ArticlePublisher preview available

Real-time frame-registered resonant fluorescence scanning laser ophthalmoscopy for quantifying static and dynamic cellular properties in the mouse retina

Optica Publishing Group
Optics Letters
Authors:
Tianqi Song
Tianqi Song
Tianqi Song
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Yao Wang
Yao Wang
Yao Wang
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Yuxiang Zhou
Yuxiang Zhou
Yuxiang Zhou
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Mingliang Zhou
Mingliang Zhou
Mingliang Zhou
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Show all 9 authorsHide
Read publisher preview
Download citation
To read the full-text of this research, you can request a copy directly from the authors.

Abstract and Figures

Fluorescence labeling offers excellent contrast for cell imaging within living mouse eyes. High-speed, high-resolution imaging with a large field of view (FOV) is always desirable. A high-speed scanning laser ophthalmoscopy (SLO) system has been developed, equipped with real-time desinusoiding correction and frame registration for fluorescence imaging of mouse retinas. Precise calibration using a standard raster grid compensates for scanning hysteresis and image lateral distortion caused by the sinusoidal motion of the resonant scanner. More importantly, a strip-based registration method has been developed to correct frame distortions induced by breathing and pupil drift, enabling effective real-time and post-processing frame averaging. This system captures images at 1024 × 1024 pixels, with a temporal resolution of 16 Hz and a lateral resolution of 1.8 µm, and a FOV of up to 50° (35 µm/degree), which facilitates accurate measurement of both static and dynamic cellular properties, such as microglia cell density, diameter, spacing, and blood hemodynamics, within living mouse eyes.
System sketch. (a) Schematic showing illumination and reflection paths along with electronic components. (b) SolidWorks 3D rendering model. (c) ZEMAX optical model. Pupil wander is shown in the bottom left corner (28.8° × 28.8°, Supplement 1). Abbreviations: CL, collimating lens; BS, beam splitter [70 (Transmission):30 (Reflectance)]; R, resonant scanner; G, galvanometer; L, lens; F, bandpass filters (45 nm band centered at 525 nm); PH, pinhole; PMT, photomultiplier tube; AMP, amplifier.
System sketch. (a) Schematic showing illumination and reflection paths along with electronic components. (b) SolidWorks 3D rendering model. (c) ZEMAX optical model. Pupil wander is shown in the bottom left corner (28.8° × 28.8°, Supplement 1). Abbreviations: CL, collimating lens; BS, beam splitter [70 (Transmission):30 (Reflectance)]; R, resonant scanner; G, galvanometer; L, lens; F, bandpass filters (45 nm band centered at 525 nm); PH, pinhole; PMT, photomultiplier tube; AMP, amplifier.
… 
Calibration for distortion and hysteresis. (a) Original image (left) and binarized plot (right) before correction and (b) after correction. (c) Scanning displacements with time in cubic polynomial and linear scan models. (d) Comparison between two methods. Abbreviations: AMP, amplitude; SD, standard deviation; MAE, mean absolute error.
Calibration for distortion and hysteresis. (a) Original image (left) and binarized plot (right) before correction and (b) after correction. (c) Scanning displacements with time in cubic polynomial and linear scan models. (d) Comparison between two methods. Abbreviations: AMP, amplitude; SD, standard deviation; MAE, mean absolute error.
… 
Real-time registration process and eye movement analysis. (a) Algorithmic flow chart including preprocessing, full-frame registration, and strip registration. (b) Result diagram. The left section shows coarse full-frame and fine horizontal and vertical strip-level motion trace extraction. The right section shows the rolling average result of every four frames and the overall average result of 1000 frames in total. (c) Eye motion analysis in time domain (left) and frequency domain (Supplement 1).
Real-time registration process and eye movement analysis. (a) Algorithmic flow chart including preprocessing, full-frame registration, and strip registration. (b) Result diagram. The left section shows coarse full-frame and fine horizontal and vertical strip-level motion trace extraction. The right section shows the rolling average result of every four frames and the overall average result of 1000 frames in total. (c) Eye motion analysis in time domain (left) and frequency domain (Supplement 1).
… 
Fluorescence images, cell distribution, and cell dynamic changes of microglia in vivo. (a) Represented average image (FOV 20°×20°). See Visualization 1. (b) Quantification, ${\mathbf{\Phi }_{\mathbf{cell}}}$ , represents the diameter of cell bodies, ${\boldsymbol{d}_\mathbf{s}}$ represents the spacing between cells, ${\mathbf{\rho }_{\mathbf{cell}}}$ represents the density of cells and the relationship among them (bottom right). (c) Microglial protrusion motility process (arrowheads in each image), location shown by the green dashed rectangle in (a), including extensions (blue) and retractions (yellow) of the process from a 250 s recording. (d) Protrusions’ lengths change over time during retractions (P1, P2) and extensions (P3, P4). (e) Average velocity in four positions.
Fluorescence images, cell distribution, and cell dynamic changes of microglia in vivo. (a) Represented average image (FOV 20°×20°). See Visualization 1. (b) Quantification, ${\mathbf{\Phi }_{\mathbf{cell}}}$ , represents the diameter of cell bodies, ${\boldsymbol{d}_\mathbf{s}}$ represents the spacing between cells, ${\mathbf{\rho }_{\mathbf{cell}}}$ represents the density of cells and the relationship among them (bottom right). (c) Microglial protrusion motility process (arrowheads in each image), location shown by the green dashed rectangle in (a), including extensions (blue) and retractions (yellow) of the process from a 250 s recording. (d) Protrusions’ lengths change over time during retractions (P1, P2) and extensions (P3, P4). (e) Average velocity in four positions.
… 
Imaging fluorescent markers in blood and interstitial fluid. (a) 2D scan of SLO and motion trajectory of the fluorescent marker in Cartesian coordinates. (b) Fluorescent image of blood vessels in the mouse retina. (c) Zoom-in scans of the green dashed rectangle area in (b) are shown in pseudo-color. Vessel boundaries are marked by the red lines. (d) Moving fluorescence object flashes along the vessel. See Visualization 2. (e) Position of a slowly moving fluorescent marker in interstitial fluid over eight consecutive frames with a maximum intensity projection. See Visualization 3. (f) Histograms of fast-moving fluorescent markers (n = 149) over 10 min: averaged blood velocity (top left) and flow (top right). Velocity plots for the marker in (e) (bottom left), the average velocity and duration of signals across 13 sets (bottom right). Abbreviations: $\overline {\mathbf{Vel}.} $ , average velocity.
Imaging fluorescent markers in blood and interstitial fluid. (a) 2D scan of SLO and motion trajectory of the fluorescent marker in Cartesian coordinates. (b) Fluorescent image of blood vessels in the mouse retina. (c) Zoom-in scans of the green dashed rectangle area in (b) are shown in pseudo-color. Vessel boundaries are marked by the red lines. (d) Moving fluorescence object flashes along the vessel. See Visualization 2. (e) Position of a slowly moving fluorescent marker in interstitial fluid over eight consecutive frames with a maximum intensity projection. See Visualization 3. (f) Histograms of fast-moving fluorescent markers (n = 149) over 10 min: averaged blood velocity (top left) and flow (top right). Velocity plots for the marker in (e) (bottom left), the average velocity and duration of signals across 13 sets (bottom right). Abbreviations: $\overline {\mathbf{Vel}.} $ , average velocity.
… 
Figures - available from: Optics Letters
This content is subject to copyright. Terms and conditions apply.
A preview of this full-text is provided by Optica Publishing Group.
Content available from Optics Letters 
This content is subject to copyright. Terms and conditions apply.
Letter Vol. 50,No. 4 /15 February 2025 / Optics Letters 1329
Real-time frame-registered resonant fluorescence
scanning laser ophthalmoscopy for quantifying
static and dynamic cellular properties
in the mouse retina
Tianqi Song,1Yao Wang,2,3 Yuxiang Zhou,1Mingliang Zhou,1Yanhong Ma,1
Donghan Ma,1Jia Qu,2,3,4,†Jinyi Zhang,3,5,†AND Pengfei Zhang1,3,†,∗
1School of Optoelectronic Engineering and Instrumentation Science, Dalian University of Technology, Dalian, Liaoning 116024, China
2National Clinical Research Center for Ocular Diseases, Eye Hospital, Wenzhou Medical University, Wenzhou, China
3Oujiang Laboratory, Zhejiang Lab for Regenerative Medicine, Vision and Brain Health, Wenzhou, Zhejiang 325000, China
4jia.qu@163.com
5zhangjinyi@ojlab.ac.cn
†These three authors are co-corresponding authors to this work.
*pfzhang@dlut.edu.cn
Received 13 November 2024; revised 1 January 2025; accepted 12 January 2025; posted 22 January 2025; published 12 February 2025
Fluorescence labeling offers excellent contrast for cell imag-
ing within living mouse eyes. High-speed, high-resolution
imaging with a large field of view (FOV) is always desirable.
A high-speed scanning laser ophthalmoscopy (SLO) system
has been developed, equipped with real-time desinusoiding
correction and frame registration for fluorescence imag-
ing of mouse retinas. Precise calibration using a standard
raster grid compensates for scanning hysteresis and image
lateral distortion caused by the sinusoidal motion of the res-
onant scanner. More importantly, a strip-based registration
method has been developed to correct frame distortions
induced by breathing and pupil drift, enabling effective
real-time and post-processing frame averaging. This system
captures images at 1024 ×1024 pixels, with a temporal res-
olution of 16 Hz and a lateral resolution of 1.8 µm, and a
FOV of up to 50°(35 µm/degree), which facilitates accurate
measurement of both static and dynamic cellular proper-
ties, such as microglia cell density, diameter, spacing, and
blood hemodynamics, within living mouse eyes. © 2025
Optica Publishing Group. All rights, including for text and data mining
(TDM), Artificial Intelligence (AI) training, and similar technologies,
are reserved.
https://doi.org/10.1364/OL.546343
The mouse is a critical biological model for both basic science
and preclinical ophthalmology research. Fluorescence labeling
is a powerful tool for visualizing the mouse retina at a cellular
resolution with high contrast. Scanning laser ophthalmoscopy
(SLO), due to its confocal nature, effectively eliminates out-of-
focus signals from the background. This enhances image contrast
and enables high-quality imaging even for weak fluorescent-
labeled cells [1,2].
SLO generates images through two-dimensional point scan-
ning; one way is to use a two-axis galvanometer scanner to create
a raster scan pattern on the retina. While galvanometer scan-
ners with linear scanning can facilitate image reconstruction,
the speed is limited. For instance, imaging at 250 ×250 pixels
with a 250 kHz sampling frequency may result in a frame rate of
less than 2 Hz due to its flyback time [3]. This frame rate is gener-
ally adequate for imaging the static structure of the mouse retina
under anesthesia; however, it poses challenges when imaging
the dynamic changes of retinal structures, or when attempting
to capture a larger FOV. Additionally, during prolonged experi-
ments, breathing and pupil drift can introduce persistent motion
artifacts that degrade image quality. The low signal-to-noise
ratio (SNR) in a single frame for weak fluorescence-labeled cells
further impedes researchers from quickly identifying optimal
regions, ultimately reducing experimental efficiency. Therefore,
high-speed fluorescence imaging with high pixel sampling and
real-time registered averaging is highly desirable.
Alternatively, replacing the fast-axis galvanometer scanner
with a resonant scanner can improve both the frame rate and
sampling density. For example, an 8 kHz resonant scanner com-
bined with a 25 MHz sampling rate can achieve a frame rate of
16 Hz at a resolution of 1562 ×1024 pixels. However, due to
its sinusoidal scanning pattern, real-time correction of scanning
distortions is necessary to produce a linear grid image, which
can be challenging in high-speed imaging. To address these
problems, here we develop a resonant fluorescence SLO sys-
tem integrated with real-time scanning distortion correction and
frame registration modules. This system can correct motion at
both the full-frame and strip levels, and simultaneously displays
corrected and averaged videos in real time.
The SLO system is shown in Fig. 1(details in Supplement 1).
The sinusoidal scanning pattern of a resonant scanner leads
to lateral distortion for a grid sample (Fig. 2(a)). In addition,
a slight delay exists between the scanning and synchronizing
signals due to the accuracy limitations of the drive circuit [4].
To accurately determine the scanning displacement curve over
time, we utilized a standard grid to obtain 40 discrete points
0146-9592/25/041329-04 Journal ©2025 Optica Publishing Group
Corrected 18 February 2025
Alignment, calibration, and validation of an adaptive optics scanning laser ophthalmoscope for high-resolution human foveal imaging
Article
Full-text available
  • Jan 2024
In prior art, advances in adaptive optics scanning laser ophthalmoscope (AOSLO) technology have enabled cones in the human fovea to be resolved in healthy eyes with normal vision and low to moderate refractive errors, providing new insight into human foveal anatomy, visual perception, and retinal degenerative diseases. These high-resolution ophthalmoscopes require careful alignment of each optical subsystem to ensure diffraction-limited imaging performance, which is necessary for resolving the smallest foveal cones. This paper presents a systematic and rigorous methodology for building, aligning, calibrating, and testing an AOSLO designed for imaging the cone mosaic of the central fovea in humans with cellular resolution. This methodology uses a two-stage alignment procedure and thorough system testing to achieve diffraction-limited performance. Results from retinal imaging of healthy human subjects under 30 years of age with refractive errors of less than 3.5 diopters using either 680 nm or 840 nm light show that the system can resolve cones at the very center of the fovea, the region where the cones are smallest and most densely packed.
View
Strip-based digital image registration for distortion minimization and robust eye motion measurement from scanned ophthalmic imaging systems
Article
Full-text available
  • Mar 2021
Retinal image-based eye motion measurement from scanned ophthalmic imaging systems, such as scanning laser ophthalmoscopy, has allowed for precise real-time eye tracking at sub-micron resolution. However, the constraints of real-time tracking result in a high error tolerance that is detrimental for some eye motion measurement and imaging applications. We show here that eye motion can be extracted from image sequences when these constraints are lifted, and all data is available at the time of registration. Our approach identifies and discards distorted frames, detects coarse motion to generate a synthetic reference frame and then uses it for fine scale motion tracking with improved sensitivity over a larger area. We demonstrate its application here to tracking scanning laser ophthalmoscopy (TSLO) and adaptive optics scanning light ophthalmoscopy (AOSLO), and show that it can successfully capture most of the eye motion across each image sequence, leaving only between 0.1-3.4% of non-blink frames untracked, while simultaneously minimizing image distortions induced from eye motion. These improvements will facilitate precise measurement of fixational eye movements (FEMs) in TSLO and longitudinal tracking of individual cells in AOSLO.
View
In vivo imaging reveals transient microglia recruitment and functional recovery of photoreceptor signaling after injury
Article
Full-text available
  • Jul 2019
Significance Microglia, the resident macrophages of the central nervous system, are critical for synaptic pruning and maintenance and for mitigating injury and neurodegeneration. Determining whether microglia–neuron interactions are beneficial in specific instances has been difficult, largely because of the local and transient nature of the interactions. Using simultaneous optical coherence tomography/scanning laser ophthalmoscopy (SLO) and adaptive optics SLO retinal imaging in mice, we show interactions of microglia and photoreceptors over time scales from seconds to months during injury, degeneration, and repair. In vivo optical assessment of photoreceptor signaling in a large neuronal field encompassing the injured area allows us to relate the time course of these microglia movements to that of the tissue remodeling and functional recovery.
View
Imaging single-cell blood flow in the smallest to largest vessels in the living retina
Article
Full-text available
  • May 2019
  • eLife
Tissue light scatter limits the visualization of the microvascular network deep inside the living mammal. The transparency of the mammalian eye provides a noninvasive view of the microvessels of the retina, a part of the central nervous system. Despite its clarity, imperfections in the optics of the eye blur microscopic retinal capillaries, and single blood cells flowing within. This limits early evaluation of microvascular diseases that originate in capillaries. To break this barrier, we use 15 kHz adaptive optics imaging to noninvasively measure single-cell blood flow, in one of the most widely used research animals: the C57BL/6J mouse. Measured flow ranged four orders of magnitude (0.0002–1.55 µL min–1) across the full spectrum of retinal vessel diameters (3.2–45.8 µm), without requiring surgery or contrast dye. Here, we describe the ultrafast imaging, analysis pipeline and automated measurement of millions of blood cell speeds.
View
Motion Detection and Compensation of Affine Deformations in Infrared Retinal Videos
Article
Full-text available
  • Oct 2016
Diabetic macular edema (DME) affects the retina and reduces the visual acuity of patients with severe diabetic retinopathy. Its conventional treatment involves laser photocoagulation combined with infrared (IR) imaging. However, the laser beam may hit healthy retinal areas and cause unintentional retinal damages if retinal motion occurs. We propose a method for retinal motion detection and compensation that relies on phase correlations in the image and in the log-polar domains, and is robust to affine retinal deformations (e.g., rotations, scales, and translations) in IR videos. The proposed method is also robust to the background noise and illumination changes commonly occurring in this retinal imaging modality. The proposed method can be used to estimate the retinal affine motion parameters and compensate for small retinal motions (nearly 50 μm). The critical method parameters are selected and adjusted optimally, which improves the robustness of the method. The experimental results suggest that the proposed method potentially can be more robust for detecting retinal motion and for estimating the parameters of affine retinal deformations than comparable methods that currently represent the state of the art, which helps to improve the reliability of laser treatments for DME.
View
Closed-loop optical stabilization and digital image registration in adaptive optics scanning light ophthalmoscopy
Article
Full-text available
  • Aug 2014
Eye motion is a major impediment to the efficient acquisition of high resolution retinal images with the adaptive optics (AO) scanning light ophthalmoscope (AOSLO). Here we demonstrate a solution to this problem by implementing both optical stabilization and digital image registration in an AOSLO. We replaced the slow scanning mirror with a two-axis tip/tilt mirror for the dual functions of slow scanning and optical stabilization. Closed-loop optical stabilization reduced the amplitude of eye-movement related-image motion by a factor of 10–15. The residual RMS error after optical stabilization alone was on the order of the size of foveal cones: ~1.66–2.56 μm or ~0.34–0.53 arcmin with typical fixational eye motion for normal observers. The full implementation, with real-time digital image registration, corrected the residual eye motion after optical stabilization with an accuracy of ~0.20–0.25 μm or ~0.04–0.05 arcmin RMS, which to our knowledge is more accurate than any method previously reported.
View
Optimization of confocal scanning laser ophthalmoscope design
Article
Full-text available
  • Jul 2013
  • J BIOMED OPT
Confocal scanning laser ophthalmoscopy (cSLO) enables high-resolution and high-contrast imaging of the retina by employing spatial filtering for scattered light rejection. However, to obtain optimized image quality, one must design the cSLO around scanner technology limitations and minimize the effects of ocular aberrations and imaging artifacts. We describe a cSLO design methodology resulting in a simple, relatively inexpensive, and compact lens-based cSLO design optimized to balance resolution and throughput for a 20-deg field of view (FOV) with minimal imaging artifacts. We tested the imaging capabilities of our cSLO design with an experimental setup from which we obtained fast and high signal-to-noise ratio (SNR) retinal images. At lower FOVs, we were able to visualize parafoveal cone photoreceptors and nerve fiber bundles even without the use of adaptive optics. Through an experiment comparing our optimized cSLO design to a commercial cSLO system, we show that our design demonstrates a significant improvement in both image quality and resolution.
View
High-speed, image-based eye tracking with a scanning laser ophthalmoscope
Article
Full-text available
  • Sep 2012
We demonstrate a high-speed, image-based tracking scanning laser ophthalmoscope (TSLO) that can provide high fidelity structural images, real-time eye tracking and targeted stimulus delivery. The system was designed for diffraction-limited performance over an 8° field of view (FOV) and operates with a flexible field of view of 1°–5.5°. Stabilized videos of the retina were generated showing an amplitude of motion after stabilization of 0.2 arcmin or less across all frequencies. In addition, the imaging laser can be modulated to place a stimulus on a targeted retinal location. We show a stimulus placement accuracy with a standard deviation less than 1 arcmin. With a smaller field size of 2°, individual cone photoreceptors were clearly visible at eccentricities outside of the fovea.
View
Real-time correction of transverse dislocation and distortion in probe-based confocal laser endomicroscopy imaging
Article
  • Jun 2022
  • OPT COMMUN
High speed scanning system based on resonant scanner can improve the imaging speed of probe-based confocal laser endomicroscopy (pCLE), which makes pCLE be a very effective tool to observe cell tissue morphology in vivo. However, since the speed of resonant scanner changes nonlinearly, transverse distortion will occur when the digitizer collects the image data at a constant rate. Moreover, accurate turning time is hard to be obtained due to the limitation of scanner hardware, although synchronous square wave signals are used for reference. Therefore, it is difficult to accurately segment the bidirectional scanned data into two adjacent lines of the image, which results in image dislocation. In this work, the offset value of each line is obtained by estimating the overall dislocation and local dislocation of the image respectively. Then, the start time of the sampling and sampling interval are adjusted based on the scanner speed to correct the image in real-time by FPGA. As a result, the dislocation and distortion of the image are corrected at video rates.
View
Geometric distortion correction for sinusoidally scanned images
Article
  • Oct 2011
A method for correcting the geometric distortion of sinusoidally scanned images was proposed. The generation mechanism of the geometric distortion in sinusoidally scanned images was analyzed. Based on the relationship between the coordinates of uniformly scanned points and those of sinusoidally scanned points, a transformation formula was obtained for correcting the geometric distortion when the sampling rate was a constant. By comparing the forward method with the inverse method, a hybrid method for correcting the geometric distortion of sinusoidally scanned images was proposed. This method takes advantage of both the forward and inverse methods and was proven to be better than either of them in terms of peak signal-to-noise ratio (PSNR). The time consumed by the hybrid method was between the other two. When a higher PSNR is desired, the hybrid method is recommended if time permits. In addition, it is a universal approach to the correction of geometric distortion of the images scanned in the sinusoidal mode.
View