Scanning Laser Ophthalmoscopy (SLO)
Since the first scanning laser ophthalmoscope (SLO) was introduced in the early 1980s, this imaging technique has been adapted and optimized for various clinical applications based on different contrast mechanism. Reflectance imaging, where the back scattered light is detected, is widely used for eye tracking and as reference image for OCT applications. But also the reflectance modality itself has several important diagnostic applications: laser scanning tomography (SLT), imaging with different laser wavelengths (Multicolor contrast) and others. Fluorescence imaging channels with different excitation wavelengths were introduced to SLOs for angiography, i.e. for the visualization of the vascular system after intravenously injecting an appropriate dye, as well as for autofluorescence imaging of endogenous fluorophores within the retina.
This chapter gives an introduction to the history of SLO technology and a general overview on its clinical applications. In the following sections the focus is directed on one distinct clinical application for each of the major imaging modalities: reflectance, angiography and autofluorescence. First, the assessment of the optic nerve head for glaucoma diagnostic by means of laser scanning tomography is presented, before in the third section the clinical aspect of wide field SLO angiography is discussed. Finally, an introduction to quantitative autofluorescence (qAF) is given, a new research tool, which is used to measure the accumulation of lipofuscin within the retinal pigment epithelium (RPE) for normal subjects and for patients with macular dystrophies.
KeywordsOphthalmoscopy Scanning laser ophthalmoscopy (SLO) Confocal scanning laser ophthalmoscope (cSLO) Resolution Cone imaging Scanning laser tomography (SLT) Optic nerve head topography Glaucoma diagnostics Wide field angiography Ultra wide field imaging Indocyanine green angiography (ICGA) Lipofuscin Autofluorescence (AF) Quantitative autofluorescence (qAF)
2.1 Introduction and Technology
In the following years Robert Webb and his team, but also the Heidelberg group of Josef Bille continuously improved and refined the performance of the then re-baptized Scanning Laser Ophthalmoscope (SLO) [3, 4], including the use of frame grabber boards allowing for conversion to digital images. Image contrast was drastically improved by confocal detection eliminating scattered light originating from outside the focal volume [5, 6]. Laser Scanning Tomography (SLT) was introduced, were a stack of 2D SLO images was acquired with each frame positioned at a different, equally spaced focal plane and from the 3D data stack a tomography image was calculated. (see [7, 8] and Sect. 2.2.2 for details).
Together with ophthalmologists the technical groups started to evaluate the clinical value of the new imaging technique in first clinical studies . A perimeter was integrated into the SLO in order to determine scotoma maps of patients [10, 11], and a study on macular holes and other macular disease was accomplished at the department of ophthalmology in San Diego in cooperation with the Heidelberg team .
Already in 1989 Josef Bille at al. proposed and demonstrated the use of adaptive optics in order to improve the axial resolution of the confocal SLO .
2.1.2 Modern Confocal SLO
2.1.3 SLO Core Components
In the following the technology of core components is discussed:
18.104.22.168 Laser Source
The early SLOs incorporated gas lasers with their superior beam profile and their stable, continuous wave (cw) laser emission. For reflectance imaging usually the red HeNe laser line (helium neon laser) was used, whereas for fluorescence applications the turquoise 488 nm laser line of the Argon laser (Ar+ laser) became very popular. This wavelength matches almost perfectly the absorption maximum of sodium fluorescein, a standard dye, which was used in clinics already since the 1960s for fluorescein angiography (FA) with fundus cameras .
22.214.171.124 Scan Unit
The biggest challenge of a fast 2D scanning system is the high line rate of 8 kHz or more, which is required for the fast scanning direction. In the following it is assumed, that the fast scan axis of the 2D scan pattern is the horizontal X-direction, and the slow axis refers to the vertical Y-scan. Such a fast line scan rate is required in order to achieve frame rates of >15 Hz while maintaining still a sufficient high density for square pixels, i.e. pixels with same separation in X- and Y-direction. In addition to the high speed, also the optical scan angle and the size of the scan pupil, i.e. the mirror size, are crucial parameters, which need to be carefully balanced. A lower primary scan angle can be compensated by appropriate magnification optics, which will, however, at the same time reduce the diameter of the pupil area defining the solid angle under which back scattering or fluorescence emission can be detected.
These demanding specifications can be met by resonant scanners, where the mirror is mounted on a beryllium torsion bar with a counter weight at the other end. The total system of mirror, torsion bar and counter weight has an intrinsic resonance frequency with high Q-value of half the line rate and is excited by a phase locked loop (PLL) circuit to the fast sinusoidal scanner oscillation.
A common alternative to resonant scanners are polygon scanners, where a polygon-shaped mirror is rotated by means of a fast DC motor at up to 60,000 rpm. For a polygon mirror with eight mirror facets, this would correspond to a line frequency of 8 kHz. Again, the size of the scan pupil and the required scan angle limit the number of facets and therefore also the scan rate.
For the slow axis (vertical direction) in general galvanometric scanners are used, which are usually controlled in a closed loop with a saw-tooth ramp.
126.96.36.199 Beam Splitter
Different beam splitter optics (plates or cubes) have been used for separating the back scattered signal light from the incident laser beam path. For fluorescence systems (angiography and autofluorescence) the use of dichroic mirrors is the optimum choice. The dielectric coating of these mirrors is usually designed in a way that the beam splitter plate has a high transmission for the excitation laser wavelength, but shows a strong reflectivity for the “red-shifted” fluorescence emission.
For reflectance imaging, a splitting ratio of 20:80 is usually a good choice: a transmission of 20% is in many cases sufficient, since the maximum laser power, which can be applied to the eye, is limited anyway by laser safety requirements. On the other hand, it is of course desirable to couple as much as possible of the back scattered signal light into the detection branch.
Another option is the use of a polarizing beam splitter: the laser beam is p-polarized and therefore passes through the polarizing beam splitter to a very high percentage. After the beam splitter, a quarter wave plate is inserted, which rotates the polarization due to the double pass (out-going and back scattering beams) by 90° in a way, that the signal light is now s-polarized and thus reflected by the polarizing beam splitter into the detection unit. It is important to note, that the birefringent structures within the eye (e.g. cornea, retinal nerve fiber layer) can cause an inhomogeneous illumination of the image, since the combined action of the quarter wave plate and the birefringent tissue will lead to an imperfect 90° rotation of the polarization.
188.8.131.52 Imaging Optics
Optics are required to relay the scan pupil to the entrance pupil of the eye. In many cases a telecentric 4f design is chosen, as shown schematically in Fig. 2.2. The advantage of such a design is, that by adjusting the distance between the two lens groups (L1 and L2), the convergence or divergence of the laser beam at the scan pupil can be adjusted to correct for the refraction error of the eye, without changing the magnification of the scan angle at the entrance pupil.
By using lens groups with different effective focal lengths, the primary scan angle can be magnified or reduced and the beam diameter at the eye pupil is correspondingly decreased resp. increased. Especially for the design of wide field imaging optics, it is important, to integrate the lens data of wide field model eye into the optics design file in order to achieve over the complete scan field a sharp fundus image.
Whereas in early SLO systems photomultiplier tubes (PMT) were used because of their at that time unique sensitivity, nowadays semiconductor based detectors as avalanche photodiodes (APD) and silicon photomultiplier (SiPM) are the first choice.
APDs consist of a semiconductor p-n junction, where the incident photons create free electrons in the absorption zone. These carriers are sucked to the multiplication zone, where a high reverse voltage leads to their avalanche-like multiplication due to impact ionization. Typically a gain of more than 100 can be achieved, and this gain is proportional to the applied reverse voltage.
APDs can be operated as well in the Geiger mode for single photon counting. In these single photon avalanche photodiodes (SAPD) the applied reverse voltage is above the breakdown voltage, and an initial charge carrier created by a single photon is extremely multiplied up to currents in the range of a milliampere. The leading edge of this current is used to trigger a photon counter. During the high current, the bias voltage drops below breakdown voltage and the device is blind for the detection of new photons. This effect limits the maximum count rate of the SAPDs.
Silicon photomultipliers (SiPM) consist of an array of APDs operated in the Geiger mode. The dimension of such an array is typically in the order of a few millimeters. The dimension of each single SAPD cell is in the range of 10 … 100 μm, i.e. the array typically consists of several 1000 elements. The output channels of the individual cells are connected in one common parallel read-out element. Although each element is operated in the digital Geiger mode, the combined output of the complete device yields an analogue output, which is proportional to the incident photon flux until the saturation level is reached.
2.1.4 Resolution of the SLO
184.108.40.206 Limitations and Numerical Aperture (NA) of the Eye
The optical resolution of the scanning laser ophthalmoscope is limited by the anatomy of the human eye itself. As in any conventional light microscope the minimum spot size of the focused laser in the object plane is limited by diffraction.
For undilated pupils the maximum NAeye is about 0.09 (D = 3 mm) and it can be increased by a factor of 2–3 by dilating the pupil with a drug to D = 6 − 8 mm. However, due to the limited optical quality of the eye and due to the strong increase of the optical aberrations in the periphery, the distortions of the wave front in contrary result in a larger focal volume on the retina and thus decrease the optical resolution compared to undilated pupils [14, 15]. In order to exploit the full diffraction limited resolution for dilated pupils, an adaptive optical (AO) element must be used to compensate the distortions of the optical wave front of the individual eye. With this concept, the lateral resolution can be increased by a factor of 2–3 and the axial resolution even by a factor of 4–9 compared to undilated pupils (see also Chaps. 16– 18). The following considerations are only valid for eyes without optical aberrations, i.e. either for undilated eyes with sufficient optical quality, or for dilated pupils with AO compensation.
In order to calculate the lateral and axial extension of the focus spot, the light propagation integral of the pupil function needs to be solved. In the following the results for two different approaches are summarized: Fraunhofer diffraction at a circular aperture and propagation of a Gaussian beam.
220.127.116.11 Fraunhofer Diffraction at a Circular Aperture
18.104.22.168 Beam Waist for Propagating Gaussian Beam
for rA ≥ 2 · ωp the truncation effect is negligible.
for rA = ωp, the truncation will result in an increase of the focus waist of ≈50%.
for rA ≪ ωp, the focus waist will approximate the result of the Fraunhofer diffraction pattern.
For SLO examination on an undilated eye the limiting aperture is usually the iris (2–3 mm diameter), which is often in the order of the beam diameter of the laser beam (rA ≈ ωp), thus a truncation factor of ×1.5 is in many applications justified.
22.214.171.124 Resolution Improvement Due to Confocal Detection
In order to determine the influence of the confocal aperture on the lateral and axial resolution, the illumination point spread function (PSF, i.e. the 3D intensity distribution of an ideal point source imaged onto the sample) needs to be multiplied with the detection point spread function defined by the image of a point source emitter within the sample onto the pinhole aperture. Thus, in theory the confocal detection could improve the resolution by a factor of 1/√2, however, this enhancement is only achieved when the pinhole size is much smaller than the Airy disk diameter projected into the pinhole plane. Usually, for intensity reasons, this is not the case in commercially available cSLOs.
Summary of lateral and axial resolution resp. focal dimensions for non-dilated and dilated pupils, under the assumption that the optical aberration is perfectly compensated e.g. by adaptive optics
Parameter, Assumption and equation
Undilated pupil (2.5 mm)
Dilated pupil w. AO (7 mm)
Resolution Rayleigh criterion [μm] based on Airy disk δxRC (Eq. 2.4)
Focus spot diameter (1/e2) [μm]
Gauss beam diam. 2ω0 (Eq. 2.8a)
Focus spot diameter for truncated Gauss beam [μm] (rA ≈ ωp)
Axial extension of the focus [μm] (confocal param. b = 2zr, Eq. 2.8b)
2.1.5 Example for High Resolution SLO Image
2.2 Laser Scanning Tomography
Gerhard Zinser and his team developed in 1991 the Heidelberg Retina Tomograph (HRT), a commercial SLO, which was from the beginning dedicated for the diagnosis of glaucoma by assessing the morphology of the optical nerve head.
2.2.1 HRTII/HRT3 Acquisition Work Flow
For laser scanning tomography a set of 2D SLO frames at equally spaced focal positions is acquired. Therefore, between two frames of a series, the focus of the laser is shifted by means of a motorized telescope within the camera head. For the HRTII/HRT3 the data acquisition work flow is as follows:
First the patient is positioned and fixates with the examined eye on a fixation target presented at about 12° nasally , such that the optic nerve head (ONH) appears centered within the 15° × 15° scan field.
The user then aligns the camera head in three dimensions, in order to make sure, that the scan pupil, i.e. the pivot point of the scanning laser beam, coincides exactly with the entrance pupil of the eye. The SLO image is displayed on the monitor and the brightness of the image, i.e. the detector sensitivity, is adjusted automatically (auto-brightness) to make sure, that the signal falls over the complete z-scan within the linear range of the detector and no saturation effects corrupt the data. Once the camera is properly aligned, the acquire button is pressed and the acquisition of three consecutive z-scans is started. The first series consists of a stack of 64 images with a focal plane distance of 62.5 μm, i.e. covering an overall z-range of 4 mm in the eye. From this first series, the required z-scan depth for the two consecutive z-scan is calculated (depending on the ONH geometry 2–4 mm), in order to avoid unnecessary long acquisition time. Finally an automatic quality control check is performed, and if o.k. the series are saved on the computer.
2.2.2 HRTII/HRT3 Data Processing
Each of the three acquired series is corrected for eye movements during the z-scan by laterally matching and shifting each image with respect to the previous image within the scan.
2.2.3 Contour Line, Reference Plane and Stereometric Parameters
After data acquisition and calculation of the reflection and topography images, a contour line around the optic disk needs to be defined, similar as it is the case in cup/disk ratio measurements on fundus images. The contour line is used to calculate stereometric parameters as the rim area, rim volume, cup shape measure and others, which describe the shape of the ONH. In addition, these parameters can be combined by means of several discriminant functions, which have been proven in clinical studies to have a high sensitivity and specificity for glaucoma detection [21, 22, 23, 24].
The ONH contour needs to be manually defined by the physician and some experience is required to correctly place the line around the optic disk. However, it has been shown in a clinical study , that the variability between contour lines drawn by different physicians has only little effect on the parameter data. In addition, the contour line needs to be defined only once for the baseline data and then is automatically transferred to the new image data acquired during follow-up examinations.
2.2.4 Analysis of HRT Optic Nerve Head (ONH) Data
126.96.36.199 ONH Classification Based on Moorfields Regression Analysis
188.8.131.52 Follow-Up and Progression Analysis
In order to differentiate between a progressing glaucomatous and a stable ONH, follow-up examinations acquired at later time points are compared with the original baseline data. Therefore, first the follow-up images are matched to the baseline images, in order to compensate for head tilts, accommodation differences and other external influences and in a second step the baseline contour line is transferred to the new images to enable the calculation of all the follow-up stereometric parameters.
One possibility to visualize the temporal development is to plot the stereo-metric parameters versus time, with the date of the baseline examination as origin of the time axis. Since the absolute changes and also the sign of the changes vary for different parameters, the temporal change of each stereometric parameter is normalized to the difference of this parameter averaged over a healthy group and over a group of glaucoma patients.
Another possibility to visualize the changes in a time series of ONH topography images was proposed by Chauhan et al. . They assessed by means of statistical methods the significance of changes of clustered super-pixels in the topography images and visualized these changes by colored overlays: red indicating a decrease of surface height (i.e. increase of excavation) and green indicating an increase of surface height.
2.2.5 Summary SLT for Glaucoma Diagnostics
Laser scanning tomography was for several years the gold standard for diagnosis and monitoring the progression of glaucoma by assessing the morphology of the optical nerve head. Numerous clinical studies, which have been published in peer-reviewed journals, have shown, that the HRT enables reproducible measurements of the morphology of the optic nerve head [7, 28], has a very high sensitivity and specificity for discriminating patients with early glaucoma from normal subjects [26, 29] and that the success of therapeutic intervention can be reliably assessed by documenting the stagnancy resp. the progression of the disease [27, 30, 31].
Only with the availability of spectral domain OCT systems (see Chap. 3), which provided on one hand a much better axial resolution compared to the confocal SLO and on the other hand a much shorter acquisition time and higher spatial accuracy compared to the previous commercially available time domain OCT device, the demand for laser scanning tomography technology slowly decreased. However, since glaucoma is a slowly progressing disease, where the progression needs to be monitored carefully over years and presently acquired data needs to be compared with baseline and follow-up data acquired years ago, the HRT is still a valuable instrument used on a daily base in clinical practices for monitoring and managing glaucoma patients.
2.3 Widefield Indocyanine Green Angiography (ICGA)
The advent of widefield imaging has for the first time offered the chance to investigate both, the central and the peripheral retina, in a single examination. A wide visualization of the retinal periphery is necessary for the screening, diagnosis, monitoring, and treatment of many diseases. Early diagnosis of peripheral retinal or choroidal disease could reduce a potential vision loss.
Fluorescein angiography (FA) and indocyanine green angiography (ICGA) are two imaging modalities that use a water-soluble dye to visualize retinal and choroidal vasculature .
Also in patients with retinal vein occlusion, ultra widefield angiography may be a powerful tool to identify therapeutic target areas for photocoagulation, allowing for efficient treatment of ischemic retina, and for potentially minimizing collateral destruction of adjacent viable perfused retina .
Widefield angiography is also used in patients with uveitis. The diagnosis and management of uveitis is challenging. Accurate diagnosis, definitions of activity and response to treatment are typically defined by clinical and angiographic appearance . Some features of posterior uveitis, such as perivascular sheathing, peripheral capillary non-perfusion, venous staining or leakage, cystoid macular edema, and disc edema, could be detected by widefield angiography . This technique allows for detecting images of both central and peripheral retina simultaneously giving more precise details then montage reconstruction.
Intraocular choroidal tumors are better visualized using wide field angiography. Differential diagnosis is supported by using multimodal imaging approach. Abnormal choroidal vessel or intrinsic vessels as in hemangioma or in melanoma are better seen in widefield angiography allowing for a more precise diagnosis.
Imaging the peripheral retina has significantly improved over the past years. Widefield angiographic technology has become an important clinical tool with regards to early diagnosis, treatment and monitoring of most sight-threatening retinal and choroidal diseases.
2.4 Quantitative Autofluorescence of the Retina
Confocal scanning laser ophthalmoscopy has been the imaging system of choice for auto-fluorescence (AF) imaging because of its high sensitivity and its image averaging capabilities that are required to record the fundus AF with acceptable signal/noise ratios using safe retinal exposures. The first clinical AF imaging systems were introduced in the mid-1990s [42, 43] using an excitation wavelength of 488 nm. Subsequent developments and the introduction of several commercial imaging platforms have further broadened the field and allowed AF imaging to become an important imaging modality for clinical diagnosis .
2.4.1 Origin and Spectral Characteristics of Fundus Auto-Fluorescence (AF)
The fluorophore responsible for AF of the fundus is principally lipofuscin residing in the retinal pigment epithelium . Lipofuscin is a byproduct of the visual cycle. The tips of the outer segments of photoreceptor are damaged by photo-oxidation and are phagocytosed on a daily basis. These materials contain poly-saturated fatty acids and byproducts of the visual cycle that are partially digested in the RPE. A small fraction is chemically incompatible for degradation and accumulates in lysosomes of the RPE as lipofuscin, a mixture of various fluorophores. Chemically some of these compounds have been identified and synthetized as bisretinoids [46, 47].
2.4.2 Quantitative Auto-Fluorescence (AF) Imaging
Successful acquisition of images for quantifying fundus AF depends in large part on the skills of the operator, and a dedicated operator is highly recommended. Detailed protocols for acquiring optimal AF images have been published  including special protocols used in quantitative work . In essence, the fundus of the dilated test eye (>6 mm diameter) is first illuminated with 488 nm light for a 20–30 s long period to reduce AF attenuation by photo-pigment absorption (bleaching). Focus and camera alignment are optimized during that period. Critical uniformity AF signal over the whole field is attained by fine axial adjustment of the camera position. The sensitivity is adjusted to avoid non-linear effects for both the fundus and the internal reference (see later). After final alignment a ‘video’ of 9 or 12 frames is acquired. After rejection of low quality frames (eye-movement, iris obstruction), the remaining frames are aligned and averaged and saved in the “non-normalized” mode (no histogram stretching) to create the AF image for analysis.
The last term of the Eq. (2.9) accounts for the absorption of the excitation light and the fluorescence emission by the ocular media. We have used the algorithm of van de Kraats and van Norren  to estimate the average optical density of the ocular media at a given age. In order to eliminate some unknown terms, this density was expressed relative to the media density at age 20 years . Thus, qAF reflects fundus AF relative to that which would be measured through the media of a 20-year-old emmetropic eye with average ocular dimensions.
2.4.3 Research Studies
qAF methodology has been used for normative studies [58, 61], and in investigations of patients affected by Best vitelliform macular dystrophy , recessive Stargardt disease , Bull’s eye maculopathy , Retinitis Pigmentosa , and Age related macular degeneration . Additionally, studies of subjects with a monoallelic ABCA4 mutation were also reported [61, 65]. The method appears robust when the image quality is good particularly in terms of uniformity. In different studies, repeatability for two sessions on the same day varied from ±6% to 10% (95%CI) and concordance between eyes was ±13–20%. These studies have demonstrated that quantification of AF with this standardized approach can aid in assessing whether specific fundus areas in pathological conditions have normal or abnormal AF levels, in providing valuable genotype-phenotype correlations and in studying the natural history of disease progression in contrast to normal aging. However, long-term repeatability will have to be systematically investigated before longitudinal studies are undertaken.
2.5 Summary and Conclusion
The scanning laser ophthalmoscope has, since its invention in 1980, undergone numerous improvements and has evolved in a sophisticated and versatile imaging modality. Today, SLO imaging is a very established tool in clinical routine. Two examples, the scanning laser tomography for glaucoma diagnostics (Sect. 2.2), as well as the wide field SLO angiography and its importance for the assessment of various retinal and choroidal diseases (Sect. 2.3), have been discussed in detail. But also other SLO applications as multicolor imaging  and autofluorescence imaging  are successfully used in clinical routine for the diagnoses and progression control of various diseases as e.g. age-related macular degeneration (AMD) and hereditary macular dystrophies.
Furthermore, in many devices SLO technology is used to provide a reference image in order to actively track eye movements and thus stabilize the area of interest for other examination technologies as optical coherence tomography (OCT) (Chap. 3), OCT angiography (Chap. 6), fluorescence lifetime imaging ophthalmoscopy (FLIO, Chap. 10) and microperimetry .
Finally, SLO autofluorescence imaging is a very active field of research, since it visualize the distribution of intrinsic fluorophores (mainly lipofuscin components), which play an important role in the visual cycle (renewal) of photoreceptors. Therefore, the quantification of the intrinsic fluorescence as described in Sect. 2.4 can provide a better understanding of the natural history of diseases and their pathological mechanisms.
- 1.Maiman T. Speech by Dr. Theodore H. Maiman, Hughes Aircraft Company. 1960. at a Press Conference at the Hotel Delmonico, New York.Google Scholar
- 8.Zinser G, et al. Confocal laser tomographic scanning of the eye. In: 33rd Annual Technical Symposium. Bellingham, WA: SPIE; 1989.Google Scholar
- 16.Niedrig H. Bergmann-Schaefer: Lehrbuch der Experimentalphysik. Optik. 1993;3:9.Google Scholar
- 17.McConnell G, Amos W, Wilson T. Confocal microscopy. In: Handbook of comprehensive biophysics. Amsterdam: Elsevier; 2011.Google Scholar
- 32.Flower RW, Hochheimer BF. A clinical technique and apparatus for simultaneous angiography of the separate retinal and choroidal circulations. Invest Ophthalmol Vis Sci. 1973;12(4):248–61.Google Scholar
- 42.von Rückmann A, Fitzke FW, Bird AC. Distribution of fundus autofluorescence with a scanning laser ophthalmoscope. Br J Ophthalmol. 1995;119:543–62.Google Scholar
- 44.Lois N, Forrester JV, editors. Fundus autofluorescence. 2nd ed. Philadelphia, PA: Wolters Kluewer; 2016.Google Scholar
- 51.Arend OA, et al. In-vivo fundus-fluoreszenz-messungen bei patienten mit alterabhangiger makulardegeneration. Ophthalmologie. 1995;92:647–53.Google Scholar
- 53.Schmitz-Valckenberg S, et al. How to obtain the optimal fundus autofluorescence image with the confocal scanning laser ophthalmoscope. In: Holz F, et al., editors. Atlas of fundus autofluorescence imaging. Berlin; New York, NY: Springer; 2007. p. 37–47.Google Scholar
- 54.Delori FC, Duncker T, Sparrow JR. Quantitative autofluorescence: the measurement of fundus autofluorescence levels. In: Lois N, Forrester JV, editors. Fundus autofluorescence. Philadelphia, PA: Wolters Kluewer; 2016.Google Scholar
- 56.Delori FC. et al. Supplementary material. Quantitative measurements of autofluorescence with the scanning laser ophthalmoscope; 2011. http://www.iovs.org/lookup/suppl/doi:10.1167/iovs.11-8319/-/DCSupplemental.Google Scholar
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.