Open Access
Add to BibSonomy
Abstract
Optoretinography has enabled noninvasive visualization of physiological changes in cone photoreceptors exposed to light. Understanding the cone optoretinogram in healthy subjects is essential for establishing it as a biomarker for cone function in disease. Here, we measure the population cone intensity optoretinogram in healthy adults, for multiple irradiance/duration combinations of visible stimuli with equal energy. We study the within and between session repeatability and reciprocity of the ORG in five healthy subjects. We find the cone optoretinogram exhibits equivalent amplitudes for equal-energy stimuli. We also find good within-subject repeatability, which allows us to show differences across the five subjects.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
The accessibility of the retina for imaging through the natural optics of the eye enables noninvasive measurements of photoreceptor structure and function. Clinical techniques such as perimetry measure perception at different locations in the visual field, although as conventionally applied this technique is limited due to the scale of the visible stimulus on the retina [1,2], the accuracy with which stimuli can by repeatedly delivered to the same retinal location, and the testing time required. The implementation of adaptive optics (AO) can improve the resolution of perimetry tests by correcting the eye’s optical aberrations and thereby confining visual stimuli to sizes commensurate with single photoreceptors [3–6]. This technique also provides for targeted delivery of stimuli to precisely-defined retinal locations. AO-guided microperimetry provides a gold-standard for assessing visual function with high resolution, as it enables measurement of visual perception at the cellular scale of individual photoreceptors [7–10]. As with conventional microperimetry, however, the testing time required per subject is high.
To quantitatively assess visual function in the retina without the need for perceptual judgments, electroretinography (ERG) [1], including focal [11,12] and multifocal ERG [13,14] has been widely used. ERG provides an objective assessment of photoreceptor and inner retinal neuronal function by measuring the electrical signals that propagate through the retina in response to photoisomerization [15]. This technique, however, requires placement of an electrode on the subject’s cornea and is thus mildly invasive.
Recently, noninvasive imaging techniques incorporating AO have been able to measure photoreceptor function by quantifying the change in cells’ optical signal after exposure to visible light. Techniques such as AO flood illumination ophthalmoscopy [16,17] and scanning laser ophthalmoscopy (SLO) [18–20] quantify changes in the optical cone intensity signal of en-face images of cone photoreceptors. AO optical coherence tomography (AO-OCT) [21,22] measures changes in the difference of the phase of light reflected from the cone inner segment/outer segment (IS/OS) and cone outer segment tip (COST), and thus changes in the optical path length (OPL) of the outer segment. Changes in reflected light or OPL after exposure to visible light occur as a result of phototransduction, and measurement of these changes is termed an optoretinogram (ORG). Multiple groups have confirmed the link between the ORG and phototransduction [18,21,23], and these changes are linked with changes in the cell’s membrane induced phototransduction [24,25].
Recently, the ORG has been used to measure functional loss in retinitis pigmentosa [25], thus showing potential as a clinically useful biomarker for retinal disease. This motivates further characterization of the ORG, to establish its properties in healthy subjects so as to allow it to be more widely used as a biomarker for disease.
Here we characterize the reliability of the cone intensity ORG, measured using AOSLO. We characterize repeatability of the ORG within and across sessions and describe between-subject variation, relative to the repeatability findings. In addition, we investigate the peak amplitude of the cone intensity ORG for a population of cone photoreceptors under multiple combinations of stimulus irradiance and exposure duration that yield equal energy. These measurements help establish optimal stimulation conditions for further study. The data presented here provide key information required for further developing the population cone ORG as a biomarker of cone function capable of evaluating cone dysfunction in retinal disease.
2. Methods
2.1 Pre-registration
The current study was pre-registered at the Open Science Foundation (https://osf.io/xufdh/) prior to initiation of data collection. Changes in both the affine registration to align more frames and the statistical analysis were made from the pre-registered experimental plan. The individual cone ORG analysis described in the pre-registration document have been reserved for a possible future report.
2.2 Subjects
This research adhered to the tenets of the Declaration of Helsinki and was approved by the Institutional Review Board at the University of Pennsylvania. Five subjects with no retinal pathologies (ages 28-41 years) participated in the study. In an independent experimental session prior to the main experiment, each subject’s best-corrected visual acuity was measured to confirm each subject had at least 20/20 best corrected visual acuity or better, and AOSLO images of the cone mosaic surrounding the fovea and along the horizontal and vertical meridians were acquired and montaged (post-imaging) to ascertain whether AO imaging was feasible that in subject (e.g., verifying steady fixation and stable wavefront correction).
2.3 AOSLO system
The optical design of the AOSLO system has been previously described [6,26]. The AOSLO imaging channel used a 785 nm laser diode (Thorlabs LP785-SF20) with a 2 nm full-width half max (FWHM) bandwidth, yielding a 95 µm coherence length and 90 µW power at the cornea, so that the retinal irradiance was 90 µW/deg2 given the 1 deg2 imaging field. The beam was scanned at a frequency of 15.1 kHz horizontally, and 17.85 Hz vertically over a 1°x1° imaging field on the retina and confocal and split-detection images [26] were captured simultaneously at a speed of 17.85 frames per second. The 545 nm visible stimulus was delivered using a super-continuum laser (NKT Photonics, SuperK EXTREME/VARIA, 10 nm FWHM bandwidth). A mechanical shutter (Thorlabs, SHB05T) was stationed in front of the exit aperture of the fiber responsible for the visible stimulus, for control of the stimulus exposure duration.
2.4 AO optoretinography imaging protocol
At the start of each imaging session, the subject’s preferred eye was dilated with one drop of 2.5% phenylephrine hydrochloride ophthalmic solution and one drop of 1% tropicamide ophthalmic solution. Preliminary imaging was used to locate the subject’s fovea, and the raster scanned image was positioned 1.5° in the temporal retina.
Before each optoretinography trial, subjects were dark adapted for 2 minutes to allow for 75% photopigment regeneration in the cone photoreceptors. A trial included 10 sequential image acquisitions, each with the same stimulus condition. Each acquisition consisted of one second of pre-stimulus recording, along with recording throughout and for four seconds after the stimulus exposure. Stimulus irradiance and duration were varied such that all stimuli provided an energy of 153 nJ/deg2 on the retina. The following retinal irradiance/duration combinations were used:
- • 0 nW/deg2 * 1 s = 0 nJ/deg2 (control).
- • 153 nW/deg2 * 1 s = 153 nJ/deg2.
- • 306 nW/deg2 * 0.5 s = 153 nJ/deg2.
- • 917 nW/deg2 * 0.167 s = 153 nJ/deg2.
- • 2.75 µW/deg2 * 0.056 s = 153 nJ/deg2.
Each type of trial was measured multiple times within an experimental session. We use the term permutation to refer to a set of trials for each stimulus condition. The order of trials within a permutation was randomized by stimulus condition. The entire experimental session consisted of five permutations; out of the five permutations in each session, only the first and last contained the control condition. Each permutation contained one trial for each stimulus condition, with the exception of the 2.75 µW/deg2 * 0.056 s condition, where two trials were collected in each permutation to enable measurement of within session variation. Thus, the five permutations in each session corresponded to 270 video acquisitions [(permutations 1 and 5 (10 acquisitions * 6 trials * 2 permutations)) + (permutations 2-4 (10 acquisitions * 5 trials * 3 permutations))]. A session lasted approximately 2.5 hours.
To measure the between-session repeatability of the cone ORG, full imaging sessions for each subject were repeated on two different days separated by at least one week. Each subject’s second imaging session was scheduled at approximately the same time of day and imaged at the same retinal location as the first session.
2.5 Image acquisition and processing
Images were processed via strip-alignment registration [27]. To correct for the static intra-frame distortion caused by the resonant scanner, the spatial distortion from images of a stationary Ronchi ruling was estimated and each image was resampled over a grid of equally spaced pixels. To maximize registration of frames during and immediately surrounding the presentation of the visual stimulus, a frame from within one-second of the start of stimulus presentation (frames 19-36) in the split-detection image sequence was used as the reference frame, and frames within the acquisition were aligned to it. Split-detection images were used as the primary sequence for image registration due to the stability of photoreceptor intensity observed throughout the acquired image sequence despite the application of a visual stimulus. The registration transforms calculated using the split-detection image sequence were then applied to the simultaneously acquired confocal images. To optimize image registration for analysis, multiple high-quality split-detection reference frames were selected, and the registration with the most frames registered was selected for analysis. If fewer than 60% of the images within the one-second stimulus duration, or if fewer than 60% of the images within an acquisition did not register to any reference image, the image acquisition was removed from analysis. Once images were registered, an affine registration was performed to remove residual torsion and the registered videos were cropped to a session-specific common area, generated for that session. All acquisitions in each session were then aligned to the respective session’s common area. Cones were identified semi-automatically within the session’s common area using a deep neural network [28] and cone locations were then matched across sessions via constellation matching [29] (Fig. 1). Cones that were identified and matched across both sessions were used for analysis, with the relative location between the two sessions known. A binary mask was created from a motion contrast algorithm [30] and applied to each aligned acquisition to highlight vasculature features. Cones underlying vasculature features were excluded from analysis.
Fig. 1. Matching cones across sessions and obtaining the population cone intensity response in Subject 11002. (A) A confocal image averaged from all acquisitions for session 1 (left) and session 2 (right) for one subject. (B) Feature matches across sessions were found using the constellation matching method [29]. The endpoints of each line indicate matched locations. The alignment procedure allows identified cones to be matched across sessions 1 (C) and 2 (D). Colored dots indicate individual cone locations that have been matched prior to identification of vasculature (2,364 matched cones). (E, F) Individual cone intensity traces from a single acquisition in one subject for session 1 (E) and session 2 (F) for the same stimulus condition (2.75 µW/deg2 * 0.056 s). Each cone’s signal was standardized to the cone’s individual pre-stimulus interval. Red, orange, and green traces correspond to the intensity traces of the individual cones labeled with red, orange and green dots in (C) and (D). 2,227 cone intensity traces are shown in (E), 2,126 for (F). The gray bar indicates the stimulus exposure duration. (G, H) The intensity response calculated from the dispersion of the standardized intensities for the 2,227/2,216 cones from the same acquisitions shown in (E)/(F). These population intensity responses were obtained by taking the standard deviation across all cones’ standardized intensity signals at each time point in (E) and (F) (G: session 1, H: session 2).
2.6 Obtaining the cone intensity ORG
The intensity ORG of cone photoreceptors was measured using methodology established previously in our lab [18], and we summarize here. In this paper, we study the population ORG obtained by aggregating intensity signals from individual cones over a retinal region containing at least 1,200 cones in each subject, with the actual number varying from subject to subject.
To obtain the raw intensity versus time signal for each cone in each acquisition, a 3 × 3 pixel column was projected within each cone throughout the aligned acquisition and then averaged over the 9 pixels, separately within each frame, to obtain the cone’s intensity frame-by-frame. The individual cone’s intensity at a given frame was then divided by the mean intensity of all cones for the same frame. This step accounted for non-specific effects on intensity caused by factors such as fluctuation in tear film thickness and the quality of the AO optical correction. This normalized individual cone signal was then standardized by the mean and standard deviation of that cone’s normalized signal during the pre-stimulus interval. This step accounted for pre-stimulus differences in intensity across cones. Standardized intensity responses are illustrated in Fig. 1(E, F).
The response to visible light in the individual cone standardized intensity is highly heterogeneous (Fig. 1(E, F)): some cone intensities increase, some decrease, and some oscillate [18]. Thus, to measure the population cone intensity response to visible light, we use a measure of dispersion taken across the standardized individual cone intensity responses. More specifically, we take the standard deviation of the standardized individual cones intensities in each frame and define this dispersion as the population cone intensity response. As seen in Fig. 1(G, H), this intensity response increases in response to a visible light stimulus.
To account for the fact that the heterogeneity of the standardized intensity of unstimulated cones increases with time because cone intensities naturally diverge from their standardized initial values [31], we also obtained the intensity response for the control acquisitions (Fig. 2). To isolate the visible stimulus effect, we then subtract the control intensity response from the stimulated intensity response to obtain the population cone intensity ORG, using the method illustrated in Fig. 2. This method also combines the data across the individual acquisitions for each stimulus condition. More specifically, for each frame, the square root of the mean of the squared cone intensity responses in the control acquisitions was subtracted from the square root of the mean of the squared intensity responses in the stimulus acquisitions. This step was done separately for each stimulus condition.
Fig. 2. Obtaining the population cone intensity ORG. (A, B) The population cone intensity response of each acquisition in the control condition for Subject 11002. 20/20 control acquisitions were included in session 1, 19/20 control acquisitions were included for session 2. (C, D) The population cone intensity response of each acquisition in the 2.75 µW/deg2 * 0.056s condition. 82/100 acquisitions were included for session 1; 93/100 acquisitions were included for session 2. Each trace in (A, B, C, D) represents a population intensity response measured from a single acquisition. (E, F) The population cone ORG (black line) for each session is obtained as described in the text; the population intensity responses from the control and stimulation conditions were pooled and the pooled control intensity response (blue line) was subtracted from the pooled stimulus intensity response (red line) to yield the population cone ORG (black line). We take the peak of a smooth spline fit to the population cone ORG to be the ORG amplitude. Gray bar indicates stimulus exposure time and duration for this stimulus condition.
As a summary measure, the peak amplitude of the population cone intensity ORG was determined by fitting a smooth spline of no theoretical significance (MATLAB MathWorks, Natick, MA, USA, R2020b, function fit: “smoothing spline” method, smoothing parameter 0.9995) and extracting the peak of the fit.
2.7 Statistical tests
Statistical analysis was implemented using MATLAB (MathWorks, Natick, MA, USA, R2022a) with a significance level set at p < 0.05. A three-way ANOVA (stimulus; session; subject; function anovan) was used to compare differences in the peak amplitude of the ORG and determine what factors affected the ORG amplitude with subject as a random effect and stimulus and session as fixed effects. Main effects and two-way interactions were included in the model, but not the three-way interaction. To measure within session variation, the two 2.75 µW/deg2 * 0.056 s trials from each permutation were randomized into two groups, and a one-way repeated-measures ANOVA (function anova1) was used to analyze differences between peak amplitudes of the population cone intensity ORG.
3. Results
The population cone intensity ORG was measurable for all stimulus conditions in both sessions for all five subjects. Of the 270 acquisitions associated with each session’s data collection, an average of 212 acquisitions (range 163-260) were included in the final analysis. The number of cones included in the analysis varied across subjects according to how well the retinal region imaged in session one was co-located with the retinal region imaged in session two. Prior to identifying the vascular mask, 2,364 cones were matched for Subject 11002, 1,921 cones were matched for Subject 11057, 3,102 cones were matched for Subject 11108, 1,269 cones were matched for Subject 11112, and 3,186 cones were matched for Subject 11115. The number of cones included for analysis was sufficiently large for all acquisitions for all subjects, such that the number of cones included is not expected to affect the population ORG [18].
To characterize the within-session repeatability of the cone intensity ORG amplitude for an individual stimulus, the 10 trials of the 2.75 µW/deg2 * 0.056 s stimulus condition were randomly sorted into two groups of 5, such that there was only 1 trial per permutation in each group. Figure 3 illustrates the cone ORG traces and amplitudes for both 2.75 µW/deg2 * 0.056 s groups, for each subject. ORG amplitudes were similar in both groups for each session in all subjects (Fig. 3). A one-way ANOVA indicated that there was no significant difference in the ORG amplitudes between groups within session 1 [F (1,49) = 0.04, p = 0.841] nor within session 2 [F (1,49) = 0.06, p = 0.815].
Fig. 3. The population cone intensity ORG (colored lines) for the two randomized groups of 2.75 µW/deg2 * 0.056 s trials for each subject for session 1 (left) and session 2 (middle). Black dotted lines indicate the smooth spline fit to the data. Population peak ORG amplitude for each group is plotted as a bar graph (right).
Each subject’s ORG amplitude was similar across equal-energy stimulus conditions within a session (Fig. 4). The three-way ANOVA (subject, session, stimulus condition; subject as random effect, session and stimulus as fixed effects) indicated that there was no main effect of stimulus condition [F (3,39) = 1.21, p = 0.349] nor a significant two-way interaction between stimulus condition and session [F (3,39) = 0.14, p = 0.932] or stimulus condition and subject [F (12,39) = 0.63, p = 0.786].
Fig. 4. The population cone ORG across stimulus conditions for each subject in both sessions. Each line graph illustrates the ORG from the four stimulus conditions separated by subject and session. Each bar plot on the right shows the corresponding peak ORG amplitudes for each condition and session as well as the average ORG amplitude taken across the four stimulus conditions within each session. The 2.75 µW/deg2 * 0.056s condition is based on the data from both randomized trial groups. Error bars for the average ORG amplitudes on the right represent ± 2 * standard error of the mean.
Given the lack of stimulus condition effect, we averaged the data across all stimulus conditions (right bars in each session group in Fig. 4). These averaged amplitudes ranged from 2.24 to 4.55. This between-subject variation was significant as shown by a main effect of subject in the three-way ANOVA [F (4,39) = 15.48, p = 0.014], indicating that it will be important to more fully characterize subject-to-subject variation in a future study.
Examination of the averaged amplitudes across sessions indicates some session variation, with the session 1 amplitude larger than the session 2 amplitude for some subjects, and the session 2 amplitude larger than the session 1 amplitude for others. This variation in sign of session effect across subjects is expected if the sessions, which were separated by at least a week, represent independent measurements of each subject’s ORG. The magnitude of the session effect was small relative to the between-subjects’ variation: the absolute value of the difference between sessions ranged between 0.003 and 0.66 across subjects, noticeably smaller than the between-subjects range of 2.31. None-the-less, this small session effect was significant, as revealed by a significant subject by session interaction in the three-way ANOVA [F (4,39) = 8.40, p = 0.002]. There was no main effect of session [F (1,39) = 3.48, p = 0.136], presumably because the sign of the session effect varied across subjects.
Although the ORG amplitude did not vary with stimulus energy, differences in the time course can be observed in Fig. 4. The ORG converges at longer durations, consistent with the common amplitude result.
Figure 5 (A) replots the average ORG amplitude for each subject and session in a form that allows easier comparison of the subject and session effects. Similarly, Fig. 5 (B) plots the averaged 2.75 µW/deg2 * 0.056s condition across the two groups of trials for this stimulus condition, again for each subject and session. The difference in ORG amplitude between subjects is larger than the error bars associated with individual subjects’ measurements. In addition, the distance of each point to the positive diagonal in the plot is smaller than the spread of the data along the diagonal, indicating that session-to-session variation is smaller than between-subject variation. The fact that the same pattern is seen in the single stimulus data in (B) as in the overall data in (A) indicates that data from a single stimulus condition and session are sufficient to characterize a subject’s ORG with good precision.
Fig. 5. Mean ORG amplitude for each subject in each session. Black data points indicate an individual subject’s ORG amplitude averaged across (A) all stimulus conditions and (B) the 2.75 µW/deg2 * 0.056s condition in each session. Error bars show ± 2 * standard error of the mean. The variation in ORG amplitude across subjects is greater than the variation observed in the ORG amplitudes for an individual subject within each session. This illustrates that one session is sufficient to obtain a reliable ORG for an individual subject. Furthermore, (B) confirms that the use of one stimulus condition is also sufficient. The black dashed line represents the 1-to-1 line of equivalence; the fact that some subjects’ data falls to the left of the line while others fall to the right illustrates the subject by session interaction described in the main text.
4. Discussion
With the spatial and temporal resolution of the AOSLO, the population cone intensity ORG can be measured reliably in single subjects. We have shown that the amplitude of the population cone intensity ORG is repeatable across varying stimuli of equal energy (Figs. 4, 5(A)). We have also shown that one stimulus condition is sufficient to acquire reliable intensity ORG measurements (Fig. 5(B)). Indeed, although there is some session-to-session variation in the measured ORG amplitude, this is small relative to the between-subject variation, indicating that a single session is sufficient to provide a reliable within-subject measurement (Figs. 4, 5). The ability to make a reliable measurement of visual function within a single session establishes the potential of the cone intensity ORG as a noninvasive and objective biomarker for the study of photoreceptor function in clinical populations. Prior studies have also measured the cone ORG in a normative group of subjects [18,19,21,32]; our work now provides quantification of its within and between subject reliability.
In addition to measuring the reliability of the cone ORG, we co-varied stimulus irradiance and exposure duration to explore how these two parameters trade off in the response. We found a reciprocal relationship in the peak amplitude, so that stimuli producing equal energy elicited the same peak ORG (Fig. 4). We believe that this effect reflects efficient temporal integration of the effects of photopigment excitation on the processes within cones that mediate the intensity changes underlying the ORG. It is possible that the temporal integration could change in response to disease or other manipulations, raising the possibility that changes in reciprocity could in the future be exploited to support more detailed inferences about the state of cone function. We do not explore this possibility here. We do note that other aspects of the ORG differed across the equal energy stimulus set. For instance, the rising slope of the cone intensity ORG increased with stimulus irradiances. As shown in Fig. 4, the 153 nW/deg2 and 306 nW/deg2 stimuli elicited smaller slopes, thus resulting in a longer time for the ORG amplitude to reach its peak level in all subjects. The change in slope is presumably related to changes in the rate of phototransduction with stimulus irradiance. Future studies may want to explore the rising slope of the ORG or other kinetic parameters, such as the time to peak or the ORG recovery time, as additional measures of cone function.
Our measurements showing peak amplitude reciprocity across a set of equal energy stimuli indicate that shorter duration stimuli may be used effectively in ORG studies. We have found qualitatively that shorter durations are more comfortable for subjects, with reduced fixational demands being a key factor. Shorter duration stimuli also help maintain tear film, which is critical for attaining high quality image sequences. Maximizing subject comfort is also important, especially as the acquisitions for a single stimulus condition can amount to 20 minutes of experimentation, with half of that time allocated to dark adaptation using our current protocol. To keep the session length reasonable, we only required the subject to dark adapt for 2 minutes between trials. This leads to recovery of ∼75% of the cone photopigment between trials. For full recovery, dark adaptation would need to extend to approximately 10 minutes between trials. Empirically, we have shown that 2 minutes of dark adaptation is sufficient for obtaining a reliable ORG. We have not established a minimum number of trials to determine a reliable ORG amplitude, though we have observed that one trial of 10 acquisitions produces a similar peak amplitude to the nominal 5 trial (50 acquisitions) protocol. Subject 11002, for example, showed an ORG amplitude of 4.75 for the first trial of the 2.75 µW/deg2 condition compared to an ORG amplitude of 4.67 for the full study. This would suggest that choosing the number of trials is a tradeoff between the amount of data collected over time, and the reliability of the data obtained. As shown, Fig. 2(C, D), reflects the variability of the ORG from individual acquisitions. Future studies evaluating the trade-offs between imaging and stimulus parameters, dark adaptation time, and subject comfort may yield optimal protocols for acquiring ORG measurements efficiently.
Our study does not address the mechanisms that underlie the cone intensity ORG, but some is known about this [16,18,25]. When photoreceptors undergo phototransduction due to light exposure, the cone outer segment (OS) is thought to deform and ultimately elongate [24]. This physiological response causes the optical path length between the reflective ends of the cone inner segment/outer segment junction (IS/OS) and the cone outer segment tip (COST) to change, as well as between other less reflective surfaces within the OS. These changes alter the interference between light reflected from these various surfaces [21,32–34]. Indeed, by using an imaging source with a coherence length longer than twice cone outer segment length (> 60 µm), the changes in interference are captured by the reflected light intensity. If the interference becomes more constructive, the intensity will increase; if it becomes more destructive, the intensity will decrease. Which of these effects is observed in a particular cone depends on the initial distances between the reflective surfaces within the outer segment. Variation in this initial state is thought to underlie the heterogeneity across cones of the intensity response. Here, we used an imaging source with a 95 µm coherence length to maximize the interference effect and hence the ORG [17,19]. Other techniques that acquire both en-face and depth-resolved images (e.g., AO-OCT) can further localize changes in the optical path length generated throughout the cone OS, providing additional information about the physiological mechanisms that undergo phototransduction [22].
Recently, several methods for measuring an ORG have been described [16,19,21,22,25,20]. Though most of these methods have utilized AO technology, it is possible to obtain an ORG without AO using OCT [35–39] to measure changes in the phase of the interference signal. To our knowledge, en face imaging techniques such as flood-illuminated fundus photography or SLO have not been able to measure an ORG without AO. As described here, the en face intensity ORG involves measuring the variability of cone intensities, with some cone intensities increasing while others are decreasing. Most non-AO en face imaging systems will not provide sufficiently high resolution to visualize individual cones. Therefore, in such systems, the blurring across cones would be expected to reduce or eliminate the ORG signal, since the variability that drives the ORG measurement would tend to be averaged out. Thus, AO provides an advantage for ORG measurements, even when examined over a population of cones.
Finally, variability in the ORG amplitude across subjects is a key finding in our study. Reasons contributing to such inter-subject variability are presently unknown. Recent studies have shown that outer segment length, dilated pupil size, and axial length can alter the overall light distributed onto the retina, contributing to inter-subject variability of the ORG [39]. Future studies investigating whether other biological variables such as age, sex, and race affect the ORG or whether the ORG changes with the amount of photoreceptor shedding over the course of the day [40] will need to be undertaken. Characterizing the normative ORG will be a critical component to establishing the ORG as a biomarker capable of detecting photoreceptor dysfunction and will enable translation of the technique to studies of retinal disease.
5. Conclusion
We have shown that the cone ORG is reliable within and across sessions. With reliability established, the current methodology now can be modified to optimize data acquisition, in order to benefit subject/experimenter experience and reduce the time allocated to the study. Further exploration of between-subject variation is needed. Future studies to characterize the cone ORG amongst a normative population of healthy subjects may provide insight into the range and causes of inter-subject variation and may ultimately lead to applications involving study of how retinal disease affects photoreceptor function.
Funding
National Institutes of Health (R01EY028601, R01EY030227, P30EY001583); Foundation Fighting Blindness; Research to Prevent Blindness; Center for Advanced Retinal and Ocular Therapeutics, Perelman School of Medicine, University of Pennsylvania; F. M. Kirby Foundation; Paul MacKall and Evanina Bell MacKall Trust.
Acknowledgments
We thank Alfredo Dubra for sharing the adaptive optics scanning laser ophthalmoscope optical design, as well as adaptive optics control, image acquisition and image registration software and we thank Robert F. Cooper for technical advice.
Disclosures
JIWM: P: US Patent 8226236 and US Patent App. 16/389,942; F: AGTC. DHB: P: US Patent App. 16/389,942. RLW declares no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. B. M. Hull and D. A. Thompson, “A review of the clinical applications of the pattern electroretinogram,” Oph. Phys. Optics 9(2), 143–152 (1989). [CrossRef]
2. K. Rohrschneider, M. Becker, N. Schumacher, T. Fendrich, and H. E. Volcker, “Normal values for fundus perimetry with the scanning laser ophthalmoscope,” Am. J. Ophthalmol. 126(1), 52–58 (1998). [CrossRef]
3. A. Roorda, F. Romero-Borja, W. Donnelly Iii, H. Queener, T. Hebert, and M. Campbell, “Adaptive optics scanning laser ophthalmoscopy,” Opt. Express 10(9), 405–412 (2002). [CrossRef]
4. S. A. Burns, A. E. Elsner, K. A. Sapoznik, R. L. Warner, and T. J. Gast, “Adaptive optics imaging of the human retina,” Prog. Retin. Eye Res. 68, 1 (2019). [CrossRef]
5. J. Liang, D. R. Williams, and D. T. Miller, “Supernormal vision and high-resolution retinal imaging through adaptive optics,” J. Opt. Soc. Am. A 14(11), 2884–2892 (1997). [CrossRef]
6. A. Dubra and Y. Sulai, “Reflective afocal broadband adaptive optics scanning ophthalmoscope,” Biomed. Opt. Express 2(6), 1757–1768 (2011). [CrossRef]
7. K. Rohrschneider, S. Bultmann, and C. Springer, “Use of fundus perimetry (microperimetry) to quantify macular sensitivity,” Prog. Retinal Eye Res. 27(5), 536–548 (2008). [CrossRef]
8. W. M. Harmening, W. S. Tuten, A. Roorda, and L. C. Sincich, “Mapping the perceptual grain of the human retina,” J Neurosci 34(16), 5667–5677 (2014). [CrossRef]
9. W. S. Tuten, G. K. Vergilio, G. J. Young, J. Bennett, A. M. Maguire, T. S. Aleman, D. H. Brainard, and J. I. W. Morgan, “Visual function at the atrophic border in choroideremia assessed with adaptive optics microperimetry,” Ophthalmol Retina 3(10), 888–899 (2019). [CrossRef]
10. W. S. Tuten, R. F. Cooper, P. Tiruveedhula, A. Dubra, A. Roorda, N. P. Cottaris, D. H. Brainard, and J. I. W. Morgan, “Spatial summation in the human fovea: Do normal optical aberrations and fixational eye movements have an effect?” J Vision 18(8), 6 (2018). [CrossRef]
11. G. E. Fish and D. G. Birch, “The focal electroretinogram in the clinical assessment of macular disease,” Ophthalmology 96(1), 109–114 (1989). [CrossRef]
12. A. Binns and T. H. Margrain, “Evaluation of retinal function using the dynamic focal cone ERG,” Oph. Phys. Optics 25(6), 492–500 (2005). [CrossRef]
13. D. C. Hood, “Assessing retinal function with the multifocal technique,” Prog. Retinal Eye Res. 19(5), 607–646 (2000). [CrossRef]
14. S. L. Ball and H. M. Petry, “Noninvasive assessment of retinal function in rats using multifocal electroretinography,” Invest. Ophth. Vis. Sci. 41(2), 610–617 (2000).
15. D. C. Hood and D. G. Birch, “Phototransduction in human cones measured using the alpha-wave of the ERG,” Vision Res. 35(20), 2801–2810 (1995). [CrossRef]
16. R. S. Jonnal, J. Rha, Y. Zhang, B. Cense, W. Gao, and D. T. Miller, “In vivo functional imaging of human cone photoreceptors,” Opt. Express 15(24), 16141–16160 (2007). [CrossRef]
17. P. Bedggood and A. Metha, “Variability in Bleach Kinetics and Amount of Photopigment between Individual Foveal Cones,” Invest. Ophthalmol. Visual Sci. 53(7), 3673–3681 (2012). [CrossRef]
18. R. F. Cooper, W. S. Tuten, A. Dubra, D. H. Brainard, and J. I. W. Morgan, “Non-invasive assessment of human cone photoreceptor function,” Biomed. Opt. Express 8(11), 5098–5112 (2017). [CrossRef]
19. R. F. Cooper, D. H. Brainard, and J. I. W. Morgan, “Optoretinography of individual human cone photoreceptors,” Opt. Express 28(26), 39326–39339 (2020). [CrossRef]
20. K. Grieve and A. Roorda, “Intrinsic signals from human cone photoreceptors,” Invest. Ophthalmol. Visual Sci. 49(2), 713–719 (2008). [CrossRef]
21. F. Zhang, K. Kurokawa, A. Lassoued, J. A. Crowell, and D. T. Miller, “Cone photoreceptor classification in the living human eye from photostimulation-induced phase dynamics,” Proc. Natl. Acad. Sci. U. S. A. 116(16), 7951–7956 (2019). [CrossRef]
22. M. Azimipour, D. Valente, K. V. Vienola, J. S. Werner, R. J. Zawadzki, and R. S. Jonnal, “Optoretinogram: optical measurement of human cone and rod photoreceptor responses to light,” Opt. Lett. 45(17), 4658–4661 (2020). [CrossRef]
23. G. Ma, T. Son, T. H. Kim, and X. Yao, “In vivo optoretinography of phototransduction activation and energy metabolism in retinal photoreceptors,” J. Biophotonics 14(5), e202000462 (2021). [CrossRef]
24. P. Zhang, R. J. Zawadzki, M. Goswami, P. T. Nguyen, V. Yarov-Yarovoy, M. E. Burns, and E. N. Pugh Jr., “In vivo optophysiology reveals that G-protein activation triggers osmotic swelling and increased light scattering of rod photoreceptors,” Proc. Natl. Acad. Sci. U. S. A. 114(14), E2937–E2946 (2017). [CrossRef]
25. A. Lassoued, F. Zhang, K. Kurokawa, Y. Liu, M. T. Bernucci, J. A. Crowell, and D. T. Miller, “Cone photoreceptor dysfunction in retinitis pigmentosa revealed by optoretinography,” Proc. Natl. Acad. Sci. U. S. A. 118(47), 1 (2021). [CrossRef]
26. D. Scoles, Y. N. Sulai, C. S. Langlo, G. A. Fishman, C. A. Curcio, J. Carroll, and A. Dubra, “In vivo imaging of human cone photoreceptor inner segments,” Invest Ophthalmol. Vis. Sci. 55(7), 4244–4251 (2014). [CrossRef]
27. A. Dubra, Z. Harvey, B. Fischer, B. Dawant, and C. Lorenz, “Registration of 2D images from fast scanning ophthalmic instruments,” Biomedical Image Registration 6204, 60–71 (2010).
28. D. Cunefare, L. Fang, R. F. Cooper, A. Dubra, J. Carroll, and S. Farsiu, “Open source software for automatic detection of cone photoreceptors in adaptive optics ophthalmoscopy using convolutional neural networks,” Sci. Rep. 7(1), 6620 (2017). [CrossRef]
29. M. Chen, R. F. Cooper, J. C. Gee, D. H. Brainard, and J. I. W. Morgan, “Automatic longitudinal montaging of adaptive optics retinal images using constellation matching,” Biomed. Opt. Express 10(12), 6476–6496 (2019). [CrossRef]
30. J. Tam, J. A. Martin, and A. Roorda, “Noninvasive visualization and analysis of parafoveal capillaries in humans,” Invest. Ophthalmol. Visual Sci. 51(3), 1691–1698 (2010). [CrossRef]
31. A. Pallikaris, D. R. Williams, and H. Hofer, “The reflectance of single cones in the living human eye,” Invest. Ophthalmol. Visual Sci. 44(10), 4580–4592 (2003). [CrossRef]
32. V. P. Pandiyan, A. Maloney-Bertelli, J. A. Kuchenbecker, K. C. Boyle, T. Ling, Z. C. Chen, B. H. Park, A. Roorda, D. Palanker, and R. Sabesan, “The optoretinogram reveals the primary steps of phototransduction in the living human eye,” Sci. Adv. 6(37), 1 (2020). [CrossRef]
33. R. S. Jonnal, “Toward a clinical optoretinogram: a review of noninvasive, optical tests of retinal neural function,” Ann. Transl. Med. 9(15), 1270 (2021). [CrossRef]
34. A. Meadway and L. C. Sincich, “Light reflectivity and interference in cone photoreceptors,” Biomed. Opt. Express 10(12), 6531–6554 (2019). [CrossRef]
35. G. Ma, T. Son, T. H. Kim, and X. Yao, “Functional optoretinography: concurrent OCT monitoring of intrinsic signal amplitude and phase dynamics in human photoreceptors,” Biomed. Opt. Express 12(5), 2661–2669 (2021). [CrossRef]
36. L. Zhang, R. Dong, R. J. Zawadzki, and P. Zhang, “Volumetric data analysis enabled spatially resolved optoretinogram to measure the functional signals in the living retina,” J. Biophotonics 15(3), e202100252 (2022). [CrossRef]
37. D. Hillmann, H. Spahr, C. Pfaffle, H. Sudkamp, G. Franke, and G. Huttmann, “In vivo optical imaging of physiological responses to photostimulation in human photoreceptors,” Proc. Natl. Acad. Sci. U. S. A. 113(46), 13138–13143 (2016). [CrossRef]
38. K. V. Vienola, D. Valente, R. J. Zawadzki, and R. S. Jonnal, “Velocity-based optoretinography for clinical applications,” Optica 9(10), 1100 (2022). [CrossRef]
39. X. Jiang, T. Liu, V. P. Pandiyan, E. Slezak, and R. Sabesan, “Coarse-scale Optoretinography (CoORG) with extended field-of-view for normative characterization,” Biomed. Opt. Express , 13, 11, 5989 (2022). [CrossRef]
40. O. P. Kocaoglu, Z. Liu, F. Zhang, K. Kurokawa, R. S. Jonnal, and D. T. Miller, “Photoreceptor disc shedding in the living human eye,” Biomed. Opt. Express 7(11), 4554–4568 (2016). [CrossRef]
