocus the beacon relative to the imaging beam,
but this difference could also adversely affect the performance of the WFS.
Albinism in mice and other small animals is a common background phenotype for
transgenic strains. In the retina, the reflectivity characteristics are different in the albino
specimen due to the different amounts of pigments in the retina layers [106,107].
Performing retinal AO on albino mice would be more similar to the case of brain imaging
since there is no good intrinsic guide star for the WFS AO, and alternatives such as WFS-
less need to be considered.We have previously demonstrated that albinism does not affect
the ability of WFS-less OCT [61]. In this chapter, we have demonstrated that WFS-less
AO for SLO also provides the flexibility to include albino mice as samples using depth-
resolved back reflectance. Albino images were not presented with the SH-WFS mode of
operation since a good wavefront measurement is difficult due to enhanced scattering
from the choroid and sclera (due to the lack of pigment in the RPE and choroid).
The imaging applications presented in this chapter span the scenarios of having
the wavefront sensing plane and the imaging plane coincident and separated. For the
case where the WFS-less approach used the fluorescently labeled cells of interest as
guide stars, it is reasonable that the aberration correction would be better based on reports
in the Literature using guide stars, for example [90,108]. Alternatively, in cases where
68
fluorescence guide stars are not available, methods such as coherence-gated wavefront
sensing [24,109] have been shown to provide depth-resolved measurements, but at the
expense of imaging system complexity.
5.5. Summary
In this chapter, we confirm previous reported WFS-less AO aberration correction
for imaging the eye using direct measurements. For retinal imaging in anesthetized and
stable small animals, our results indicate that exclusively using WFS-less methods is
reliable given enough time for the optimization method to find the best correction.
However, when imaging time is limited, WFS-based methods have a large advantage in
achieving the optimal aberration correction, especially for the correction of time-varying
aberrations. The ideal AO imaging for small animal imaging should use both methods in
order to find the best aberration correction across different mouse strains, different retinal
layers and eccentricities, and for different levels of sedation. Finally, our results also
suggest that AO without a well-defined guide star requires the use of WFS-less methods
for the optimal aberration correction.
69
Chapter 6. Multi-modal imaging
6.1. Introduction
The study of small animal models of human diseases causing blindness is
important to understand the mechanisms of vision loss and to develop novel therapies.
Conventional histological approaches require sacrificing the animal at each study time
point. Non-invasive imaging is highly desirable for longitudinal studies, reducing the effects
of inter-animal variation and reducing the number of animals required for a study. There
would be potential benefits and advancements if more researchers had access to high
resolution in vivo imaging systems with the functional and structural detection capabilities
that were previously only attainable through histology [3]. Furthermore, in vivo imaging
allows for the study of physiological processes such as the dynamics of microglia [50,110].
Theoretically, the Numerical Aperture (NA) through the pupil of the mouse eye
permits sub-micrometer imaging of the retina. However, optical aberrations introduced by
the tear film, cornea, and lens reduce the actual resolution. In order to approach diffraction-
limited imaging, these aberrations can be corrected with Adaptive Optics (AO) using a
wavefront corrector such as a Deformable Mirror (DM) [4,5].
AO has been implemented in many ophthalmic imaging modalities such as Optical
Coherence Tomography (OCT), Scanning Laser Ophthalmoscopy (SLO), and fundus
photography, which have been well documented in References [2–4,51–53,111]. The
traditional approach to AO is to use a Wavefront Sensor (WFS) to measure the ocular
aberrations directly. For example, AO SLO has been demonstrated for in vivo imaging
with cellular resolution of Green Fluorescent Protein (GFP) labelled cells [4,14,62].
Performing accurate wavefront measurements for WFS AO imaging in a small animal
retina requires a high level of system complexity due to the short length of the eye creating
an optically thick sample with multiple scattering surfaces [5]. Alternatively, Sensorless
AO (SAO) has the potential to allow for systems to be compact, easily operated, robust,
and inexpensive. SAO does not require direct measurements of the optical wavefront but
instead uses an image-based aberration correction approach, such as a multi-dimensional
optimization or pupil segmentation [20,86]. SAO methods have the ability to provide depth
70
resolved aberration correction by using images acquired at specific layers within the retina.
For example, AO OCT has been demonstrated using en face projections extracted from
three dimensional OCT volumes to drive the optimization algorithm on the selected retinal
layers [61].
The multi-modal system in this chapter was designed to incorporate SAO with
multiple modalities including Optical Coherence Tomography (OCT), OCT based
Angiography (OCT-A), confocal Scanning Laser Ophthalmoscopy (SLO), and
fluorescence detection. In this chapter, we present a compact lens-based design of a
imaging system for multi-purpose imaging of the small animal retina, which has
significantly improved performance and functionality since previous reports
[15,61,112,113]. The en face and cross-sectional imaging enable visualization of the
retinal structure while the fluorescence imaging has the ability to visualize the biological
function of the retina through labelled reporter cells. OCT and SLO can be combined to
employ a multi-modal system for simultaneous and co-localized structural and functional
imaging. We present representative images and analyses to demonstrate the
performance, versatility, and usability of the system for small animal imaging. Images
acquired prior to SAO aberration correction demonstrate the widefield and standard
resolution imaging in a mouse eye. After performing SAO optimization, our results
demonstrate high resolution imaging featuring in vivo volumetric and time-lapse imaging
of fluorescently labelled microglia.
6.2. Methods
6.2.1. Optical design
A schematic of the optical layout of the system is presented in Figure 6.1(a). The
system components were assembled with off-the-shelf optomechanics and custom
mounts designed with SolidWorks (Dassault Systèmes, France) as shown in the Figure
6.1(b). The light sources for the imaging system included a near infrared (NIR)
Superluminescent Diode (SLD, BLM2-D, Superlum Diodes Ltd., Ireland) for OCT using a
central wavelength of 840 nm with a spectral bandwidth of ~80 nm, and a 488 nm laser
(0488L-13A, Integrated Optics, Lithuania) for confocal SLO and fluorescence excitation.
71
The OCT subsystem was based on a fiber optic Michelson interferometer. The
OCT light was split by a 70:30 single mode optical fiber coupler (AC Photonics Inc, CA,
USA). The 70% portion of the light was connected to a reference arm consisting of a fiber
collimator, a dispersion compensation block and a mirror. The OCT probe beam was the
30% portion of light from the coupler, which was launched from a reflective collimator
(RC04FC-F01, Thorlabs Inc., NJ, USA) and transmitted through a cold mirror (ZT670rdc-
xxrxt, Chroma Technology Corp, VT, USA) for combination with the 488 nm SLO light.
In the SLO subsystem, another reflective collimator (RC08FC-F01, Thorlabs Inc.,
NJ, USA) was used to launch the SLO light from a fiber with a polarization controller, such
that the horizontally polarized light was reflected from a Polarization Beam Splitter (PBS,
PBS251, Thorlabs Inc., NJ, USA). The light was then reflected from a dichroic mirror
(ZT405/488/561rpc-UF1, Chroma Technology Corp, VT, USA) to the cold mirror, and then
co-aligned with the OCT light.
The first pupil plane of both subsystems was defined by the Variable Focus Lens
(VFL, Arctic 39N0, Corning, NY, USA) with an aperture of 3.9 mm. The imaging beams
were relayed and magnified to a continuous membrane DM (DM69, Alpao, France) with a
10.5 mm aperture, and then to a mounted pair of Galvanometer-scanning Mirrors (GM,
6210H, Cambridge Technology Inc., MA, USA) with a clear aperture of 3.0 mm. Finally,
the light was reduced to a beam diameter of 1.0 mm to be focused by the mouse eye and
relayed from the GM to be scanned across the retina with a maximum scanning angle of
50 degrees. The optical relays were constructed using achromatic doublets with an
antireflection coating for both visible and near infrared light (VIS-NIR Coated Achromatic
Lenses, Edmund Optics). Each relay used two off-the-shelf lenses, except the final relay
to the mouse eye. See L5 and L6 in Figure 6.1. These elements were constructed from
two achromatic lenses that were placed symmetrically with a <1 mm air gap [114,115]
This design enabled shorter optical relay required for our desired scanning angles without
introducing significant aberrations.
72
Figure 6.1. (a) Schematic of Optical Coherence Tomography (OCT) and confocal Scanning Laser Ophthalmoscopy (SLO) system. The cyan represents the beam path of only 488 nm light, the green represents the beam path of only the fluorescence emission and the red represents the beam path of only the SLD light. The pink represents the co-aligned beam path of the 488 nm light, fluorescence emission, and SLD light. System components: Superluminescent diode (SLD), fiber coupler (FC), polarization controller (PC), polarization beam splitter (PBS), dichroic mirror (DC), emission filter (EF), cold mirror (CM), variable focus lens (VFL), deformable mirror (DM), galvanometer-scanning mirrors (GM), quarter wave plate (QWP), photomultiplier tube (PMT), dispersion compensation block (DCB), mirror (M). Achromatic doublet lenses: L1=50mm, L2=150mm, L3=300mm, L4=75mm, L5=2x125mm, L6=2x50mm. (b) Computer simulation of optical layout on custom optical mounts using OpticStudio and SolidWorks.
73
The final element before the mouse eye was a Quarter Wave Plate (QWP,
WPQ10E-488, Thorlabs Inc., NJ, USA), which rotated the polarization state of the back-
scattered SLO light from the sample so that the light would be transmitted by the PBS for
detection [15,56]. Although this technique does not entirely remove the strongest reflection
in the center of the back-scattered SLO images (see the Results section), this method
does remove other reflections from the optical elements. The optical design was simulated
in OpticStudio (Zemax, WA, USA) and Figure 6.2(a) presents the spot diagrams with an
ideal model eye. The expected resolution and calculations were performed for mouse eyes
using an NA of 0.25. The OCT and SLO spot diagrams are presented across a 15-degree
(~500 µm for mouse eyes) Field of View (FOV) with 0 D of vergence at the sample pupil
plane. The black circle in the top row represents the Airy disk with a ~2.1 µm radius for
820 nm, 840 nm, and 860 nm. The middle and bottom row have a 1.2 µm Airy disk radius
for 488 nm SLO light. The bottom row represents the 488 nm spots scanned across 7
degrees (~230 µm for mouse eyes) with a 20 D vergence at the pupil plane for the eye
produced by the simulated VFL. These FOVs (or smaller) are typically used for AO in the
mouse eye [4], whereas for imaging a larger FOV, it may not be necessary to have a spot
size on the order of microns. The axial resolution for OCT was estimated to be ~3 µm. For
the SLO axial resolution, the FWHM of the axial point spread function was ~18 µm. For
the OCT-A imaging, the system was reconfigured to have a smaller beam into the eye,
which reduced the NA to 0.15 and the Airy disk radius to 3.5 µm.
In order to reduce the overall size and simplicity of the system, we did not use an
optical relay between the fast and slow scanning mirrors, which prohibited perfect
conjugation to the pupil of the mouse eye with both scan axes. In Figure 6.2(b), the amount
of pupil wander in the mouse eye was simulated for the FOVs that are suitable for
diffraction-limited imaging.
74
Figure 6.2. (a) Spot diagrams of the OCT light at 820 nm (red), 840 nm (pink) and 860 nm (purple) across a 15-degree FOV, where the black circle represents the Airy disk with a 2.1 µm radius. Spot diagrams of the 488 nm (blue) SLO light scanned across a 15-degree FOV with 0 D of vergence at the sample pupil plane and 7-degress with 20 D of vergence at the sample pupil plane where the black circle represents the Airy disk with a 1.2 µm radius. (b) The boundary of the imaging beam at the final pupil plane of the system. The black circle represents a 2 mm aperture. Each color represents a different scan position across a 15-degree and 7-degree FOV to simulate the pupil wander due to the space between the scanning mirrors in the optical design.
The back-scattered OCT light from the sample was recombined with the reference
arm light at the fiber coupler and directed to a spectrometer (Cobra 800, Wasatch
Photonics, NC, USA). The A-scans were acquired with a frame grabber (PCIe-1433,
National Instrument, Austin, TX) at 100 kHz and the OCT volumes were sampled at 1024
x 400 x 200 points. For OCT-A, two B-scans were acquired at the same location in the
slow scan direction to calculate changes due to blood flow. The OCT/OCT-A cross-section
and en face view were processed for real-time display using our custom GPU accelerated
acquisition software [89,112,113] written in C/C++.
The fluorescence emission was transmitted through the dichroic mirror, and the
emission filter (ZET405/488/561m-TRF, Chroma Technology Corp, VT, USA), and then
focused into a multimode fiber with a core diameter of ~2 Airy disk diameters (ADD) that
directed the light to a photo-multiplier tube (PMT, H10723-20, Hamamatsu Photonics K.
K., Japan). The back-scattered 488 nm laser light was reflected from the dichroic mirror,
transmitted through the PBS, focused into a multimode fiber with a core diameter ~5 ADD
or ~20 ADD, and detected by another PMT (H7827-002, Hamamatsu Photonics K. K.,
Japan). We used a 5 ADD fiber core when performing image-based SAO with the back-
75
scattered images, or else we used a 20 ADD fiber core, which provided the higher SNR
for general navigation on the mouse retina. The PMT signal gain could be adjusted on the
power supply depending on the amount of signal from the sample. The digitization (PCIe-
6361, National Instrument, Austin, TX) of the PMTs was synchronized to the acquisition
of the OCT A-scans for simultaneous imaging, otherwise the SLO could be operated alone
at a 2 kHz line rate with a sampling density of 400 x 200 points.
6.2.2. Sensorless adaptive optics
The SAO could be performed on the en face projection of the OCT volumes, the
back-scattered confocal SLO, or the fluorescence SLO images. We implemented a hill
climbing Coordinate Search (CS) algorithm presented in Chapter 5, which provided an
exhaustive search to find the optimal Zernike coefficients. The merit function for
optimization was determined by the highest image sharpness (Simg) [94,95], defined by
the sum of the intensity squared of each image pixel I(x,y) in Equation 6.1.
𝑆𝑖𝑚𝑔 = ∑[𝐼(𝑥, 𝑦)]2
𝑥,𝑦
, (6.1)
The CS algorithm started with a flat wavefront with an RMS ~0.05 µm, which was
calibrated using a SH-WFS in the location of the GM scanners. Then, for the first mode
(𝑘) in a sequence, a range of coefficients (±α) was applied to the DM. The coefficient (𝑎𝑛∗ )
that corresponded to highest metric value was applied to the DM and the algorithm moved
onto the next mode. For the first iteration, the sequence of modes began with a defocus
(k = 4) value, then the astigmatisms, and continuing in ascending order up to the 21st mode
for a total of 18 modes. The Zernike polynomials were ordered and reported using the
mode number according to the OSA/ANSI standard [68]. The sequence of 18 modes was
usually repeated for multiple iterations, typically 2 or 3 times, until the metric value no
longer increased. Successive iterations would search coefficients ranges (±β) around the
previously selected coefficients. Between iterations, the imaging FOV or location could be
adjusted, as the features of interest became visible.
For high resolution imaging, SAO could be used to correct wavefront aberrations
from the mouse eye using the output from of the different imaging modalities for the image-
based optimization. During optimization, the sampling density of the OCT was decreased
76
to 1024 x 400 x 20 which resulted in ~20 seconds for each iteration of the optimization.
When the SLO was used for optimization, the sampling was set to 400 x 100 and each
iteration took a total of 12 seconds.
6.2.3. Animal handling
The animal imaging sessions were performed under protocols compliant to the
Canadian Council on Animal Care and the approval of the University Animal Care
Committee at Simon Fraser University. The mice were anesthetized with a subcutaneous
injection of ketamine (100 mg/kg of body weight) and dexmedetomidine (0.1 mg/kg of
body weight). A drop of topical solution (Tropicamide, 1%) was applied to dilate the ocular
pupils. A rigid 0-Diopter contact lens was placed on the animal eyes to prevent the cornea
from dehydration and then the animal was aligned without any further contact to the
imaging system [19]. For fluorescein angiography, the mice were subcutaneously injected
with 100 µL of 100 mg/mL fluorescein. Mice were purchased from the Jackson Laboratory,
including wild type strain (C57BL/6J) and transgenic strain with Enhanced Green
Fluorescent Protein (EGFP) labelled retinal ganglion cells (Tg(Thy1-EGFP)MJrs/J) and
microglia (B6.129P-Cx3cr1tm1Litt/J).
For retinal imaging, the OCT imaging light was limited to 620 µW. The SLO imaging
light did not exceed 230 µW in this chapter and was limited to 100 µW when operating
simultaneously with the OCT.
6.2.4. Image processing
Images in this chapter were generated by standard post-processing techniques,
including steps to register and align frames to a template for averaging, using a
combination of Matlab (MathWorks Inc, MA, USA) and ImageJ (National Institutes of
Health (NIH), MD, USA) toolkits. The number of volumes and frames that were saved
could be easily changed in the acquisition software. For the images presented in this
chapter, we used the following parameters: for OCT, we recorded 5 volumes per
acquisition in 4 seconds; for OCT-A images, only one volume was recorded per acquisition
in 1.6 seconds; and for SLO, we recorded 50 to 100 frames per acquisition in 5 to 10
seconds. The OCT B-scans were aligned with a vertical translation to remove axial motion
77
of the animal. Most of the B-scans presented in this chapter were an average of 5 adjacent
B-scans within one volume with an exception that is explained in the results section.
The en face OCT images were generated using a Maximum Intensity Projection
(MIP) between two manually selected horizontal lines corresponding to depths in the
retina. Then, the en face OCT projections and the SLO frames were processed with the
following procedure: 1) The registration process was initialized by manually selecting a
single frame as the template to align the other frames; 2) Each frame was globally
translated horizontally and vertically to maximize the cross-correlation with the template;
3) The frames were broken up into horizontal strips and each strip was translated
horizontally and vertically to maximize the cross-correlation with the template [70,97]; 4)
The frames were non-rigidly aligned to the template with a sum of squared differences
similarity metric along cubic B-splines using the Medical Image Registration Toolbox
(MIRT) [69]; 5) After registration, the frames were averaged and the black and white
thresholds were adjusted to enhance the image brightness and contrast for presentation.
All the B-scans in this chapter are presented in a linear intensity scale; 6) The images
were scaled so that the vertical and horizontal dimensions have the same scale.
SLO frames from the back-scattered and fluorescence channels were
simultaneously acquired, which would allow for co-registration if the fluorescence signal
was insufficient [24]. However, in this chapter, the fluorescence images had sufficient
signal to use directly for registration.
6.3. Results
6.3.1. Imaging without adaptive optics
For imaging large retinal features, a widefield image is preferred and it may not be
necessary to perform SAO. Figure 6.3(a) demonstrates a 50-degree OCT B-scan and a
44-degree en face projection of the Outer Plexiform Layer (OPL) of a wild type mouse
retina. Unlike the other B-scan images in this chapter, in Figure 6.3(a), the vertical
scanning was disabled and 200 B-scans were acquired, aligned, and averaged. In Figure
6.3(b) and (c), the sampling density is increased with a 22-degree FOV and the focus was
shifted with the VFL from the OPL to the Nerve Fiber Layer (NFL). The B-scans and en
78
face images were registered and averaged as described in the previous section. The
location of the B-scan is indicated by the red dashed line.
Figure 6.3. (a) OCT B-scan across 50 degrees in the mouse retina and en face projection of the outer plexiform layer (OPL) across 44 degrees. The B-scan is an average of 200 consecutively acquired cross-sectional frames and the en face OCT image is an average of 5 frames. (b,c) Average of 5 adjacent OCT B-scans and an average of 5 en face OCT frames of the OPL. The B-scans are located at the position of the red dashed lines. Vertical scale bar: 50 µm. Horizontal scale bars: 100 µm.
After acquiring OCT volumes, the OCT was disabled in the software so that the
SLO could be operated at a faster speed. The structural SLO image in Figure 6.4(a) was
generated by the 488 nm back-scattered channel light from 50 averaged frames with the
imaging light focused on the NFL and a 5 ADD confocal aperture. Fluorescence SLO
images were generated from an average of 50 frames acquired a few minutes after a
fluorescein injection. Images were acquired from three different vascular layers in the inner
retina, including the NFL, inner plexiform layer (IPL), and OPL. These images were
combined with a MIP for presentation in Figure 6.4(b), which is demonstrated in further
detail in the following results with AO.
79
Figure 6.4. Confocal SLO images of a mouse retina with 488 nm light. (a) Structural image of the nerve fiber layer from back-scattering. (b) Fluorescein angiography composited with a MIP from images of three different vascular layers. Scale bar: 100 µm.
OCT-A B-scans were created by calculating the difference between two intensity
B-scans in the same location. Figure 6.5(a) shows the en face OCT-A image generated
from the MIP of the OPL layer in the B-scans of a single volume. For comparison, Figure
6.5(b) shows the en face OCT intensity image that was generated from the same OPL
region. Figure 6.5(c) was created by coloring the en face OCT-A images that were
extracted from the OPL with red, the IPL with green, and the NFL with blue.
Figure 6.5. (a) En face OCT-A images of the OPL in a mouse retina. (b) En face OCT intensity image from the same image data. (c) En face OCT-A images that were generated from the OPL (red), IPL (green), and NFL (blue). Scale bar: 50 µm.
6.3.2. Structural imaging with sensorless adaptive optics OCT and SLO
For SAO OCT, the retinal layer of interest was selected and the image quality
metric was calculated on the en face image to drive the optimization. Figure 6.6 represents
an example of an imaging sequence. The imaging plane was focused on the OPL layer
with a FOV of ~250 µm. En face OCT images were used for optimization, and then OCT
80
volumes were acquired for presentation. Figure 6.6(a) shows en face OCT images before
and after SAO at different focal planes. Figure 6.6(b) demonstrates the two-iteration
optimization with a plot of the image quality metric for each step in the optimization and
the coefficients selected for each iteration. Overall, there was a 1.9-fold improvement in
the image quality metric reported from the merit function of the optimization algorithm.
Figure 6.6. (a) En face images of the outer plexiform layer (OPL, top row, ~250 µm FOV) and nerve fiber layer (NFL, bottom row, ~280 µm FOV) retinal layers before and after Sensorless Adaptive Optics (SAO). SAO-OCT B-scans with the imaging focal plane on the OPL (red arrows) and NFL (blue arrows). (b) The normalized image quality for each step in the SAO optimization over two iterations and the Zernike coefficients selected for each iteration. Vertical scale bars: 50 µm. Horizontal scale bars: 20 µm.
Similarly, for SAO SLO, a typical imaging procedure is presented in Figure 6.7.
The confocal pinhole was 5 ADD for the structural SLO images in Figure 6.7(a). In this
case, the imaging light was focused on the NFL layer and the back-scattered SLO images
were used for the optimization metric. Figure 6.7(a) shows the averaged SLO images
before SAO and after SAO images of the NFL. Then the focus was shifted with the VFL
to image different retinal layers including the OPL as shown. Figure 6.7(b) demonstrates
the improvement in the image quality metric during the optimization. With a FOV of ~250
µm, the first iteration improved the image quality 2.2-fold. Then the FOV was reduced by
81
half in the second iteration, which further improved the image quality 1.6-fold. The FOV
could be changed for the successive iterations because new image quality values would
be determined each iteration. Therefore, the plot of the image quality metric over the entire
optimization had to be normalized independently for the metric values in each of these
iterations. Figure 6.7(b) also presents the coefficients selected for each iteration.
Figure 6.7. (a) Confocal SLO images before and after Sensorless Adaptive Optics (SAO) of the nerve fiber layer (NFL) with a FOV ~250 µm. Images of the outer plexiform layer (OPL) after SAO. (b) The normalized image quality metric values for each step used for the SAO optimization for each iteration. The Zernike coefficients selected for each iteration. Scale bar: 20 µm.
For both SAO OCT (Figure 6.6) and SAO SLO (Figure 6.7), dark circular “holes”
are revealed in between the nerve fiber bundles after aberration correction similar to other
AO-SLO images [4], which are speculated to be retinal ganglion cell soma due to the size
and shape.
6.3.3. Fluorescence imaging with sensorless adaptive optics
The ability to image EGFP labelled cells with the SAO SLO fluorescence detection
further increases the functionality of the imaging system. The results in this section
demonstrate the SAO SLO image quality and the abilities of the fluorescence detection
82
channel. The structural confocal SLO images in this section were acquired with the
detection pinhole ~20 ADD.
For Figure 6.8, a larger FOV (~750 µm) was used to locate the EGFP labelled
Retinal Ganglion Cell (RGC). The imaging FOV was reduced to ~250 µm to perform SAO
on the fluorescence imaging channel, then a second iteration was performed on a further
smaller FOV ~100 µm with only dendrites of the RGCs in view. Figure 6.8(a) presents a
comparison of the images acquired before and after SAO, including a line plot across
between the blue (before SAO) and red arrows (after SAO). Figure 6.8(b) presents the
structural images that were acquired when the imaging plane was focused on the RGC
axon. In the right column, the structural image was colored in magenta and the
fluorescence image was overlaid in green in order to better localize the RGC.
Figure 6.8. Confocal SLO images of a mouse retina with labelled retinal ganglion cells (Tg(Thy1-EGFP)MJrs/J). (a) Fluorescence images before and after Sensorless Adaptive Optics (SAO) and an intensity line plot between the blue arrows (before SAO) and red arrows (after SAO). (b) The left column presents structural images focused on the nerve fiber layer at a ~750 µm FOV (top) and ~230 µm FOV (bottom). The right column presents the structural image in magenta overlaid by the fluorescence image in green. The fluorescence image was composited from two different focal planes for the axon and the dendrites of the RGC. Scale bars: 50 µm.
We performed fluorescein angiography to demonstrate the confocal capability to
discriminate different layers in the inner retina. Figure 6.9 presents SAO images of three
distinct vascular layers, including the NFL, IPL and OPL. The images were composited
83
using a MIP, with the NFL in red, IPL in green, and OPL in blue to show the vessel
connections in the axial direction.
Figure 6.9. Confocal SLO fluorescein angiography of a mouse retinal vasculature after Sensorless Adaptive Optics. Images (left to right) of the nerve fiber layer (NFL), inner plexiform layer (IPL), outer plexiform layer (OPL), and the MIP with the NFL in red, IPL in green, and NFL in blue. Scale bar: 50 µm.
To demonstrate the volumetric imaging ability of the system, we imaged a mouse
with EGFP labelled microglia, which are located in many retinal layers. SLO images were
acquired at 18 different focal positions between the OPL and NFL layer, and the axial
location of the fluorescence was determined by the structural images. The depth fly-
through of the back-scattered SLO images with the co-localized fluorescence SLO image
is presented in Visualization 1 of reference [16]. Figure 6.10 presents images from the
sequence, where the two right-most images are the structural and fluorescence images
from the NFL layer. The fluorescence image in the middle-left (Fig. 6.10) was located
immediately below the NFL layer and the image on the far-right was located deeper into
the retina at the OPL. The images were color-coded in depth from the OPL to the NFL and
presented with 3D shadowing that was rendered by ImageJ in Visualization 2 of reference
[16].
84
Figure 6.10. Confocal SLO images with Sensorless Adaptive Optics of EGFP labelled microglia in the mouse retina (B6.129P-Cx3cr1{tm1Litt}/J) acquired at different focal position between the outer plexiform layer (OPL) and the nerve fiber layer (NFL) selected from Visualization 1 of reference [16]. The microglia images were color-coded in depth between the OPL and the NFL of the retina and rendered in 3D for Visualization 2 reference [16]. Scale bar: 20 µm.
Microglia are known to constantly survey the surrounding environment and time-
lapse imaging can reveal the dynamics of the cellular branches [110,116]. The microglia
in Figure 6.11(a) were located just below the NFL and these images were selected from
a 1-hour time-lapse video that acquired images in 20 second intervals. The SAO was
performed periodically throughout the imaging to ensure optimal image quality. For each
optimization, the FOV was reduced to 52 µm across, containing only the microglia
branches. The image in Figure 6.11(b) was annotated and color-coded at these time
points to highlight areas of growth and retraction. Visualization 3 in reference [16] presents
the entire time-lapse with the time stamp of acquisition.
85
Figure 6.11. (a) Confocal SLO fluorescence images with Sensorless Adaptive Optics of EGFP labelled microglia in the mouse retina (B6.129P-Cx3cr1{tm1Litt}/J) from three time points in the time-lapse video from Visualization 3 reference [16]. (b) The microglia images color-coded with time. The white arrows 1-4 note areas of significant growth and retraction. Scale bar: 20 µm.
The central microglia in Figure 6.11(b) had a branch (white arrow #1) that retracted 24
µm, with an average velocity of 4.8 µm/min from minute 3 to minute 8, and a branch (white
arrow #2) that retracted 38 µm, with an average velocity of 1.3 µm/min during minute 19
to minute 49. The microglia branch on the right of the image (white arrow #3) generally
retracted over 50 minutes but also had periods of extension during that time. The microglia
branch on the left (white arrow #4) appears to move towards another microglia branch
(white arrow #2) at minute 24.
We performed further time-lapse imaging of microglia using the same methods for
SAO in order to investigate the potential effect of the imaging light. The 488 nm imaging
light was reduced to 100 µW for 39 minutes, then the exposure was increased to 230 µW
and imaging proceeded for another 50 minutes, as shown in Visualization 4 of reference
[16]. Figure 6.12(a) shows time-points before and after the laser power was increased.
The image in Figure 6.12(b) was annotated and color-coded image at these time points to
highlight areas of growth and retraction.
86
Figure 6.12. (a) Confocal SLO fluorescence images with Sensorless Adaptive Optics of EGFP labelled microglia in the mouse retina (B6.129P-Cx3cr1{tm1Litt}/J) from three time points in the time-lapse video from Visualization 4 of reference [16] with an increase in laser power at 39 minutes. (b) The microglia images color-coded with time. The white arrows 1-2 note areas of significant growth and retraction. Scale bar: 20 µm.
6.4. Discussion
In this chapter, we have demonstrated a multi-modal en face imaging system with
diverse functionality for vision scientists needing a variety of imaging requirements. The
system imaging modalities include en face OCT, OCT-A, SAO OCT, as well as SLO and
SAO SLO with fluorescence detection. Our system uses lens-based optical relays
between the active elements, which include the VFL, the DM, and the GMs. Our results
demonstrate state-of-the-art AO imaging for the mouse retina and represent improvement
from our previous reported systems for each individual modality. Our results demonstrated
the variety of in vivo imaging abilities that included structural imaging, angiography,
volumetric and time-lapse imaging of microglia cells.
The imaging system primarily used an NA of 0.25 into the mouse eye, which only
represents about half of the theoretical maximum. However, with 488 nm light, this still
has a calculated resolution of ~1 µm. For the purpose of this chapter, the image quality
87
was sufficient for clearly imaging the microglia branches and to report metrics, such as
movement speeds, yet maintains good quality imaging without requiring AO for imaging
large features. The system was initially designed and tested for mouse imaging; however,
it is also capable of imaging the rat retina as well, which is often required by many vision
researchers for longitudinal studies [117]. Since the rat eye is larger than the mouse eye,
this decreases the maximum attainable resolution. However, it was still beneficial to have
the SAO to correct for aberrations.
During the time-lapse imaging of microglia cells, we only illuminated the retina with
488 nm since we did not require the use of a beacon for WFS measurements. The
microglia time-lapse in Figure 6.11 appears to have more retraction than the microglia
time-lapse from Figure 6.12, despite the increase in laser intensity. It is possible that this
was normal microglia surveillance of a healthy retina or a response to the 488 nm imaging
light. If the 488 nm imaging light itself has an effect on microglia, then it may be difficult to
conclude the reason for a microglia response when investigating their role in immunity
studies. There is no established maximum permissible exposure (MPE) for the mouse
eyes; however, other groups have scaled the MPE for SLO in human eyes [4,64,97]. The
MPE for human SLO imaging decreases with imaging FOV [118], so as we image small
features in small animal experiments, it will be important to continue to consider laser
irradiance as a potential factor.
The imaging system was designed to be used by a non-specialist and future
improvements could improve the reliability and robustness of the SAO. For example, a
Region of Interest (ROI) within the display could be selected by the user instead of
reducing the entire imaging FOV, which further increases the exposure during the ~10 to
20 seconds required for the optimization iteration. Real-time image tracking on the ROI
would also enable the optimization algorithm to follow an object of interest or reject frames
with a large amount of motion artifact [71,85,119]. In this chapter, we were using a multi-
iteration exhaustive search, which was robust to the occasional motion artifact over the
~30 to 60 seconds required for the entire optimization. However, accurate image tracking
would encourage the use of faster optimization algorithms, such as model-based
approaches [100,101] that require much fewer measurements, thereby decreasing
optimization time. This would be advantageous to further reduce the exposure of the entire
imaging process and the potential for damage over time.
88
In this chapter, we optimized up to the 21st Zernike mode for 18 modes in total.
The improvement in the image quality after each mode is optimized is represented in
Figure 6.7(b), which demonstrates that there is an increase in the metric value in the 5th
radial order in the first and second iteration. Using higher orders in the optimization
algorithm could improve results but it would come at the cost of algorithm time. Since time
is limited for in vivo imaging, the algorithmic execution time is better spent on further
iterations [120]. For example, the step sizes between coefficients can be reduced to
improve the wavefront correction. Furthermore, successive iterations have an improved
SNR, which will also improve the performance of the AO correction.
6.5. Summary
In conclusion, we have demonstrated a lens-based system, capable of high-
resolution en face small animal imaging with multiple modalities. The compactness and
simplicity of the system allow for the potential translation to vision scientists that require
tools for in vivo and longitudinal studies. Our results demonstrate the potential for studying
individual cells, such as RGCs and microglia, in healthy and diseased animal models.
89
Chapter 7. Non-invasive cellular-resolution imaging of the retina with two-photon excited fluorescence
7.1. Introduction
Non-invasive retinal imaging is a valuable tool that is used in both clinical and
preclinical vision research to aid the development of novel therapies for preventing
irreversible vision loss. More sensitive assessment of the physiological and biochemical
processes within the retina could be used to detect earlier signs of disease in order to
preserve sight [12,121]. Fluorescence can be used to image many biomarkers, since
fluorophores can be added to the retina to provide contrast or fluorophores that are
intrinsic to the retina can be used for measurements. For example, measuring
autofluorescence from the Retinal Pigment Epithelium (RPE) is of high interest for
investigating diseases such as Age-Related Macular Degeneration (AMD) and Stargardt
disease [3,17,122].
For many fluorophores intrinsic to the retina, the single-photon excitation is in the
ultraviolet (UV) range and the fluorescence cannot be excited through the eye of many
species due to the ocular transmission window [65,81]. Imaging the retina non-invasively
with Two-Photon Excited Fluorescence (TPEF) could enable novel in vivo studies of
disease and retinal physiology [65,81,123,124]. Furthermore, the multiphoton process
suppresses out-of-focus background signal, which improves the axial sectioning without
a confocal aperture in the optical detection path. Imaging the retina with near-infrared
(NIR) light has advantages since the retina is less sensitive to NIR than visible light. NIR
light is also less scattered within biological tissue than the equivalent visible light required
to excite the same fluorophores [37,38].
The difficulty of retinal TPEF imaging in vivo is that high energy is typically required
to generate the TPEF, while minimizing the incident exposure energy is required for the
imaging to be non-invasive. An active area of research is the development of technology
that reduces the average laser power required for TPEF imaging in the eye [12,97,122].
The TPEF signal intensity is highly sensitive to the focused spot size, which leads to high
90
sensitivity to aberrations [38]. It has been demonstrated that the TPEF signal for retinal
imaging can be increased by using Adaptive Optics (AO) for aberration correction
[24,97,123].
For high-resolution imaging of the mouse retina, optical aberrations introduced by
the eye must be corrected for diffraction-limited imaging. AO has been demonstrated to
correct ocular aberrations extensively [3,4,13,39]. The traditional approach to AO is to use
a Wavefront Sensor (WFS) to directly measure the aberrations. Alternatively, it has also
been demonstrated that Sensorless AO (SAO) can provide depth-resolved aberration
correction using an image-based approach. SAO has the advantage of avoiding the
system complexities that are required due to the short length of the mouse eye, which
creates an optically thick sample with multiple scattering surfaces [5].
In this chapter, we present a multi-modal imaging system that uses SAO and
Optical Coherence Tomography (OCT) to achieve non-invasive TPEF imaging of cellular
features in the retina. We used the same light source to simultaneously generate OCT
and TPEF, where both imaging modalities could be used to drive the SAO optimization
algorithm. We demonstrate the high-resolution OCT/TPEF system with a variety of
samples, which included fluorescein angiography (FA) and fluorescently labelled cells. We
also demonstrate the ability to non-invasively image intrinsic fluorescence from the RPE
of the mouse retina in various strains.
7.2. Methods
7.2.1. System setup
The SAO OCT TPEF imaging system used a femtosecond pulsed laser (Mai Tai
HP, Spectra-Physics, CA, USA) for both the OCT and TPEF excitation. The laser had a
tuning range from 690 nm to 1040 nm, where the central wavelength was chosen for each
of the sample fluorophores. Table 7.1 summarizes the calculations to estimate the axial
resolution of the OCT in tissue [125] and the maximum laser power used for each
fluorophore imaged in this report. The laser power could be adjusted with neutral density
filters. A schematic of the imaging system is presented in Figure 7.1.
91
Figure 7.1. Schematic of the Sensorless Adaptive Optics (SAO) Optical Coherence Tomography (OCT) and Two-Photon Excitation Fluorescence (TPEF) imaging system. The imaging system was constructed with a pellicle beam splitter (PeBS), a variable focus lens (VFL), a deformable mirror (DM), a dichroic mirror (DcM), galvanometer-scanning mirrors (GM), emission filters (EF), a photo-multiplier tube (PMT), dispersion compensation (DC), and the following lenses: L1=100 mm, L2=300 mm, L3=400 mm, L4=100 mm, L5=2×125 mm, L6=2×50 mm. The reference arm denoted as a dashed line.
For the OCT/TPEF system, we used a pellicle beam splitter to separate the light
into the sample arm and the reference arm. In the sample arm, we used a variable focus
lens (VFL, A-39N1, Corning, NY, USA) to control the focal plane within the sample, a
continuous membrane deformable mirror (DM, DM-69, ALPAO, France) to correct the
wavefront aberrations, and XY mounted pair of galvanometer-scanning mirrors (GM,
6215H, Cambridge Technology Inc., MA, USA) to scan the light across the sample. The
scanning angles of the GMs could be adjusted to change the imaging field of view (FOV)
on the retina and the maximum FOV for this system was ~25 degrees (~850 µm). Three
lens-based telescopes were used to optically relay the pupil planes at the VFL, DM, and
GMs to the mouse eye. The telescopes were constructed with achromatic doublets that
had focal lengths listed in Figure 7.1. The imaging beam entered the mouse eye with an
NA of ~0.25. Table 7.1 summarizes the calculations to estimate the spot size in tissue
using the Airy disk radius and the FWHM of the axial point spread function for each center
wavelength [26]. The reference arm included dispersion compensating glass to match the
sample arm.
92
Table 7.1. Laser specifications used for each fluorescent sample and the calculated resolution.
Fluorescent Sample
Center Wavelength
(nm)
Spectral Bandwidth
(nm)
Laser Power (mW)
Lateral Spot Size (µm)
TPEF Axial
Resolution (µm)
OCT Axial Resolution
(µm)
FA 800 13 < 3 2.0 29 16 GFP 910 16 < 9 2.2 33 17 YFP 940 15 < 3 2.3 34 19 RPE 740 8.2 < 8 1.8 26 22
For OCT, back-scattered light from the eye and the light returning from the
reference arm were combined at the PeBS and coupled into a single mode fiber. The
spectral interference signal was then detected by a custom-built spectrometer (1024
pixels, 100 kHz, Bioptigen Inc., NC, USA) with a spectral range from ~730 nm to ~995 nm.
The data was acquired through a Camera Link frame grabber board (PCIe-1433, National
Instrument, Austin, TX) and processed using a custom GPU-accelerated program for real-
time processing and display [89]. The two-dimensional transverse area (en face) was
scanned and sampled with 400 × 200 A-scans, which resulted in an acquisition rate of 1
volume per second. For OCT-based Angiography (OCTA), we used two intensity B-scans
(two BM-scans) to calculate the angiography at 400 × 200 (200 × 2) A-scans. During the
SAO-OCT optimization process, the density in the direction of the slow scan was reduced
to 400 × 20 A-scans for aberration correction using the en face OCT images.
TPEF emission from the sample was de-scanned by the GMs and reflected by the
dichroic mirror (DcM) to the photo-multiplier tube (PMT, H7422P-40, Hamamatsu
Photonics, Japan) detector. The short-pass filters (FF01-650/SP-25, FF01-720/SP-25,
IDEX Health & Science LLC, NY, USA), a lens, and an aperture were positioned before
the PMT to reject stray light. The electric current from the PMT was converted to a voltage
with a transimpedance amplifier (LCA-400K-10M, FEMTO Messtechnik GmbH, Germany)
and digitized by a DAQ device (PCIe-6361, National Instrument, TX, USA). The digitization
of the PMT signal was synchronized with the OCT A-scans for simultaneous operations,
which ensured that both OCT and TPEF images were co-registered. The TPEF could also
be operated without the OCT to acquire at 10 frames per second for averaging in post-
processing.
The SAO optimization used a hill-climbing coordinate-search algorithm, which was
recently reported for mouse retinal imaging [16,18]. In brief, the optimization used the merit
function defined by the image sharpness of the en face OCT images that were extracted
93
at the user-selected depth or the TPEF images. The algorithm began by finding the best
initial defocus value, then the astigmatisms, and then continuing in ascending order for a
total of 18 modes. Multiple iterations could be performed to ensure optimal the aberration
correction. The optimization required ~20 seconds per iteration using either the OCT or
the TPEF images.
7.2.2. Animal handling and image processing
The mouse imaging was performed under protocols compliant to the Canadian
Council on Animal Care and the approval of the University Animal Care Committee at
Simon Fraser University. A subcutaneous injection of ketamine (100 mg/kg of body
weight) and dexmedetomidine (0.1 mg/kg of body weight) was used to anesthetize the
mouse prior to imaging. The pupils of the mouse were dilated with a drop of topical solution
(Tropicamide, 1%). A rigid 0-Diopter contact lens was placed on the mouse eyes to
prevent dehydration of the cornea. Then, the animal was aligned to the imaging system
without any contact [19]. All of the mice imaged in this work were purchased from The
Jackson Laboratory, ME, USA, which included B6 mice (C57BL/6J), albino B6 mice
(B6(Cg)-Tyrc-2J/J), mice with GFP labelled microglia (B6.129P-Cx3cr1tm1Litt/J), mice with
GFP labelled retinal ganglion cells (Tg(Thy1-EGFP)MJrs/J), mice with YFP labelled neural
cells (B6.Cg-Tg(Thy1-YFP)16Jrs/J), and mice with a mutated rpe65 gene (B6(A)-
Rpe65rd12/J). Table 7.2 summarizes the that mice used for the imaging presented in this
report. In addition, all the mice were female and weighed 25 – 35 grams. For fluorescence
angiography, the mice were anesthetized and then subcutaneously injected with 100 μL
of 100 mg/mL fluorescein.
Table 7.2. Summary of mice that were used in this report.
Mouse Strain Stock Number Pigmentation Number
C57BL/6J 000664 Pigmented 1
B6(Cg)-Tyr{c-2J}/J 000058 Albino 3
B6.129P-Cx3cr1{tm1Litt}/J 005582 Pigmented 3
Tg(Thy1-EGFP)MJrs/J 007788 Pigmented 1
B6.Cg-Tg(Thy1-YFP)16Jrs/J 003709 Pigmented 1
B6(A)-Rpe65{rd12}/J 005379 Pigmented 2
The images in this report were generated with post-processing steps that were
performed using Matlab (MathWorks Inc, MA, USA) and ImageJ (National Institutes of
94
Health, MD, USA), which included steps to register, average, and contrast stretch the
images [16]. For OCT images, we recorded 5 volumes per acquisition. For TPEF images
of fluorescently labelled cells and FA, we recorded 100 frames per acquisition. For TPEF
images of the RPE layer, we recorded 300 to 900 frames per acquisition. The OCT images
presented in this chapter are either cross-sectional B-scans in the fast scanning direction,
or the en face view at a user-selected depth. The OCT B-scans were aligned to remove
axial motion with a vertical translation to maximize the cross-correlation between the
images. The B-scans presented in this report were an average of 5 adjacent B-scans
within one of the acquired volumes. The en face OCT and OCTA images were generated
with a Maximum Intensity Projection (MIP) between two manually selected axial positions
on the B-scans, which corresponded to the focal plane within the sample. From five en
face OCT/OCTA volumes that were acquired sequentially, one image was selected as the
template image and the other images were registered to the template image prior to
averaging. In the registration process, vertical and horizontal image translations were first
performed to maximize the cross-correlation between the template and moving image.
Then, each image was non-rigidly aligned to the template with a sum of squared
differences similarity metric along cubic B-splines using the Medical Image Registration
Toolbox (MIRT) [69]. For the TPEF images, the same rigid and non-rigid steps were used
to register each frame to the template. However, the unregistered average image was
used as a template [97]. After registration, the OCT/OCTA and TPEF images were
averaged, the image pixel intensity was scaled for presentation if required, and the images
were resized to have the same vertical and horizontal scale. The images that were
acquired at multiple depths from the same eye and the images before/after SAO were
processed identically for comparisons. We estimated the peak signal-to-noise ratio (SNR)
defined by Eq. 7.1 [126,127]:
𝑆𝑁𝑅 = 10 log
max (𝐼)2
𝜎𝑏2 , (7.1)
Where I is the image pixel intensities and σb2 is the variance, which was calculated from
a region within the image that only contained background noise.
95
7.3. Results
The results are split into three parts for each type of fluorophore that was imaged
in the mouse eye. Please see Section 7.3.1 for fluorescein angiography (FA), where we
used OCT to guide the aberrations correction to reduce the laser power to < 3 mW. Section
7.3.2 we demonstrate the imaging performance with fluorescently labelled cells and
Section 7.3.3 we demonstrate imaging of the RPE layer with SAO in a variety of mouse
strains.
7.3.1. Fluorescein angiography
The murine vasculature of the inner retina is stacked in three distinct layers,
including the Outer Plexiform Layer (OPL), Inner Plexiform Layer (IPL), and Nerve Fiber
Layer (NFL). This hierarchical structure provides an opportunity to demonstrate the depth-
resolved aberration correction and the ability of TPEF to provide axial sectioning without
a confocal aperture. In Figure 7.2, the laser power was adjusted to 2.5 mW, and the focal
plane was positioned at the OPL of a 6-month-old albino B6 mouse (B6(Cg)-Tyrc-2J/J).
Then, the SAO wavefront optimization was performed using the en face OCT images of
the OPL, which were extracted from the B-scans between the two depth positions that are
marked by yellow arrows in Figure 7.2. The SAO optimization was performed with 2
iterations for a total algorithmic execution time of ~40 seconds. The top and bottom rows
of Figure 7.2 are before and after the aberration correction, respectively. The first column
shows the improvement in the OCT B-scans (linear scale) at the OPL layer. The OCT B-
scans were selected from the cross-section between the blue arrows on the en face OCT.
The middle column shows the improvement in the sharpness and brightness in the en face
OCT. The third column shows improvement of the TPEF images that resulted from the
wavefront optimization on the en face OCT images. The TPEF images were estimated to
have an SNR of 26.3 dB before aberration correction and 31.0 dB after OCT-guided
aberration correction using Equation 7.1.
96
Figure 7.2. Optical Coherence Tomography (OCT) and Two-Photon Excited Fluorescence (TPEF) images of the mouse retina before (top row) and after (bottom row) OCT-guided Sensorless Adaptive Optics (SAO). The improvement in the OCT B-scan is shown in the left column, the improvement in the en face OCT is shown in the middle column, and the improvement in the TPEF is shown in the right column. The yellow arrows represent the imaging focal position and the line between the blue arrows represents the cross-sectional location of the OCT B-scans. Scale bars: 50 µm.
After SAO at the OPL, using the same aberration correction, OCT and OCTA
volumes were acquired while shifting the focus to the OPL, IPL, and NFL using the
tunable lens, as shown in Figure 7.3(a). The OCT B-scans in top row were from the
same location as in Figure 7.2, where the shift in focal plane can be seen by the change
in image intensities along the depth of the B-scan. The B-scans focused on the OPL,
IPL, and NFL were color-coded in depth and composited with a Maximum Intensity
Projection (MIP), as shown in Figure 7.3(a) (top right). Similarly, TPEF images were
acquired while shifting the focus with the tunable lens through the inner retinal layers
from the OPL to the NFL to create a depth-stack (z-stack). TPEF images were acquired
at 25 depth location with increments of ~5 µm through the inner retina. The bottom row
of Figure 7.3(a) shows the TPEF images focused at the OPL, IPL, and NFL. The TPEF
images were composited with color corresponding to the relative location between the
OPL and NFL (bottom right), which matched the color-coded OCT B-scan (top right).
The red arrows in Figure 7.3(a) point to a few connecting or diving vessels between the
97
retinal layers [128,129], which appear clearly in the TPEF FA. However, these vessels
are not clearly identifiable as connecting vessels in the OCTA.
In order to demonstrate the effect of the SAO-OCT optimization on the TPEF
imaging in a cross-sectional view, TPEF z-stacks were resliced to generate a cross-
section in the same direction as the OCT B-scans (fast scan direction). Prior to the
extraction of TPEF cross-sectional images from the TPEF volume, the 25 TPEF images
were interpolated to 75 pixels in depth to scale with the OCT B-scans. Figure 7.3(b) shows
the TPEF cross-sections before and after SAO optimization. The image intensity was
plotted between the blue arrows and the red arrows in Figure 7.3(b) to show the
improvement due to the aberration correction. The TPEF cross-sectional slices presented
in Figure 7.3(b) had two distinctive blood vessels on top of each other, which
demonstrations the axial sectioning ability of the TPEF.
98
Figure 7.3. (a) OCT B-scans (top row), OCTA en face (middle row), and TPEF (bottom row) with the focal plane at the Outer Plexiform Layer (OPL), Inner Plexiform Layer (IPL), and Nerve Fiber Layer (NFL). In the right column, the images of the vascular layers were composited with a MIP. The red arrows point out connecting vessels in the TPEF. (b) Cross-sectional TPEF images (left) of the inner retinal vasculature before and after Adaptive Optics (SAO) acquired with a 25-step z-stack that was interpolated to 75 image pixels. The axial intensity profile plot between the red and blue arrows of the TPEF cross-sectional images. Scale bars: 50 µm.
7.3.2. GFP and YFP labelled cells
GFP labelled cells have small features that can be used to demonstrate the
imaging resolution of the TPEF system. In this section, we demonstrate the ability to
visualize cellular features after SAO optimization.
99
We imaged a mouse strain with GFP labelled microglia (B6.129P2(Cg)-
Cx3cr1tm1Litt/J) with a center wavelength of 910 nm. Figure 7.4(a) shows a single TPEF
image frame (left), and then an average of 100 frames (right). First, OCT B-scans were
used to navigate to the focal plane where we would expect to find the GFP labelled cells.
Then, as shown in Figure 7.4(b), we decreased the imaging FOV to ~70 µm to performed
the SAO optimization using the fluorescence images. The SAO optimization was
performed with 5 iterations for a total algorithmic execution time of ~100 seconds. The
example in Figure 7.4(b) demonstrates an overall image improvement from an SNR of
33.6 dB before aberration correction to 36.6 dB after aberration correction using Equation
7.1. There was also an improvement in the features of the microglia branches that are
visualized after aberration correction. The images in Figure 7.4(a) and 7.4(b) were from
two mice that were both 18 months of age. In Figure 7.4(c), we show another image after
aberration correction with a FOV that was ~100 µm from another mouse that was also 18
months of age.
Figure 7.4. TPEF imaging of GFP labelled microglia (B6.129P2(Cg)-Cx3cr1{tm1Litt}/J) in the mouse retina. (a) Single TPEF frame (left) and an average of 100 frames (right) at a ~0.8 mm FOV. The red square represents a 100 µm FOV to represent the scale of the microglia. Scale bar: 100 µm. (b) TPEF images of a GFP labelled microglia cells before (left) and after (right) Sensorless Adaptive Optics (SAO). (c) TPEF image after SAO. Scale bars: 20 µm.
We also demonstrate the ability to image GFP labelled retinal ganglion cells (RGC)
of a transgenic mouse strain (Tg(Thy1-EGFP)MJrs/J) with SAO-TPEF and we compared
100
the images to single photon excited fluorescence (SPEF) with SAO. Figure 7.5 (left) shows
an image of a RGC from a 14-month-old mouse after SAO-TPEF with a FOV of ~200 µm.
Figure 7.5 (middle) shows the same cell imaged with SAO-SPEF with a matching FOV.
The larger FOV of ~1.3 mm is presented in Figure 7.5 (right) to show the location of the
RGC relative to the optic nerve head. The imaging system used for SPEF is described in
a recent report [16], which used a 488 nm laser for fluorescence excitation and the same
numerical aperture (NA) into the mouse eye. By comparing these results, it appears that
the TPEF imaging can resolve similar features in the lateral plane. However, the apparent
size of the soma in the SPEF image is 30% larger than the soma in the TPEF image. The
‘glow’ of the soma is reduced in the TPEF image due to improved TPEF axial sectioning
in this example.
Figure 7.5. Comparison of a GFP labelled retinal ganglion cell that was imaged using SAO TPEF (left) and using SAO SPEF with the same 200 µm FOV (middle). A SPEF image is also shown at a ~1.3 mm FOV (right), where the red square represents the 200 µm FOV that was used for the other images. Left scale bar: 20 µm. Right scale bar: 100 µm.
The Thy-1 YFP-16 Line (B6.Cg-Tg(Thy1-YFP)16Jrs/J) mouse retina was also
imaged using SAO-TPEF with only 2.5 mW at 940 nm . From a 7-month-old mouse, a z-
stack of TPEF images were acquired throughout the inner retina to visualize the
fluorescently labelled cells. In Figure 7.6, the OCT B-scans on the top row are presented
in linear scale to show the focal plane at the same depth location as the TPEF images in
the middle row, which are focused at the NFL, IPL, and OPL.
101
Figure 7.6. OCT B-scans (top row) and TPEF (middle row) imaging with the focal plane at the Nerve Fiber Layer (NFL), Inner Plexiform Layer (IPL), and Outer Plexiform Layer (OPL) of a Thy-1 YFP-16 Line (B6.Cg-Tg(Thy1-YFP)16Jrs/J) transgenic mouse. The blue arrow and yellow arrow point at fluorescently labelled cell bodies. The red arrow points at fluorescently labelled axons. In the bottom row, the OCTA en face image (magenta) was composited with the TPEF image (green). Vertical scale bar: 50 µm. Horizontal scale bars: 20 µm.
As shown by the OCT B-scans, some of the fluorescently labelled cell bodies (blue
arrow) and axons (red arrow) appear to located near the NFL, whereas others cell bodies
(yellow arrow) appear to be located near the OPL. The fluorescently labelled cells in the
OPL could be a different type of neural cell, such as horizontal cells [130]. In the bottom
row of Figure 7.6, the OCTA en face images (magenta) were combined with the TPEF
(green) with the focal plane at the NFL (left) and the OPL (right) to co-localize of the
fluorescently labelled cells with the blood vessels
7.3.3. RPE imaging
In the following results (Figure 7.7 and 7.8), we demonstrate TPEF imaging of the
RPE with the assistance of SAO-OCT in an albino B6 mouse (B6(Cg)-Tyrc-2J/J). For Figure
7.7, we imaged a mouse that was 5.5 months of age. First, the OCT was used to ensure
102
the focal plane was at the RPE layer of the retina. In Figure 7.7(a), the OCT B-scans (top
row) are presented in linear scale to show the intensity change as the focal plane was
shifted from the NFL (left), to the OPL (middle), and to the RPE (right). Here, we used the
TPEF images to perform one iteration of the SAO optimization for a total algorithmic
execution time of ~20 seconds. In Figure 7.7(b) and 7.7(c), there was an apparent
increase in signal after SAO and an improved image quality of the RPE mosaic, as shown
by the line plots between the blue arrows and the red arrows. From the line plot, we can
calculate a 25 - 30 µm spacing between peaks, which corresponds the spacing of the RPE
cells.
Figure 7.7. (a) The SAO-OCT B-scans in linear scale (top row) and the en face OCT (bottom row) with the focal plane at the Nerve Fiber Layer (NFL), Outer Plexiform Layer (OPL), and Retinal Pigment Epithelium (RPE) in the mouse retina. The en face OCT images were extracted between the cyan arrows (NFL), yellow arrows (OPL), and green arrows (RPE). The OCT B-scans were located between the red arrows on the en face OCT image. (b) TPEF images of the RPE of the mouse retina before and after SAO. (c) An intensity line plot between the blue arrows and the red arrows on the TPEF images of the RPE mosaic. Scale bars 50 µm.
103
We imaged the RPE of one mouse (5.5 months of age) four days apart to
demonstrate the ability to perform non-invasive longitudinal imaging of the same eye over
time. This interval was sufficient time for photo-chemical and photo-thermal damage to
appear in the OCT [131,132]. The TPEF from the RPE and the OCT images are shown
on day 1 in Figure 7.8(a). The OCT B-scans are presented in log-scale and show that
there was no damage immediately after using an 8 mW exposure for a few minutes. In
Figure 7.8(b), we imaged the same area four days later, where there was no apparent
damage from the two imaging sessions. In Figure 7.8(c), we have digitally zoomed into a
small area of the TPEF images to show a similar RPE pattern on each day. TPEF images
from day 1 (green) and day 4 (magenta) were combined with a MIP.
Figure 7.8. (a) TPEF images of the RPE (left), en face OCT (middle), and OCT B-scans (right). (b) TPEF images of the RPE (left), en face OCT (middle), and OCT B-scans (right) from the same mouse four days later. (c) The digital enlargement of the TPEF images on day 1 (green) and day 4 (magenta), which were combined with a MIP. Scale bars 50 µm.
An advantage of working with mice is the well-established ability to manipulate
their genetics. However, here, we show differences in the TPEF images from the RPE
layer that resulted from a variety of mouse strains. In Figure 7.9, using 740 nm excitation
light, we imaged the popular pigmented B6 mouse (C57BL/6J) at 16 months of age, an
albino B6 mouse (B6(Cg)-Tyrc-2J/J) at 5.5 months of age, and a pigmented mouse strain
with a disruption of the rpe65 gene (B6(A)-Rpe65rd12/J)) at 2.5 months of age. The TPEF
images from the B6 mouse had an evenly distributed signal from the RPE with only faint
104
structures visible, as compared the albino B6 mouse, where the RPE mosaic was clear.
In the pigmented rpe65 mouse, we can also visualize the RPE mosaic, which is likely due
to the build-up of retinyl esters storage particles from the dysfunction in the visual system
[123,124]. Fluorescent compounds in the RPE are known to accumulate with age [133],
which would increase the autofluorescence in older mice and could be an additional factor
that affects the visualization of the RPE mosaic.
Figure 7.9. TPEF from the RPE layer of the mouse retina in three different mouse strains, including a pigmented B6 mouse (C57BL/6J), an albino B6 mouse (B6(Cg)-Tyr{c-2J}/J), and a pigmented rpe65 mouse (B6(A)-Rpe65{rd12}/J). Scale bar 100 µm.
We further investigated the ability of TPEF to provide compound-specific contrast.
Figure 7.10 shows TPEF images at different central wavelengths from the pigmented
rpe65 mouse at 18 months of age. At 760 nm, we are able to visualize the fluorophores
that have accumulated near the cell membrane of the RPE cells. However, after 800 nm
the TPEF signal is dominated by a more evenly distributed fluorescence signal in the RPE
layer. These results agree with other Literature [81,133,134], which demonstrate that the
fluorescence from the retinyl ester storage particles is more efficient at 760 nm than at
longer wavelengths. At the longer wavelengths used in these results, the fluorescence
emission is largely due to the compound A2E in lipofuscin. Since autofluorescent
compounds, including A2E, accumulate in the RPE with age [133], there was likely more
fluorescence emission with the older mouse than the rpe65 mouse imaged for Figure 7.9.
105
Figure 7.10. TPEF image of a pigmented rpe65 mouse (B6(A)-Rpe65{rd12}/J) with different central wavelengths, including 760 nm, 780 nm, 800 nm, and 820 nm. The red arrow highlights an RPE cell where the fluorescence near the cell membrane is reduced with longer wavelengths. Scale bar 50 µm.
7.4. Discussion
In this chapter, we have demonstrated OCT guided TPEF imaging with image-
based wavefront optimization to improve the fluorescence signal. Since the TPEF is very
sensitive to the focal spot size, aberration correction is crucial for high-resolution imaging
with minimal laser power. We used a tunable femtosecond laser source to enable retinal
imaging from a variety of fluorophores. The system also used a tunable lens to shift the
focal plane within the retina, and a deformable mirror for aberration correction with a hill-
climbing coordinate-search algorithm. Our results included high-resolution volumetric
fluorescein angiography of mouse retinal vasculature, sub-cellular imaging resolution of
fluorescently labelled cells, and intrinsic fluorescence imaging from the RPE. We have
demonstrated that some samples can be imaged with < 3 mW of laser power, including
YFP labelled cells and fluorescein angiography. We have also demonstrated RPE imaging
in various mouse strains and we have shown that the amount of pigment in RPE of the
particular strain can reduce the ability to visualize the RPE mosaic with TPEF.
The optical layout in this chapter was similar to the imaging systems for the mouse
retina [15,61] in Chapter 6, which we were able to compact into a 1.5x2 ft footprint, as
described in reference [16]. Using a similar design, the compact system would enable
easier collaboration with other vision scientists.
In the mouse model, common methods such as genetic manipulation or viral
injection can be used to label cells with GFP. In the Literature, TPEF imaging of GFP
106
labelled cells without AO has been demonstrated [135,136], as well as with WFS-based
AO methods [65]. However, it can be difficult to accurately measure the aberrations from
a mouse eye without a high degree of system complexity, whereas SAO has the ability to
perform depth-specific aberration correction at the cost of algorithmic execution time and
the potential for errors due to sample motion [18]. This is important since high-resolution
GFP imaging with NIR light could be used to investigate the visual response of RGCs with
calcium indicators, such as GCaMP [64,65,135]. While the single photon excitation at 488
nm would provide a higher SNR, the NIR light used for TPEF is much less likely to activate
the visual system. In this chapter, we used 9 mW of laser power at 910 nm to image the
GFP labelled RGCs. It may be possible to measure a calcium response in the RGC by
only imaging soma [136]. This could be performed with a significant reduction in the laser
power, which may be similar or less than the 2.5 mW of laser power at 940 nm that was
required for imaging the YFP labelled cells.
For TPEF, SAO is well-suited for good aberration correction performance due to
the nonlinear relationship between the signal and the spot size [38]. Hence, the TPEF
signal strength is significantly degraded in the presence of aberrations. The overall
optimization convergence speed and performance could be improved by using a
combination of aberration correction with the OCT signal, increasing the TPEF signal
enough that it could be used for a second iteration of the SAO process for a more precise
correction. Furthermore, even in the cases where there are poor structural features in the
OCT at the location of the fluorophore, the SAO could potentially be improved with low-
order aberrations using the OCT images before further optimization using the TPEF.
Similarly, at a large imaging FOV, a faster initial optimization could have been used for
partial aberration correction before reducing the FOV to an area interest where the
aberration correction would have better performance [18].
The images of the RPE in this chapter had strong differences for each strain and
appear to be affected by the type of accumulated compounds, pigmentation, and age of
the mouse. In order to visualize the cellular mosaic of the RPE, the TPEF signal from the
features at the edges of the cells had to be greater than the background signal. This was
the case for the for the albino B6 mice and the pigmented rpe65 mouse, where fluorescent
compounds were visualized near the RPE cell membrane [123,124]. Palczewska et al.
[12] has recently reported results for imaging RPE of albino rpe65 mice with laser powers
down to 1 mW by averaging many frames, and using ultra short light pulses down to ~20
107
femtoseconds into the eye. Our results use similar methods to average many frames, but
we could potentially visualize the RPE mosaic with similar power levels by imaging the
albino rpe65 mouse and by pre-compensating for the group velocity dispersion through
the optical components. However, we would still be limited by our source at 70 to 100 fs
without sacrificing tunability with a broader spectral bandwidth laser. As shown in this
chapter, the strain of mouse is important for visualizing the RPE mosaic at low power
levels. However, the lack of pigment in albino mouse strains could reduce the performance
of WFS-based aberration correction.
Reducing the average laser power on the retina is important not only for robust
longitudinal animal studies, but also for potential applications in human imaging [121,137].
Although our laser power levels for these results are not safe for human eye imaging,
mouse imaging can be used to further develop imaging techniques to enable TPEF at safe
power levels. Recently, it has been demonstrated that averaging hundreds of OCT
volumes can produce high-resolution retinal imaging of transparent features, including
RGCs [138,139]. This so called ‘super-averaging’ could enable all 100 – 900 TPEF frames
to be co-registered with the OCT, which could further reduce the TPEF signal
requirements.
7.5. Summary
In conclusion, we have demonstrated that SAO and OCT can be used to achieve
non-invasive cellular-resolution TPEF imaging from a variety of fluorophores in the retina.
The increase in TPEF signal after aberration correction enables imaging to be performed
with less laser power than without AO. Finally, if safe levels exposure levels are
determined, the techniques shown here could be developed for human imaging.
108
Chapter 8. Future work and conclusion
8.1. Technology refinement
Improvements and refinement of the current imaging technology will further
increase the reliability and usability of the SAO imaging. In order to enable widespread
use SAO imaging, the future objectives are to increase the image optimization speed and
reduce motion artifacts. I envision that SAO aberration correction will be quick enough that
it is not be cumbersome to the user or even continuous. In this section, I will outline some
proposed solutions for future investigations.
Although SAO imaging has been demonstrated to be able to provide similar
performance as WFS-based AO, the major disadvantage is the time required for SAO. A
method for potentially improving the sensorless AO is to search an ideal set of basis
functions in the optimization algorithm. An idea basis can be generated from system
measurements in order to orthogonalize the basis used by corrective element to the effect
the modes have on the image quality metric [98]. Generating these modes has been
demonstrated to improved the performance of the sensorless optimization using a modal-
search, especially in single iteration [140,141]. Iterating multiple times through the Zernike
modes is the simple solution that we have implemented to the SAO algorithm, which also
benefits from the improved image SNR as the aberration correction proceeds. An
alternative basis was used for WFS-based AO and described in Section 5.2.4. The basis
was generated by the SVD of the influence matrix measured by the wavefront sensor and
could potentially also be use as modes for the sensorless AO algorithm.
Another solution is to use faster converging algorithms that require fewer imaging
frames to determine the optimal correction. The fastest algorithms develop a mathematical
model using the effect on the images due to known aberrations put on the deformable
element. These are known as model-based algorithms, which can converge is as few as
N+1 measurements, with N being the number of modes that are used in the model.
Examples of model-based algorithms include the DONE algorithm [21,101], and sphere
packing [100].
109
Since motion artifact can be detrimental to the SAO performance and to the final
averaged image, real-time tracking systems could provide solutions [71,85,119]. We have
started with simple image-based registration, as mentioned in Section 6.4. For SAO, this
has allowed a region of interest (ROI) to be followed and used during the course of an
optimization. The next logical step is to have the imaging system respond to the registered
motion, potentially in three dimensions using the OCT for volumetric localization. For the
motion of an anesthetized mouse, it may be sufficient to follow the sample with the
galvanometer mirrors and tunable lens. Furthermore, this would reduce the reliance on
registration in post processing.
Imaging the human retina with two-photon excited fluorescence will not be possible
without reducing the laser exposures currently required for imaging. As discussed in
Section 7.5, further reducing the imaging power required for TPEF imaging of the mouse
RPE seems possible through the development image processing tools that allow the OCT
and TPEF to be concurrently registered and averaged. Preliminary results, shown in
Figure 8.1, demonstrate that registration and averaging of >150 OCT volumes acquired
from the mouse eye using the multi-modal system described in Chapter 5 can produce
high quality images the nerve fiber bundles.
Figure 8.1. Volumetric averaging of 150 OCT volumes. Scale bar: 50 µm.
8.2. Non-confocal Scanning Laser Ophthalmoscopy
For structural imaging purposes, non-invasive in vivo visualization of the cellular
mosaics in the outer retina in the mouse has been scarce in the Literature, most likely due
to difficultly correcting aberrations. Notable examples include, rod photoreceptor mosaic
110
imaging reported by Y. Geng et al. [4], and P. Zhang et al. [19] using a confocal AO SLO
system and the RPE mosaic was imaged by Palczewska et al. and Stremplewski et al.
[12,81,123] using two-photon excited fluorescence. Another approach uses a split-
aperture detector that can image the photoreceptor mosaic and horizontal cell bodies of
the outer plexiform layer, demonstrated by A. Guevara-Torres et al. [130]. Other non-
confocal regimes are also possible in order to favor multiply scattered light and reject the
dominating back-scattered light. These methods have been demonstrated to enable the
visualization of transparent cells without fluorescent contrast agents.
A split-detector system or other non-confocal configuration replaces the confocal
aperture in an SLO with an annulus, a filament, or a knife edge. For the split-detector
configuration, the PSF is divided into two detectors with a knife edge in the detection path.
The intensity received in each detector is subtracted to reveal the asymmetric intensity
variations that arise in the PSF from the multiply scattered light.
Combining SAO with non-confocal SLO is an interesting prospect. However, the
image quality metric may require the center portion of the aperture in order to optimize the
image. The simplest solution is to allow the center of the aperture to transmit to one
detector and an outer annulus could be reflected to another device. Then, the confocal
channel could be used for sensorless AO methods. Other approaches could use a
configurable aperture, such as a spatial light modulator (SLM) or digital-micro-device
(DMD) in the detection path. This has been demonstrated by S. A. Burns et al. [142] with
an AO SLO system for enhanced retinal vasculature imaging. There are also the
possibilities to take advantage of multiple detectors for improving the speed and
performance of sensorless AO, which has been demonstrated by Pozzi et al. [92].
8.3. Extensions of two-photon excited florescence technology
Future technology developments of the two-photon imaging capabilities are
important because of the potential for the modality to provide functional insight to the
retina. Imaging with near infrared light is preferred since the retina is less sensitive to the
longer wavelengths.
111
Genetically encoded calcium indicators, such as GCaMP, can be expressed in
retinal ganglion cells of mice. The fluorescence intensity from a labelled ganglion cell will
increase when the cell is activated by an electrochemical signal originating from stimulated
photoreceptors. Some studies have demonstrated that the activity of the cell expressing
a calcium indicator can be optically measured [64,65,136,143–145]. Imaging the calcium
indicators with infrared light would be advantageous, since the photoreceptors are less
likely to be activated by the imaging light. This could enable novel studies of the signal
pathways via the interconnected neural cells in the retina by the response to light in
different areas of retina [64,65].
For imaging autofluorescence of the retina, often the source of the fluorescence
intensity is a result of many fluorophores with overlapping emission spectra. A technique
known as fluorescence lifetime imaging microscopy (FLIM) has been combined with
ophthalmoscopy to produces contrast from the fluorescence decay rate, which is
characteristic of the fluorophore [146–148]. This enables fluorescence imaging to be
mostly independent of the concentration of the fluorophores, as well as the ability to
distinguish emission from different fluorophores are mixed together. Furthermore, FLIM
has been demonstrated to provide information about the health of the cells. Recently, J.
A. Feeks et al. [149] demonstrated two-photon excited FLIM with AO in the mouse retina.
Using two-photon excited FLIM with cellular resolution provides an opportunity to
distinguish intrinsic fluorophores that are important for visual function.
8.4. Conclusion
Imaging the mouse retina with optimized resolution presents challenges that
require the development of specialized equipment. This is a worthwhile endeavour since
the mouse model is used ubiquitously in biology and health research. This thesis has
presented novel imaging systems (with an emphasis on fluorescence detection) that are
capable of high-resolution imaging for mouse retina imaging using the flexibility provided
by Sensorless Adaptive Optics (SAO) for aberration correction.
The SAO optimization algorithm developed throughout this thesis used a multi-
iteration hill-climbing coordinate search algorithm, with a decreasing search range to
enable diffraction-limited imaging. Furthermore, pupil segmentation methods for image-
based adaptive optics were demonstrated to be feasible for imaging the retina.
112
The work presented in this thesis has resulted in a compact multi-modal imaging
design, which included Optical Coherence Tomography (OCT), OCT-Angiography,
Scanning Laser Ophthalmoscopy (SLO), and fluorescence detection. The imaging
capability of these imaging modalities were demonstrated by high resolution fluorescence
and structural imaging of the mouse retina. Also based on the results in Chapter 7, it
seems straightforward to use the same compact design for Two-Photon Excited
Fluorescence and OCT imaging with a few modifications to the detection scheme. The
flexibility of the SAO has been demonstrated by a range of imaging scenarios that may be
desired by a vision scientist, including imaging the structures in the different layers of the
retina with OCT and SLO, fluorescein angiography, GFP labelled cells, and
autofluorescence of the RPE mosaic.
113
References
1. "Blindness In Canada," https://cnib.ca/en/sight-loss-info/blindness/blindness-canada?region=bc. Accessed Feb. 2019.
2. D. R. Williams, "Imaging single cells in the living retina," Vision Res. 51(13), 1379–1396 (2011).
3. S. Marcos, J. S. Werner, S. A. Burns, W. H. Merigan, P. Artal, D. A. Atchison, K. M. Hampson, R. Legras, L. Lundstrom, G. Yoon, J. Carroll, S. S. Choi, N. Doble, A. M. Dubis, A. Dubra, A. Elsner, R. Jonnal, D. T. Miller, M. Paques, H. E. Smithson, L. K. Young, Y. Zhang, M. Campbell, J. Hunter, A. Metha, G. Palczewska, J. Schallek, and L. C. Sincich, "Vision science and adaptive optics, the state of the field," Vision Res. 132, 3–33 (2017).
4. Y. Geng, A. Dubra, L. Yin, W. H. Merigan, R. Sharma, R. T. Libby, and D. R. Williams, "Adaptive optics retinal imaging in the living mouse eye," Biomed. Opt. Express 3(4), 715–734 (2012).
5. Y. Geng, L. A. Schery, R. Sharma, A. Dubra, K. Ahmad, R. T. Libby, and D. R. Williams, "Optical properties of the mouse eye," Biomed. Opt. Express 2(4), 717–738 (2011).
6. M. Land, "Photoreception," Encycl. Br. (2018). https://www.britannica.com/science/photoreception. Accessed Feb. 2019.
7. L. Li, N. Eter, and P. Heiduschka, "The microglia in healthy and diseased retina," Exp. Eye Res. 136, 116–130 (2015).
8. "Webvision – The Organization of the Retina and Visual System," https://webvision.med.utah.edu/. Accessed Feb. 2019.
9. P. D. Kiser, M. Golczak, and K. Palczewski, "Chemistry of the retinoid (visual) cycle.," Chem. Rev. 114(1), 194–232 (2014).
10. R. Sharma, D. R. Williams, G. Palczewska, K. Palczewski, and J. J. Hunter, "Two-Photon Autofluorescence Imaging Reveals Cellular Structures Throughout the Retina of the Living Primate Eye.," Invest. Ophthalmol. Vis. Sci. 57(2), 632–646 (2016).
11. J. J. Hunter, B. Masella, A. Dubra, R. Sharma, L. Yin, W. H. Merigan, G. Palczewska, K. Palczewski, and D. R. Williams, "Images of photoreceptors in living primate eyes using adaptive optics two-photon ophthalmoscopy," Biomed. Opt. Express 2(1), 139–148 (2011).
12. G. Palczewska, P. Stremplewski, S. Suh, N. Alexander, D. Salom, Z. Dong, D. Ruminski, E. H. Choi, A. E. Sears, T. S. Kern, M. Wojtkowski, and K. Palczewski, "Two-photon imaging of the mammalian retina with ultrafast pulsing laser.," JCI insight 3(17), 121555 (2018).
13. J. Schallek, Y. Geng, H. Nguyen, and D. R. Williams, "Morphology and topography of retinal pericytes in the living mouse retina using in vivo adaptive optics imaging and ex vivo characterization," Invest. Ophthalmol. Vis. Sci. 54(13), 8237–8250 (2013).
14. R. J. Zawadzki, P. Zhang, A. Zam, E. B. Miller, M. Goswami, X. Wang, R. S.
114
Jonnal, S. H. Lee, D. Y. Kim, J. G. Flannery, J. S. Werner, M. E. Burns, and E. N. Pugh Jr., "Adaptive-optics SLO imaging combined with widefield OCT and SLO enables precise 3D localization of fluorescent cells in the mouse retina," Biomed. Opt. Express 6(6), 2191–2210 (2015).
15. D. J. Wahl, Y. Jian, S. Bonora, R. J. Zawadzki, and M. V. Sarunic, "Wavefront sensorless adaptive optics fluorescence biomicroscope for in vivo retinal imaging in mice," Biomed. Opt. Express 7(1), 1–12 (2016).
16. D. J. Wahl, R. Ng, M. J. Ju, Y. Jian, and M. V. Sarunic, "Sensorless adaptive optics multimodal en-face small animal retinal imaging," Biomed. Opt. Express 10(1), 252–267 (2019).
17. L. L. Molday, D. Wahl, M. V Sarunic, and R. S. Molday, "Localization and functional characterization of the p.Asn965Ser (N965S) ABCA4 variant in mice reveal pathogenic mechanisms underlying Stargardt macular degeneration," Hum. Mol. Genet. 27(2), 295–306 (2018).
18. D. J. Wahl, P. Zhang, J. Mocci, M. Quintavalla, R. Muradore, Y. Jian, S. Bonora, M. V. Sarunic, and R. J. Zawadzki, "Adaptive optics in the mouse eye: wavefront sensing based vs. image-guided aberration correction," Biomed. Opt. Express (In Production) (2019).
19. P. Zhang, J. Mocci, D. J. Wahl, R. K. Meleppat, S. K. Manna, M. Quintavalla, R. Muradore, M. V. Sarunic, S. Bonora, E. N. Pugh, and R. J. Zawadzki, "Effect of a contact lens on mouse retinal in vivo imaging: Effective focal length changes and monochromatic aberrations," 172, 86–93 (2018).
20. D. J. Wahl, C. Huang, S. Bonora, Y. Jian, and M. V. Sarunic, "Pupil segmentation adaptive optics for invivo mouse retinal fluorescence imaging," Opt. Lett. 42(7), 1365–1368 (2017).
21. H. R. G. W. Verstraete, M. Heisler, M. J. Ju, D. Wahl, L. Bliek, J. Kalkman, S. Bonora, Y. Jian, M. Verhaegen, and M. V Sarunic, "Wavefront sensorless adaptive optics OCT with the DONE algorithm for in vivo human retinal imaging [Invited]," Biomed. Opt. Express 8(4), 2261–2275 (2017).
22. M. J. Ju, M. Heisler, D. Wahl, Y. Jian, and M. V. Sarunic, "Multiscale sensorless adaptive optics OCT angiography system for in vivo human retinal imaging," J. Biomed. Opt. 22(12), 1 (2017).
23. M. J. Ju, C. Huang, D. J. Wahl, Y. Jian, and M. V. Sarunic, "Visible light sensorless adaptive optics for retinal structure and fluorescence imaging," Opt. Lett. 43(20), 5162–5165 (2018).
24. M. Cua, D. J. Wahl, Y. Zhao, S. Lee, S. Bonora, R. J. Zawadzki, Y. Jian, and M. V. Sarunic, "Coherence-Gated Sensorless Adaptive Optics Multiphoton Retinal Imaging," Sci. Rep. 6(1), 32223 (2016).
25. D. J. Wahl, M. J. Ju, Y. Jian, and M. V. Sarunic, "Non-invasive cellular-resolution retinal imaging with two-photon excited fluorescence," Biomed. Opt. Express (In Production) (2019).
26. R. H. Webb, "Confocal optical microscopy," Reports Prog. Phys. 59(3), 427–471 (1996).
27. J. B. Pawley, ed., Handbook Of Biological Confocal Microscopy (Springer US, 2006).
115
28. T. Wilson1 and A. R. Carlini, "Size of the detector in confocal imaging systems," Opt. Lett. 12(4), 227–229 (1987).
29. J. F. de Boer, C. K. Hitzenberger, and Y. Yasuno, "Polarization sensitive optical coherence tomography – a review [Invited]," Biomed. Opt. Express 8(3), 1838–1873 (2017).
30. Y. Jia, O. Tan, J. Tokayer, B. Potsaid, Y. Wang, J. J. Liu, M. F. Kraus, H. Subhash, J. G. Fujimoto, J. Hornegger, and D. Huang, "Split-spectrum amplitude-decorrelation angiography with optical coherence tomography," Opt. Express 20(4), 4710–4725 (2012).
31. I. Gorczynska, J. V. Migacz, R. J. Zawadzki, A. G. Capps, and J. S. Werner, "Comparison of amplitude-decorrelation, speckle-variance and phase-variance OCT angiography methods for imaging the human retina and choroid," Biomed. Opt. Express 7(3), 911–942 (2016).
32. A. Zhang, Q. Zhang, C.-L. Chen, and R. K. Wang, "Methods and algorithms for optical coherence tomography-based angiography: a review and comparison.," J. Biomed. Opt. 20(10), 100901 (2015).
33. M. S. Mahmud, D. W. Cadotte, B. Vuong, C. Sun, T. W. H. Luk, A. Mariampillai, and V. X. D. Yang, "Review of speckle and phase variance optical coherence tomography to visualize microvascular networks," J. Biomed. Opt. 18(5), 050901 (2013).
34. T. Liu, H. Jung, J. Liu, M. Droettboom, and J. Tam, "Noninvasive near infrared autofluorescence imaging of retinal pigment epithelial cells in the human retina using adaptive optics," Biomed. Opt. Express 8(10), 4348 (2017).
35. S. Schmitz-Valckenberg, D. Lara, S. Nizari, E. M. Normando, L. Guo, A. R. Wegener, A. Tufail, F. W. Fitzke, F. G. Holz, and M. F. Cordeiro, "Localisation and significance of in vivo near-infrared autofluorescent signal in retinal imaging.," Br. J. Ophthalmol. 95(8), 1134–1139 (2011).
36. U. Kellner, S. Kellner, B. H. F. Weber, B. Fiebig, S. Weinitz, and K. Ruether, "Lipofuscin- and melanin-related fundus autofluorescence visualize different retinal pigment epithelial alterations in patients with retinitis pigmentosa," Eye 23(6), 1349–1359 (2009).
37. F. Helmchen and W. Denk, "Deep tissue two-photon microscopy.," Nat. Methods 2(12), 932–940 (2005).
38. W. Denk, J. Strickler, and W. Webb, "Two-photon laser scanning fluorescence microscopy," Science 248(4951), 73–76 (1990).
39. J. Porter, H. Queener, J. Lin, K. Thorn, and A. Awwal, Adaptive Optics for Vision Science: Principles, Practices, Design and Applications (Wiley, 2006).
40. A. Roorda, F. Romero-Borja, W. Donnelly, III, H. Queener, T. Hebert, and M. Campbell, "Adaptive optics scanning laser ophthalmoscopy," Opt. Express 10(9), 405 (2002).
41. R. J. Zawadzki, S. M. Jones, S. S. Olivier, M. Zhao, B. A. Bower, J. A. Izatt, S. Choi, S. Laut, and J. S. Werner, "Adaptive-optics optical coherence tomography for high-resolution and high-speed 3D retinal in vivo imaging," Opt. Express 13(21), 8532–8546 (2005).
116
42. D. T. Miller, J. Qu, R. S. Jonnal, and K. E. Thorn, "Coherence gating and adaptive optics in the eye," in V. V. Tuchin, J. A. Izatt, and J. G. Fujimoto, eds. (International Society for Optics and Photonics, 2003), 4956, pp. 65–72.
43. D. G. Voelz, Computational Fourier Optics: A MATLAB Tutorial (SPIE, 2011).
44. D. E. Milkie, E. Betzig, and N. Ji, "Pupil-segmentation-based adaptive optical microscopy with full-pupil illumination.," Opt. Lett. 36(21), 4206–4208 (2011).
45. N. Ji, D. E. Milkie, and E. Betzig, "Adaptive optics via pupil segmentation for high-resolution imaging in biological tissues.," Nat. Methods 7(2), 141–147 (2010).
46. C. Wang and N. Ji, "Pupil-segmentation-based adaptive optical correction of a high-numerical-aperture gradient refractive index lens for two-photon fluorescence endoscopy," Opt. Lett. 37(11), 2001 (2012).
47. B. C. Chauhan, K. T. Stevens, J. M. Levesque, A. C. Nuschke, G. P. Sharpe, N. O’Leary, M. L. Archibald, and X. Wang, "Longitudinal in vivo imaging of retinal ganglion cells and retinal thickness changes following optic nerve injury in mice," PLoS One 7(6), e40352 (2012).
48. L. Chen, P. Yang, and A. Kijlstra, "Distribution, markers, and functions of retinal microglia," Ocul. Immunol. Inflamm. 10(1), 27–39 (2002).
49. P. Zhang, A. Zam, Y. Jian, X. Wang, M. E. Burns, M. V. Sarunic, E. N. Pugh, and R. J. Zawadzki, "Multispectral scanning laser ophthalmoscopy combined with optical coherence tomography for simultaneous in vivo mouse retinal imaging," Proc. SPIE 9307, 93070H (2015).
50. C. Alt, J. M. Runnels, L. J. Mortensen, W. Zaher, and C. P. Lin, "In vivo imaging of microglia turnover in the mouse retina after ionizing radiation and dexamethasone treatment," Invest. Ophthalmol. Vis. Sci. 55(8), 5314–5319 (2014).
51. A. Roorda, "Adaptive optics for studying visual function: a comprehensive review," J. Vis. 11(5), 6 (2011).
52. R. Tyson, Principles of Adaptive Optics (CRC Press, 2010).
53. J. A. Kubby, Adaptive Optics for Biological Imaging (CRC Press, 2013).
54. Y. N. Sulai and A. Dubra, "Non-common path aberration correction in an adaptive optics scanning ophthalmoscope," Biomed. Opt. Express 5(9), 3059–3073 (2014).
55. J. W. Evans, R. J. Zawadzki, S. M. Jones, S. S. Olivier, and J. S. Werner, "Error budget analysis for an adaptive optics optical coherence tomography system," Opt. Express 17(16), 13768–13784 (2009).
56. F. Felberer, J. S. Kroisamer, C. K. Hitzenberger, and M. Pircher, "Lens based adaptive optics scanning laser ophthalmoscope," Opt. Express 20(16), 17297–17310 (2012).
57. A. Dubra and Y. Sulai, "Reflective afocal broadband adaptive optics scanning ophthalmoscope," Biomed. Opt. Express 2(6), 1757–1768 (2011).
58. M. J. Booth, "Adaptive optics in microscopy," Philos. Trans. A. Math. Phys. Eng. Sci. 365(1861), 2829–2843 (2007).
59. Y. Jian, R. J. Zawadzki, and M. V Sarunic, "Adaptive optics optical coherence tomography for in vivo mouse retinal imaging," J. Biomed. Opt. 18(5), 56007 (2013).
117
60. K. S. K. Wong, Y. Jian, E. M. C. Cua, J. Xu, S. Bonora, R. J. Zawadzki, M. V. Sarunic, Y. Jian, M. Cua, S. Bonora, R. J. Zawadzki, and M. V. Sarunic, "In vivo imaging of human photoreceptor mosaic with wavefront sensorless adaptive optics optical coherence tomography," Biomed. Opt. Express 6(2), 580–590 (2015).
61. Y. Jian, J. Xu, M. A. Gradowski, S. Bonora, R. J. Zawadzki, and M. V. Sarunic, "Wavefront sensorless adaptive optics optical coherence tomography for in vivo retinal imaging in mice," Biomed. Opt. Express 5(2), 547–559 (2014).
62. D. P. Biss, D. Sumorok, S. A. Burns, R. H. Webb, Y. Zhou, T. G. Bifano, D. Côté, I. Veilleux, P. Zamiri, and C. P. Lin, "In vivo fluorescent imaging of the mouse retina using adaptive optics," Opt. Lett. 32(6), 659–661 (2007).
63. C. Alt, D. P. Biss, N. Tajouri, T. C. Jakobs, and C. P. Lin, "An adaptive-optics scanning laser ophthalmoscope for imaging murine retinal microstructure," Proc. SPIE 7550, 755019 (2010).
64. L. Yin, Y. Geng, F. Osakada, R. Sharma, A. H. Cetin, E. M. Callaway, D. R. Williams, and W. H. Merigan, "Imaging light responses of retinal ganglion cells in the living mouse eye," J. Neurophysiol. 109(9), 2415–2421 (2013).
65. R. Sharma, L. Yin, Y. Geng, W. H. Merigan, G. Palczewska, K. Palczewski, D. R. Williams, and J. J. Hunter, "In vivo two-photon imaging of the mouse retina," Biomed. Opt. Express 4(8), 1285–1293 (2013).
66. S. A. Burns, R. Tumbar, A. E. Elsner, D. Ferguson, and D. X. Hammer, "Large-field-of-view, modular, stabilized, adaptive-optics-based scanning laser ophthalmoscope," J. Opt. Soc. Am. A 24(5), 1313–1326 (2007).
67. M. Martínez-Corral, A. Doblas, E. Sánchez-Ortiga, J. Sola-Pikabea, and G. Saavedr, "Static real-time capture of 3D microscopy images," SPIE Newsroom (2015).
68. L. N. Thibos, R. A. Applegate, J. T. Schwiegerling, R. Webb, and VSIA Standards Taskforce Members, "Standards for reporting the optical aberrations of eyes," J. Refract. Surg. 18(5), 652–660 (2002).
69. A. Myronenko and X. Song, "Intensity-based image registration by minimizing residual complexity," IEEE Trans. Med. Imaging 29(11), 1882–1891 (2010).
70. A. Dubra and Z. Harvey, "Registration of 2D Images from Fast Scanning Ophthalmic Instruments," in Biomedical Image Registration, Vol. 6204 of Lecture Notes in Computer Science (Springer, 2010), pp. 60–71.
71. C. K. Sheehy, Q. Yang, D. W. Arathorn, P. Tiruveedhula, J. F. de Boer, and A. Roorda, "High-speed, image-based eye tracking with a scanning laser ophthalmoscope," Biomed. Opt. Express 3(10), 2611–2622 (2012).
72. O. P. Kocaoglu, R. D. Ferguson, R. S. Jonnal, Z. Liu, Q. Wang, D. X. Hammer, and D. T. Miller, "Adaptive optics optical coherence tomography with dynamic retinal tracking," Biomed. Opt. Express 5(7), 2262–2284 (2014).
73. W. Sun, N. Li, and S. He, "Large-scale morphological survey of mouse retinal ganglion cells," J. Comp. Neurol. 451(2), 115–126 (2002).
74. J. Coombs, D. van der List, G. Y. Wang, and L. M. Chalupa, "Morphological properties of mouse retinal ganglion cells," Neuroscience 140(1), 123–136 (2006).
118
75. W. Zou, X. Qi, and S. A. Burns, "Wavefront-aberration sorting and correction for a dual-deformable-mirror adaptive-optics system," Opt. Lett. 33(22), 2602–2604 (2008).
76. C. Li, N. Sredar, K. M. Ivers, H. Queener, and J. Porter, "A correction algorithm to simultaneously control dual deformable mirrors in a woofer-tweeter adaptive optics system," Opt. Express 18(16), 16671–16684 (2010).
77. S. M. Jones, S. Olivier, D. Chen, S. Joeres, S. Sadda, R. J. Zawadzki, J. S. Werner, and D. T. Miller, "Adaptive optics ophthalmologic systems using dual deformable mirrors," Proc. SPIE 6467, 64670H (2007).
78. X. Zhou, P. Bedggood, and A. Metha, "Limitations to adaptive optics image quality in rodent eyes," Biomed. Opt. Express 3(8), 1811–1824 (2012).
79. S. Zommer, E. N. Ribak, S. G. Lipson, and J. Adler, "Simulated annealing in ocular adaptive optics," Opt. Lett. 31(7), 939–941 (2006).
80. H. Hofer, N. Sredar, H. Queener, C. Li, and J. Porter, "Wavefront sensorless adaptive optics ophthalmoscopy in the human eye," Opt. Express 19(15), 14160–14171 (2011).
81. G. Palczewska, Z. Dong, M. Golczak, J. J. Hunter, D. R. Williams, N. S. Alexander, and K. Palczewski, "Noninvasive two-photon microscopy imaging of mouse retina and retinal pigment epithelium through the pupil of the eye," Nat. Med. 20(7), 785–789 (2014).
82. Y. Geng, K. P. Greenberg, R. Wolfe, D. C. Gray, J. J. Hunter, A. Dubra, J. G. Flannery, D. R. Williams, and J. Porter, "In vivo imaging of microscopic structures in the rat retina," Invest. Ophthalmol. Vis. Sci. 50(12), 5872–5879 (2009).
83. S. Bonora, Y. Jian, P. Zhang, A. Zam, E. N. Pugh, R. J. Zawadzki, and M. V. Sarunic, "Wavefront correction and high-resolution in vivo OCT imaging with an objective integrated multi-actuator adaptive lens," Biomed. Opt. Express 23(17), 21931–21941 (2015).
84. A. Jesacher and M. J. Booth, "Sensorless Adaptive Optics for Microscopy," in Adaptive Optics for Biological Imaging, S. S. Olivier, T. G. Bifano, and J. A. Kubby, eds. (2013), pp. 177–190.
85. X. Zhou, P. Bedggood, B. Bui, C. T. O. Nguyen, Z. He, and A. Metha, "Contrast-based sensorless adaptive optics for retinal imaging," Biomed. Opt. Express 6(9), 3577–3595 (2015).
86. M. J. Booth, "Adaptive optical microscopy: the ongoing quest for a perfect image," Light Sci. Appl. 3(4), 165 (2014).
87. R. Navarro and M. A. Losada, "Aberrations and relative efficiency of light pencils in the living human eye," Optom. Vis. Sci. 74(7), 540–547 (1997).
88. C. Wang, R. Liu, D. E. Milkie, W. Sun, Z. Tan, A. Kerlin, T.-W. Chen, D. S. Kim, and N. Ji, "Multiplexed aberration measurement for deep tissue imaging in vivo.," Nat. Methods 11(10), 1037–40 (2014).
89. Y. Jian, K. Wong, and M. V Sarunic, "Graphics processing unit accelerated optical coherence tomography processing at megahertz axial scan rate and high resolution video rate volumetric rendering," J. Biomed. Opt. 18(2), 26002 (2013).
90. K. Wang, W. Sun, C. T. Richie, B. K. Harvey, E. Betzig, and N. Ji, "Direct
119
wavefront sensing for high-resolution in vivo imaging in scattering tissue," Nat. Commun. 6(1), 7276 (2015).
91. J. Mocci, M. Quintavalla, C. Trestino, S. Bonora, and R. Muradore, "A multi-platform CPU-based architecture for cost-effective adaptive optics systems," IEEE Trans. Ind. Informatics 14(10), 4431–4439 (2018).
92. P. Pozzi, D. Wilding, O. Soloviev, H. Verstraete, L. Bliek, G. Vdovin, and M. Verhaegen, "High speed wavefront sensorless aberration correction in digital micromirror based confocal microscopy," Opt. Express 25(2), 949–959 (2017).
93. S. Bonora and R. J. Zawadzki, "Wavefront sensorless modal deformable mirror correction in adaptive optics: optical coherence tomography.," Opt. Lett. 38(22), 4801–4804 (2013).
94. R. A. Muller and A. Buffington, "Real-time correction of atmospherically degraded telescope images through image sharpening," J. Opt. Soc. Am. 64(9), 1200–1210 (1974).
95. D. Debarre, M. J. Booth, and T. Wilson, "Image based adaptive optics through optimisation of low spatial frequencies," Opt. Express 15(13), 8176–8190 (2007).
96. M. Kasper, E. Fedrigo, D. P. Looze, H. Bonnet, L. Ivanescu, and S. Oberti, "Fast calibration of high-order adaptive optics systems.," J. Opt. Soc. Am. A. Opt. Image Sci. Vis. 21(6), 1004–1008 (2004).
97. N. S. Alexander, G. Palczewska, P. Stremplewski, M. Wojtkowski, T. S. Kern, and K. Palczewski, "Image registration and averaging of low laser power two-photon fluorescence images of mouse retina," Biomed. Opt. Express 7(7), 2671–2691 (2016).
98. B. Wang and M. J. Booth, "Optimum deformable mirror modes for sensorless adaptive optics," Opt. Commun. 282(23), 4467–4474 (2009).
99. T. DuBose, D. Nankivil, F. LaRocca, G. Waterman, K. Hagan, J. Polans, B. Keller, D. Tran-Viet, L. Vajzovic, A. N. Kuo, C. A. Toth, J. A. Izatt, and S. Farsiu, "Handheld adaptive optics scanning laser ophthalmoscope," Optica 5(9), 1027–1036 (2018).
100. M. Booth, "Wave front sensor-less adaptive optics: a model-based approach using sphere packings," Opt. Express 14(4), 1339–1352 (2006).
101. H. R. G. W. Verstraete, S. Wahls, J. Kalkman, and M. Verhaegen, "Model-based sensor-less wavefront aberration correction in optical coherence tomography," Opt. Lett. 40(24), 5722–5725 (2015).
102. N. Ji, "Adaptive optical fluorescence microscopy," Nat. Methods 14(4), 374–380 (2017).
103. J. Polans, B. Keller, O. M. Carrasco-Zevallos, F. LaRocca, E. Cole, H. E. Whitson, E. M. Lad, S. Farsiu, and J. A. Izatt, "Wide-field retinal optical coherence tomography with wavefront sensorless adaptive optics for enhanced imaging of targeted regions.," Biomed. Opt. Express 8(1), 16–37 (2017).
104. A. Facomprez, E. Beaurepaire, and D. Débarre, "Accuracy of correction in modal sensorless adaptive optics," Opt. Express 20(3), 2598–2612 (2012).
105. C. Bourgenot, C. D. Saunter, G. D. Love, and J. M. Girkin, "Comparison of closed loop and sensorless adaptive optics in widefield optical microscopy," J. Eur. Opt.
120
Soc. Rapid Publ. 8, 13027 (2013).
106. P. Zhang, M. Goswami, R. J. Zawadzki, E. N. Pugh, and Jr, "The Photosensitivity of Rhodopsin Bleaching and Light-Induced Increases of Fundus Reflectance in Mice Measured In Vivo With Scanning Laser Ophthalmoscopy.," Invest. Ophthalmol. Vis. Sci. 57(8), 3650–3664 (2016).
107. G. Huber, S. C. Beck, C. Grimm, A. Sahaboglu-Tekgoz, F. Paquet-Durand, A. Wenzel, P. Humphries, T. M. Redmond, M. W. Seeliger, and M. D. Fischer, "Spectral Domain Optical Coherence Tomography in Mouse Models of Retinal Degeneration," Investig. Opthalmology Vis. Sci. 50(12), 5888–5895 (2009).
108. X. Tao, O. Azucena, M. Fu, Y. Zuo, D. C. Chen, and J. Kubby, "Adaptive optics microscopy with direct wavefront sensing using fluorescent protein guide stars," Opt. Lett. 36(17), 3389–3391 (2011).
109. M. Rueckel, J. A. Mack-Bucher, and W. Denk, "Adaptive wavefront correction in two-photon microscopy using coherence-gated wavefront sensing.," Proc. Natl. Acad. Sci. U. S. A. 103(46), 17137–17142 (2006).
110. J. B. Schallek and A. Joseph, "Time-lapse imaging of retinal microglia in vivo show dynamic process motility at rest," Invest. Ophthalmol. Vis. Sci. 58(8), 316 (2017).
111. M. Pircher and R. J. Zawadzki, "Review of adaptive optics OCT (AO-OCT): principles and applications for retinal imaging [Invited]," Biomed. Opt. Express 8(5), 2536–2562 (2017).
112. J. Xu, K. Wong, Y. Jian, and M. V. Sarunic, "Real-time acquisition and display of flow contrast using speckle variance optical coherence tomography in a graphics processing unit," J. Biomed. Opt. 19(2), 26001 (2014).
113. J. Xu, S. Han, C. Balaratnasingam, Z. Mammo, K. S. K. Wong, S. Lee, M. Cua, M. Young, A. Kirker, D. Albiani, F. Forooghian, P. Mackenzie, A. Merkur, D.-Y. Yu, and M. V Sarunic, "Retinal angiography with real-time speckle variance optical coherence tomography," Br. J. Ophthalmol. 99(10), 1315–1319 (2015).
114. A. Negrean and H. D. Mansvelder, "Optimal lens design and use in laser-scanning microscopy," Biomed. Opt. Express 5(5), 1588–1609 (2014).
115. A. Nagler, "Plossl type eyepiece for use in astronomical instruments," U.S. patent 4482217 (February 18, 1983).
116. A. Nimmerjahn, F. Kirchhoff, and F. Helmchen, "Resting microglial cells are highly dynamic surveillants of brain parenchyma in vivo," Science 308(5726), 1314–1318 (2005).
117. M. V Sarunic, A. Yazdanpanah, E. Gibson, J. Xu, Y. Bai, S. Lee, H. U. Saragovi, and M. F. Beg, "Longitudinal study of retinal degeneration in a rat using spectral domain optical coherence tomography," Opt. Express 18(22), 23435–23441 (2010).
118. F. C. Delori, R. H. Webb, and D. H. Sliney, "Maximum permissible exposures for ocular safety (ANSI 2000), with emphasis on ophthalmic devices," J. Opt. Soc. Am. A 24(5), 1250–1265 (2007).
119. D. W. Arathorn, Q. Yang, C. R. Vogel, Y. Zhang, P. Tiruveedhula, and A. Roorda, "Retinally stabilized cone-targeted stimulus delivery," Opt. Express 15(21),
121
13731–13744 (2007).
120. D. J. Wahl, P. Zhang, Y. Jian, S. Bonora, R. J. Zawadzki, and M. V. Sarunic, "Image-based adaptive optics compared to wavefront sensing methods for retinal imaging," Proc. SPIE 10502, 105020G (2018).
121. C. Schwarz, R. Sharma, W. S. Fischer, M. Chung, G. Palczewska, K. Palczewski, D. R. Williams, and J. J. Hunter, "Safety assessment in macaques of light exposures for functional two-photon ophthalmoscopy in humans.," Biomed. Opt. Express 7(12), 5148–5169 (2016).
122. D. Merino and P. Loza-Alvarez, "Adaptive optics scanning laser ophthalmoscope imaging: technology update," Clin. Ophthalmol. 10, 743–755 (2016).
123. P. Stremplewski, K. Komar, K. Palczewski, M. Wojtkowski, and G. Palczewska, "Periscope for noninvasive two-photon imaging of murine retina in vivo.," Biomed. Opt. Express 6(9), 3352–3361 (2015).
124. Y. Imanishi, M. L. Batten, D. W. Piston, W. Baehr, and K. Palczewski, "Noninvasive two-photon imaging reveals retinyl ester storage structures in the eye.," J. Cell Biol. 164(3), 373–383 (2004).
125. W. Drexler and J. G. Fujimoto, Optical Coherence Tomography: Technology and Applications, 2nd ed. (Springer, 2015).
126. M. Heisler, S. Lee, Z. Mammo, Y. Jian, M. Ju, A. Merkur, E. Navajas, C. Balaratnasingam, M. F. Beg, and M. V. Sarunic, "Strip-based registration of serially acquired optical coherence tomography angiography," J. Biomed. Opt. 22(3), 036007 (2017).
127. D. C. Adler, T. H. Ko, and J. G. Fujimoto, "Speckle reduction in optical coherence tomography images by use of a spatially adaptive wavelet filter," Opt. Lett. 29(24), 2878–2880 (2004).
128. M. Salas, M. Augustin, L. Ginner, A. Kumar, B. Baumann, R. Leitgeb, W. Drexler, S. Prager, J. Hafner, U. Schmidt-Erfurth, and M. Pircher, "Visualization of micro-capillaries using optical coherence tomography angiography with and without adaptive optics," Biomed. Opt. Express 8(1), 207–222 (2017).
129. M. T. Bernucci, C. W. Merkle, and V. J. Srinivasan, "Investigation of artifacts in retinal and choroidal OCT angiography with a contrast agent," Biomed. Opt. Express 9(3), 1020–1040 (2018).
130. A. Guevara-Torres, D. R. Williams, and J. B. Schallek, "Imaging translucent cell bodies in the living mouse retina without contrast agents," Biomed. Opt. Express 6(6), 2106–2119 (2015).
131. J. I. W. Morgan, J. J. Hunter, B. Masella, R. Wolfe, D. C. Gray, W. H. Merigan, F. C. Delori, and D. R. Williams, "Light-Induced Retinal Changes Observed with High-Resolution Autofluorescence Imaging of the Retinal Pigment Epithelium," Investig. Opthalmology Vis. Sci. 49(8), 3715–3729 (2008).
132. W. T. Ham, H. A. Mueller, J. J. Ruffolo, and A. M. Clarke, "Sensitivity of the retina to radiation damage as a function of wavelength," Photochem. Photobiol. 29(4), 735–743 (1979).
133. G. Palczewska, T. Maeda, Y. Imanishi, W. Sun, Y. Chen, D. R. Williams, D. W. Piston, A. Maeda, and K. Palczewski, "Noninvasive multiphoton fluorescence
122
microscopy resolves retinol and retinal condensation products in mouse eyes," Nat. Med. 16(12), 1444–1449 (2010).
134. S. He, C. Ye, Q. Sun, C. K. S. Leung, and J. Y. Qu, "Label-free nonlinear optical imaging of mouse retina," 6(3), 1055–1066 (2015).
135. T. Kamali, J. Fischer, S. Farrell, W. H. Baldridge, G. Zinser, and B. C. Chauhan, "Simultaneous in vivo confocal reflectance and two-photon retinal ganglion cell imaging based on a hollow core fiber platform," J. Biomed. Opt. 23(9), 091405 (2018).
136. A. S. Bar-Noam, N. Farah, and S. Shoham, "Correction-free remotely scanned two-photon in vivo mouse retinal imaging," Light Sci. Appl. 5(1), 16007 (2016).
137. C. Schwarz, R. Sharma, S. K. Cheong, M. Keller, D. R. Williams, and J. J. Hunter, "Selective S Cone Damage and Retinal Remodeling Following Intense Ultrashort Pulse Laser Exposures in the Near-Infrared," Invest. Ophthalmol. Vis. Sci. 59(15), 5973–5984 (2018).
138. Z. Liu, K. Kurokawa, F. Zhang, J. J. Lee, and D. T. Miller, "Imaging and quantifying ganglion cells and other transparent neurons in the living human retina.," Proc. Natl. Acad. Sci. U. S. A. 114(48), 12803–12808 (2017).
139. Z. Liu, J. Tam, O. Saeedi, and D. X. Hammer, "Trans-retinal cellular imaging with multimodal adaptive optics.," Biomed. Opt. Express 9(9), 4246–4262 (2018).
140. P. Pozzi, O. Soloviev, D. Wilding, G. Vdovin, and M. Verhaegen, "Optimal model-based sensorless adaptive optics for epifluorescence microscopy," PLoS One 13(3), e0194523 (2018).
141. A. Thayil and M. J. Booth, "Self calibration of sensorless adaptive optical microscopes," J. Eur. Opt. Soc. 6, 11045 (2011).
142. K. A. Sapoznik, T. Luo, A. de Castro, L. Sawides, R. L. Warner, and S. A. Burns, "Enhanced retinal vasculature imaging with a rapidly configurable aperture," Biomed. Opt. Express 9(3), 1323–1333 (2018).
143. S. N. Blandford, M. L. Hooper, T. Yabana, B. C. Chauhan, W. H. Baldridge, and S. R. M. Farrell, "Retinal Characterization of the Thy1-GCaMP3 Transgenic Mouse Line After Optic Nerve Transection," Investig. Opthalmology Vis. Sci. 60(1), 183 (2019).
144. W. Wei, J. Elstrott, and M. B. Feller, "Two-photon targeted recording of GFP-expressing neurons for light responses and live-cell imaging in the mouse retina," Nat. Protoc. 5(7), 1347–1352 (2010).
145. B. G. Borghuis, L. Tian, Y. Xu, S. S. Nikonov, N. Vardi, B. V Zemelman, and L. L. Looger, "Imaging light responses of targeted neuron populations in the rodent retina.," J. Neurosci. 31(8), 2855–67 (2011).
146. D. Schweitzer, M. Hammer, F. Schweitzer, R. Anders, T. Doebbecke, S. Schenke, E. R. Gaillard, and E. R. Gaillard, "In vivo measurement of time-resolved autofluorescence at the human fundus," J. Biomed. Opt. 9(6), 1214 (2004).
147. C. Dysli, M. Dysli, V. Enzmann, S. Wolf, and M. S. Zinkernagel, "Fluorescence Lifetime Imaging of the Ocular Fundus in Mice," Investig. Opthalmology Vis. Sci. 55(11), 7206–7215 (2014).
148. C. Dysli, S. Wolf, M. Y. Berezin, L. Sauer, M. Hammer, and M. S. Zinkernagel,
123
"Fluorescence lifetime imaging ophthalmoscopy," Prog. Retin. Eye Res. 60, 120–143 (2017).
149. J. A. Feeks and J. J. Hunter, "Adaptive optics two-photon excited fluorescence lifetime imaging ophthalmoscopy of exogenous fluorophores in mice," Biomed. Opt. Express 8(5), 2483 (2017).
