Flicker-Induced Retinal Arteriole Dilation Is Reduced by Ambient Lighting

Abstract

Purpose.: To investigate the impact of ambient room lighting on the magnitude of flicker light–induced retinal vasodilations in healthy individuals.

Methods.: Twenty healthy nonsmokers participated in a balanced 2 × 2 crossover study. Retinal vascular imaging was performed with the dynamic vessel analyzer under reduced or normal ambient lighting, then again after 20 minutes under the alternate condition. Baseline calibers of selected arteriole and venule segments were recorded in measurement units. Maximum percentage dilations from baseline during 20 seconds of luminance flicker were calculated from the mean of three measurement cycles. Within-subject differences were assessed by repeated measures analysis of variance with the assumption of no carryover effects and pairwise comparisons from the fitted model.

Results.: Mean (SD) maximum arteriole dilations during flicker stimulation under reduced and normal ambient lighting were 4.8% (2.3%) and 4.1% (1.9%), respectively (P = 0.019). Maximum arteriole dilations were (mean ± 95% confidence interval) 0.7% ± 0.6% lower under normal ambient lighting compared with reduced lighting. Ambient lighting had no significant effect on maximum venular dilations during flicker stimulation or on the baseline calibers of arterioles or venules.

Conclusions.: Retinal arteriole dilation in response to luminance flicker stimulation is reduced under higher ambient lighting conditions. Reduced responses with higher ambient lighting may reflect reduced contrast between the ON and OFF flicker phases. Although it may not always be feasible to conduct studies under reduced lighting conditions, ambient lighting levels should be consistent to ensure that comparisons are valid.

Introduction

Retinal blood vessels dilate when the retina is stimulated with flickering light. ¹ This phenomenon is thought to reflect an increase in blood flow to meet the increased metabolic needs of activated neurons. The precise mechanism is not completely understood, but is thought to involve increased ganglion cell activity ² and local nitric oxide (NO) production. ^3–5  

Repetitive modulation of the luminance or chromatic characteristics of a light source produces an increase in retinal blood flow. Luminance modulation is relatively selective for the magnocellular visual pathway, derived from parasol ganglion cells, whereas chromatic modulation is selective for the parvocellular pathway and midget ganglion cells. ⁶ Parasol ganglion cells receive mixed input from L- (red) and M-cones (green) in their receptive fields and carry achromatic visual information. In contrast, midget cells receive L- and M-cone antagonistic input and mediate red-green color opponency. ^7–9 Maximal luminance or chromatic flicker-induced retinal hyperemia occurs when the flicker phases are directed at opposing ganglion cell populations. ¹⁰  

Flicker light-induced retinal vasodilation has received particular attention as a marker of retinal function in diabetes since the development in vivo imaging techniques. For example, studies have consistently found that people with diabetes have significantly reduced responses compared with their nondiabetic counterparts, ^11–13 even in the absence of diabetic retinopathy. ^14,15 Measurement of retinal vasodilation during flicker stimulation may therefore represent a sensitive biomarker of retinal damage in diabetes. 

Measurement of retinal vasodilation during flicker stimulation is generally conducted under reduced ambient lighting conditions. ^1,12,13 However, it may not always be feasible to examine people in the dark, such as when examinations must be conducted in a room with limited ability to reduce the ambient lighting or when other procedures or monitoring dictate that good lighting is required. It is unclear whether ambient lighting could affect the magnitude of retinal vasodilations during flicker stimulation. If true, direct comparisons between tests conducted under different ambient lighting conditions would not be valid. 

The purpose of this study was to investigate whether flicker light–induced retinal arteriole or venule dilation is affected by ambient lighting. We hypothesized that tests conducted under reduced ambient lighting would yield the same results as those conducted under normal room lighting. In addition, we hypothesized that baseline retinal arteriole and venule calibers would be unchanged between reduced and normal ambient lighting conditions. 

Methods

Participants

We studied the right eyes of 20 healthy nonsmokers aged at least 18 years recruited from the Centre for Eye Research Australia (Melbourne, Australia). Exclusion criteria were any self-reported systemic condition, epilepsy, significant eye pathology other than a need for glasses or contact lenses, and tropicamide sensitivity. Participants were not excluded on the basis of myopia. However, one participant with a refractive error (spherical equivalent) of −6.75 was excluded because we could not identify an arteriole segment near the optic disc greater than 90 measurement units (MU) wide. This is the limit of accuracy for dilation measurements with the dynamic vessel analyzer (DVA; IMEDOS Systems UGI, Jena, Germany). ¹⁶ All participants underwent measurement of their best-corrected visual acuity (BCVA), refractive error (spherical equivalent), intraocular pressure (IOP), blood pressure, heart rate, and body mass index. 

The study protocol followed the tenets of the Declarations of Helsinki and received institutional review board approval (12/1094H). Informed consent was obtained from all participants after explanation of the nature and possible consequences of the study. 

IOP and Hemodynamic Measurements

The IOP was measured in the right eye after topical instillation of oxybuprocaine hydrochloride, 0.4%, and fluorescein sodium with a slit lamp–mounted Goldmann applanation tonometer (Haag-Streit, Bern, Switzerland). The blood pressure and heart rate were recorded after 5 minutes in a seated position with an automatic upper arm sphygmomanometer (HEM-7000-C1L; Omron Healthcare, Lake Forest, IL, USA). 

Crossover Study

Flicker light–induced retinal vasodilation was measured in two periods with at least 20 minutes between consecutive periods. We previously found that sensitivity of the retinal arteriole response to flicker light stimulation recovers within 5 to 30 minutes, probably as a result of adaptation in cones after withdrawal of the light source. ¹ We chose a rest period of 20 minutes for the present study, given that dark adaptation in cones after exposure to bright light is complete within 5 to 10 minutes. ¹⁷ Tests were conducted under reduced ambient lighting and normal ambient lighting. Normal ambient lighting was with three sets of two ceiling-mounted, 1.2 m-long, 36-W fluorescent lights positioned approximately 2 m above the equipment. Reduced ambient lighting was with these lights turned off. Ten participants were tested under reduced followed by normal ambient lighting (sequence 1), while 10 participants were tested under normal followed by reduced ambient lighting (sequence 2). 

Flicker Light–Induced Retinal Vasodilation

We measured flicker light–induced retinal vasodilation in the right eye after dilation with topical tropicamide 1% with the DVA (IMEDOS Systems UGI) as described previously. ¹ Briefly, participants were seated and instructed to fixate on the tip of a fixation bar inside the camera while the fundus was examined under green light (530–600 nm) with the mydriatic camera (FF 450^plus; Carl Zeiss AG, Jena, Germany) set to a viewing angle of 30°. The light source was set to an average luminance of 130 cd/m² for both ambient lighting conditions, measured with a research radiometer (ILT1700; International Light Technologies, Peabody, MA, USA). A unique high-contrast region (e.g., a vessel branch) was selected as a fixation target to allow the DVA software (IMEDOS Systems UGI) to compensate for eye movements and blinking. Next, we selected for analysis a straight segment of a temporal arteriole and venule located between 0.5 to 2 disc diameters from the optic disc margin and at least one vessel diameter from any bifurcation or neighboring vessel. Superior vessels were chosen where possible to reduce interference from the upper eyelid. An example of vessel selection is provided in Figure 1. 

Figure 1

 

An example of arteriole and venule segment selection.

Vessel diameters were automatically and continuously measured in real-time for 350 seconds. This consisted of 50 seconds of constant light, followed by three cycles of 20 seconds of diffuse luminance flicker at 12.5 Hz and 80 seconds of constant light. Repetition mode was used for the second test to ensure that the same vessel segments were studied. If repetition mode was unable to automatically reidentify the vessel segments, the same segments were manually reselected using an image of the previous examination location. The distance of the camera from the participant was adjusted during the test to maintain the software brightness indicator between 30% and 40% on an arbitrary scale from 0% to 100%. Baseline vessel diameters were reported in MU, where 1 MU is equivalent to 1 μm for the Gullstrand eye. ¹⁶ Maximum vessel dilation was calculated from the mean of the three measurement cycles as the maximum percentage increase in vessel diameter relative to baseline during 20 seconds of flicker stimulation. In addition, the time of maximum dilation and the area under the curve (AUC) during flicker stimulation was measured for each participant. 

Retinal illumination was estimated with a digital light meter (EasyView Digital Light Meter; EA30, Extech Instruments, Waltham, MA, USA) placed at the location of the eye. Illumination was 170 lux and 0.1 lux during the ON and OFF flicker phases under reduced ambient lighting, respectively. Illumination under normal ambient lighting was 190 and 30 lux during the ON and OFF flicker phases, respectively. This corresponded to a modulation depth of 1.00 and 0.73 under reduced and normal ambient lighting, respectively. The modulation depth was calculated by the Michelson formula as (I ₁ – I ₂)/(I ₁ + I ₂), where I ₁ and I ₂ represent the illuminance during the ON and OFF flicker phases, respectively. ¹⁰  

Statistical Analyses

The maximum dilations during flicker stimulation and the baseline calibers of retinal arteriole and venule segments were compared by repeated measures ANOVA. Sequence (the order of lighting conditions) was the between-subject factor. Period and lighting were the within-subject factors. This method determined whether our results were impacted by: the order of lighting conditions; the chronological order of the tests; and ambient lighting. A technical limitation of the 2 × 2 crossover study is that carryover effects cannot be assessed with the other three factors, but our study was designed to minimize such effects. Pairwise comparisons of the dependent variables between the alternative lighting conditions were performed from the fitted ANOVA model. Data were analyzed in statistical software (STATA version 12.1; StataCorp LP, College Station, TX, USA). Values of P < 0.05 were considered significant. Based on previous data, ¹ our crossover study with 20 participants would have a power of 85% to detect a 0.5% difference in mean arteriole dilations at the 0.05 significance level, assuming a within-subject standard deviation of 0.5%. 

Results

The baseline characteristics of participants are described in Table 1. Participants were relatively young (age 32.8 [6.5] years); had healthy systolic (112.6 [13.0] mm Hg) and diastolic (75.0 [9.2] mm Hg) blood pressures; and had excellent vision (best corrected visual acuity [BCVA]: −0.1 [0.1] by LogMAR chart). Participants were generally emmetropic or mildly myopic (spherical equivalent −0.9 [1.8] diopters) and most were female (75%). 

Table 1

 

Baseline Characteristics of Participants

Parameter	Value (n = 20)
Age, y	32.8 (6.5)
Sex, male/female	5/15
Systolic blood pressure, mm Hg	112.6 (13.0)
Diastolic blood pressure, mm Hg	75.0 (9.2)
Heart rate, beats/min	68.6 (12.4)
Body mass index, kg/m2	25.1 (4.5)
Intraocular pressure, mm Hg	13.9 (2.1)
BCVA, logMAR chart	−0.1 (0.1)
Refractive error, spherical equivalent	−0.9 (1.8)

 Data are expressed as mean (SD) unless otherwise indicated.

The mean arteriole and venule responses over time under reduced and normal ambient lighting are expressed graphically in Figure 2. Maximum dilations and their timings, the AUC of vessel responses and baseline calibers under reduced and normal ambient lighting conditions are summarized in Table 2. Mean (SD) maximum arteriole dilations under reduced and normal ambient lighting were 4.8% (2.3%) and 4.1% (1.9%), respectively (P = 0.019). The corresponding values for venule dilations were 5.4% (1.8%) and 5.1% (2.1%), respectively (P = 0.397). Maximum arteriole dilations were (mean ± 95% confidence interval [CI]) 0.7% ± 0.6% lower under normal ambient lighting compared with reduced lighting. In addition, the AUC of arteriole responses were reduced by 8.4 ± 7.1% × s. Ambient lighting had no significant effect on the AUC of the venule response or the timing of maximum vessel dilations during flicker stimulation. The maximum relative dilations of retinal arterioles and venules during flicker stimulation under the two different ambient lighting conditions are expressed graphically in Figure 3. No significant sequence or period effects were identified for maximum arteriole or venule dilations. 

Figure 2

 

Mean relative dilations over time under reduced and normal ambient lighting. (A) Arterioles. (B) Venules.

Figure 3

 

Maximum relative retinal arteriole and venule dilations during flicker stimulation. (A) Responses under reduced and normal ambient lighting. (B) Differences in individual responses under normal compared with reduced ambient lighting. Data are mean (SEM). *P < 0.05.

Table 2

 

Dynamic Imaging Parameters Under Reduced and Normal Ambient Lighting

Parameter	Reduced Lighting	Normal Lighting	P
Maximum dilations (%)
Arterioles	4.8 (2.3)	4.1 (1.9)	0.019
Venules	5.4 (1.8)	5.1 (2.1)	0.397
Time of maximum dilations (s)
Arterioles	15.5 (4.4)	13.5 (5.3)	0.179
Venules	16.4 (2.2)	16.9 (2.7)	0.553
AUC during flicker (% × s)
Arterioles	57.6 (30.5)	49.2 (28.3)	0.023
Venules	58.9 (23.2)	54.1 (21.2)	0.156
Baseline calibers (MU)
Arterioles	123.0 (16.0)	123.1 (15.7)	0.817
Venules	150.4 (14.6)	151.1 (13.9)	0.678

 Data are expressed as mean (SD).

The baseline calibers of selected retinal arteriole and venule segments are shown in Figure 4. Mean (SD) baseline arteriole calibers under reduced and normal ambient lighting were 123.0 (16.0) MU and 123.1 (15.7) MU, respectively. Baseline venule calibers under reduced and normal ambient lighting were 150.4 (14.6) MU and 151.1 (13.9) MU, respectively. Ambient lighting had no significant effect on baseline arteriole or venule calibers. Further, no significant sequence or period effects were identified for baseline arteriole or venule calibers. 

Figure 4

 

Baseline retinal arteriole and venule calibers. (A) Calibers under reduced and normal ambient lighting. (B) Differences in individual calibers under normal compared with reduced ambient lighting. Data are mean (SEM).

Discussion

The results of our study indicate that retinal arteriole dilations during luminance flicker stimulation are smaller when tests are conducted under higher ambient lighting conditions, in contrast to our original hypothesis. Any effect of ambient lighting on venule dilations appears to be negligible. As predicted, we found no suggestion that ambient lighting affects baseline retinal arteriole or venule calibers. This result is important, as it shows that comparisons of flicker light–induced retinal vasodilations are only valid when tests are conducted under the same ambient lighting conditions. 

Retinal vasodilation during flicker stimulation is an indirect marker of neuronal activity-dependent hyperemia. However, it is currently more reliable than direct measures of hyperemia, such as with laser Doppler velocimetry, which are unable to take real-time measurements during flicker stimulation and produce more variable results. ^1,15 According to the Hagen-Poiseuille equation, laminar flow through a cylinder is proportional to the fourth power of the radius. A small increase in the radius of a blood vessel therefore leads to a relatively large increase in blood flow. For these reasons, vasodilation is generally considered to be an acceptable surrogate marker of hyperemia. 

It is curious that reduced arteriole dilations during flicker stimulation under normal ambient lighting were not associated with a similar reduction in venule dilations, given that blood flow through the arterial and venous circulations should match. We previously observed reduced arteriole dilations during flicker stimulation on short-term retesting, probably due to light adaptation over the course of the test. ¹ As with the present study, we did not observe a significant effect on venule dilations. In addition, we found that while arteriole dilations had excellent reproducibility, venule dilations were less consistent. Blood flow is dependent on vessel caliber and blood velocity. It is likely that arteriole flow is strongly related to arteriole caliber, whereas blood velocity, which is not measured by the DVA, may be a more important determinant of venule flow. 

Retinal ganglion cells are thought to be the main neuronal mediators of hyperemia during flicker stimulation. ² The two main ganglion cells in the primate retina are parasol cells, which give rise to the magnocellular pathway, and midget cells of the parvocellular pathway. ¹⁸ These cells are activated by flickering light to different degrees according to characteristics such as the flicker frequency, modulation depth, luminance contrast, and chromatic contrast. ^19,20 Parasol cells are most sensitive to luminance contrast modulation at frequencies between 10 and 20 Hz, whereas midget cells are most sensitive to chromatic contrast modulation at frequencies below 5 Hz. ⁶ Our flicker stimulus had a constant wavelength of 530 to 600 nm at a frequency of 12.5 Hz. Thus, our stimulus was most selective for parasol cells of the magnocellular pathway. 

With some exceptions, ganglion cells have a central receptive field with a concentrically arranged antagonistic surround. ²¹ For luminance signals, both parasol and midget cells can be either ON center/OFF surround or OFF center/ON surround. Parasol cells receive mixed input from L- and M-cones in both areas. Midget cells receive overlapping red-green opponent signals and are either L-ON/M-OFF or L-OFF/M-ON. ON center midget cells receive a dominant ON signal from L- or M-cones in the center, whereas OFF center midget cells receive a dominant OFF signal from L- or M-cones in the center. ⁹ Signals from S-cones are carried by small bistratified ganglion cells and do not appear to involve the magnocellular or parvocellular pathways. ²² Luminance flicker with green light at a frequency of 12.5 Hz, as in our study, might therefore increase net ganglion cell activity by stimulating both ON and OFF center parasol cells. Midget cells might also contribute to a small increase in ganglion cell activity despite being less sensitive to our stimulus. 

An important determinant of retinal hyperemia during luminance flicker stimulation is the modulation depth. ¹⁰ That is, a greater change in luminance between the ON and OFF flicker phases produces a greater increase in retinal blood flow. Lower flicker light–induced arteriole dilations under higher ambient lighting could therefore be explained in terms of a reduction in the modulation depth. With higher ambient lighting, retinal luminance would remain higher during the OFF flicker phase and the stimulation of OFF center ganglion cells would be reduced. This probably accounts for why arteriole dilations were lower under normal ambient lighting in our study. 

Data from humans and cats indicate that NO is important for retinal vasodilations during flicker stimulation. ^3–5 Nitric oxide is synthesized by three isoforms of nitric oxide synthase (NOS): neuronal (nNOS), endothelial, and inducible. Mouse data indicate that nNOS is found in retinal bipolar, amacrine, and ganglion cells. ^23–25 Further, recent evidence indicates that nNOS is the dominant isoform in the retina and that light-induced NO production occurs primarily in the two plexiform layers away from most blood vessels. ²⁶ Endothelial NOS is approximately half as prevalent as neuronal NOS and is primarily found in blood vessels. ^26,27 Inducible NOS is not present in significant quantities under normal conditions. ²⁶ Luminance flicker might increase nNOS activity in ganglion cells, with local diffusion of NO to neighboring inner retinal blood vessels, vasodilation and increased blood flow. A smaller change in retinal luminance between flicker phases could therefore produce smaller arteriole dilations via a reduction in NO release from parasol ganglion cells. 

The main strengths of this study were its simple crossover design and clear distinction between the two ambient lighting conditions under investigation. In addition, all measurements were taken by a single experienced investigator (JEN). The reliability of our measurements meant that we had good statistical power to detect small differences in our main outcomes with a relatively small number of participants. We recognize that our study was limited in its ability to explain a reduction in arteriole dilation during flicker stimulation with higher ambient lighting. Although this may involve reduced ganglion cell activity, our study was unable to provide direct evidence to support this hypothesis. Another limitation was our overrepresentation of female participants, even though we have not previously observed an effect of sex on DVA measurements. ¹  

Our study demonstrates that ambient lighting is an important determinant of arteriole dilation during luminance flicker stimulation. Studies of this response should therefore ensure that comparisons are only made between tests conducted under the same ambient lighting conditions. It may not always be feasible to conduct tests in the dark, but if higher ambient lighting is unavoidable, it is important to maintain consistent lighting throughout each study. 

Acknowledgments

Supported with funding from the Royal Victorian Eye and Ear Hospital; the Australian National Health and Medical Research Council (JEN) Postgraduate Medical Scholarship (ID1038701); and Australian NHMRC Research Fellowships (ID1045280 and 1003113; ELL, GJD). The Centre for Eye Research Australia receives Operational Infrastructure Support from the Victorian Government. 

Disclosure: J.E. Noonan, None; G.J. Dusting, None; T.T. Nguyen, None; R.E.K. Man, None; W.J. Best, None; E.L. Lamoureux, None