Retinal Arteriolar Dilation to Flicker Light Is Reduced on Short-Term Retesting

Abstract

Purpose.: To investigate the impact of retesting frequency over a short period on flicker light–induced retinal vasodilation.

Methods.: Twenty healthy participants were included. The retinal vascular response to flicker light stimulation was assessed three times (at baseline and after 5 and 30 minutes of rest [tests 1, 2, and 3, respectively]) in each participant using the Dynamic Vessel Analyzer. Relative dilations of selected arteriole and venule segments during flicker stimulation and resting diameters were measured automatically. The mean vessel dilations and resting diameters were compared using repeated-measures analysis of variance.

Results.: Participants were young (mean [SD] age, 33.1 [5.7] years) and mostly female (70%). The mean (SD) maximum arteriolar dilations during flicker stimulation were 3.23% (2.06%), 2.44% (1.62%), and 3.36% (2.11%) in tests 1, 2, and 3, respectively. The mean (SD) venular dilations were 4.26% (1.28%), 3.81% (1.61%), and 4.43% (1.73%) in tests 1, 2, and 3, respectively. The mean arteriolar dilations were significantly different across the three tests (P < 0.001). Compared with test 1, arteriolar dilations were significantly reduced after 5 minutes (P = 0.008) but not 30 minutes (P = 0.437) of rest. No significant differences were found over time for the mean venular dilations (P = 0.128). Resting diameters of selected vessels were not significantly different between tests.

Conclusions.: Retinal arteriolar dilation during flicker stimulation is reduced on short-term retesting, without a significant change in baseline vessel diameter, indicating decreased responsiveness to the flicker stimulus. Researchers should allow at least 30 minutes between consecutive tests to minimize suppression of the flicker response.

Introduction

Retinal arterioles and venules dilate when stimulated with flickering light, a phenomenon described as functional hyperemia or neurovascular coupling. ^1,2 This response reflects an increase in blood flow to meet the increased metabolic demands of activated retinal neurons. Vasodilation occurs within seconds of flicker light stimulation and quickly reverses upon its cessation. Transient constriction is observed in arterioles but not venules for several seconds after withdrawal of the stimulus. 

Interest in the retinal vascular response to flicker stimulation has been generated from several cross-sectional studies ^3–5 that showed reduced vasodilations in people with type 1 and type 2 diabetes mellitus compared with nondiabetic control subjects. Furthermore, reduced vasodilations have been observed in patients with diabetes before the evolution of typical diabetic retinopathy lesions. ^6,7 The retinal response to flicker stimulation may therefore represent a useful marker of retinal function in these patients. 

Initial work demonstrated good reproducibility of the flicker response in healthy people across ethnic groups. ^8,9 However, these studies incorporated a rest period of at least 30 minutes between consecutive tests to allow sufficient recovery of retinal function. Anecdotal evidence from our department in Melbourne suggests that test participants experience a transient purple discoloration of central vision following measurement of the retinal flicker response. It is unclear if this visual disturbance may represent a short-term disturbance in flicker light–activated retinal pathways. 

The purpose of this study was to investigate whether retinal vasodilation during flicker stimulation is reduced on retesting after 5 and 30 minutes of rest compared with baseline tests. We hypothesized that flicker-induced responses would be reduced for several minutes after the test but would recover by 30 minutes. 

Methods

Participants

We conducted a repeated-measures observational study of retinal vascular responses to flicker light stimulation on the right eye of 20 healthy volunteers aged at least 18 years recruited from the Centre for Eye Research Australia, Melbourne, Australia. Exclusion criteria were smoking, any self-reported medical conditions, eye pathology other than a need for glasses or contact lenses, and tropicamide sensitivity. All participants received measurement of height, weight, blood pressure, heart rate, and IOP to screen for any significant undiagnosed pathology. In addition, participants had their refractive error and best-corrected visual acuity (BCVA) measured with a logarithm of the minimum angle of resolution (logMAR) chart. Abstention from alcohol and caffeine-containing products was requested from all participants in the 12-hour period before the study. 

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

IOP and Hemodynamic Measurements

At the beginning of the study, the IOP and hemodynamic measurements were recorded. The IOP was measured in the right eye after topical instillation of oxybuprocaine hydrochloride, 0.4%, and fluorescein sodium with a slitlamp-mounted Goldmann applanation tonometer (Haag-Streit, Bern, Switzerland). Blood pressure was recorded with a manual sphygmomanometer (Tycos; Welch Allyn, Inc., Skaneateles Falls, NY) and stethoscope (3M Littmann Master Cardiology; 3M, St. Paul, MN). Heart rate was calculated from palpation of the radial artery for 15 seconds. 

Flicker Light–Induced Retinal Vasodilation

Flicker light–induced retinal vasodilation was measured in the right eye with the Dynamic Vessel Analyzer machine (DVA; IMEDOS Systems UGI, Jena, Germany), modified to improve reliability in Asian people. ⁸ This is a commercially available machine that permits reliable continuous recordings of retinal vessels with a minimum width of 90 μm in conscious human subjects. ¹⁰ The system uses green light illumination (530–600 nm) to provide the greatest contrast between retinal blood vessels and the surrounding tissue. Erythrocytes within the vessels have a maximum absorbance of between 400 and 620 nm, whereas most surrounding tissue reflects light in this range. ¹¹ The reflected light is detected by a charge-coupled device camera and used to estimate the width of the vessels. Diffuse luminance flicker is delivered by the machine over the entire 30° visual field by way of an optoelectronic shutter that interrupts the light source with a ratio of bright to dark of 25:1 at a frequency of 12.5 Hz. This frequency is close to values previously shown to maximize retinal vasodilation and blood flow during flicker stimulation. ^12–14 The temporal resolution of the machine was 40 milliseconds, corresponding to 25 vessel diameter recordings per second. 

Topical tropicamide was used for mydriasis, and examinations were conducted in a dimly lit room. Participants focused on the tip of a fixation bar for the duration of each test, while the fundus was examined under green light with an average luminance of 130 cd/m² (ILT1700 Research Radiometer; International Light Technologies, Inc., Peabody, MA). A unique high-contrast region of the fundus (e.g., a vessel branch) was chosen as a fixation target to allow the DVA software (IMEDOS Systems UGI) to track eye movements. Next, we selected for analysis a straight superior temporal arteriolar and venular segment located between 0.5 and 2 disc diameters from the optic disc margin and at least one vessel diameter from any bifurcation or neighboring vessel. 

Vessel diameters were measured for 50 seconds and then during flicker stimulation for 20 seconds and a nonflicker period for 80 seconds. A flicker period of 20 seconds was chosen to allow vessels to dilate maximally without causing unnecessary distress to our participants. ^14,15 The flicker and nonflicker cycle was repeated twice, for a total duration of 350 seconds. The test was repeated after 2 minutes and 50 seconds to allow 5 minutes of rest between flicker periods. A third test was performed 27 minutes and 50 seconds after the second test to allow 30 minutes of rest between flicker periods. Repetition mode was used for retests so that the same vessel segments were used in each participant. Baseline vessel diameters were calculated automatically and expressed in measurement units (MU). Maximum vessel dilation was calculated as the percentage increase in vessel diameter relative to baseline after 20 seconds of flicker stimulation, averaged over the three measurement cycles. 

Statistical Analysis

The mean resting diameters of arteriole and venule segments before flicker stimulation were compared across all tests. We then compared the mean maximum retinal arteriolar and venular dilations relative to baseline during flicker stimulation. Group means were compared with repeated-measures analysis of variance. Where means were significantly different, paired Student's t-tests were used to compare retests with baseline tests using a Bonferroni correction factor of 2. The post hoc intraclass correlation coefficients (ICCs) and 95% confidence intervals (CIs) were calculated for outcomes in tests 1 and 3 using a two-way random-effects model. Multiple linear regression models were used to investigate for any effect of age, sex, refractive error, and timing of tests on vessel diameters and flicker responses. Data were analyzed in STATA (version 12.1; StataCorp LP, College Station, TX). P < 0.05 was considered significant. 

Results

The baseline characteristics of participants are given in the Table. Participants were young (mean [SD] age, 33.1 [5.7] years) and normotensive (mean [SD] systolic blood pressure of 108.4 [10.1] mm Hg and diastolic blood pressure of 67.0 [11.5] mm Hg), with normal IOP (mean [SD], 14.8 [2.8] mm Hg); they were mostly female (70%) and Caucasian (75%). Mild myopia was present (mean [SD], −1.35 [2.05] spherical equivalent), but BCVA was excellent (mean [SD], −0.03 [0.06] with the logMAR chart). 

Table

 

Baseline Characteristics of DVA Participants

Characteristic	Value, n = 20
Age, y	33.1 (5.7)
Ratio of men to women	6:14
Ratio of Asian to Caucasian to other ethnicities	4:15:1
Ratio of brown to other iris pigmentations	10:10
Systolic blood pressure, mm Hg	108.4 (10.1)
Diastolic blood pressure, mm Hg	67.0 (11.5)
Heart rate, beats/min	73.0 (11.3)
Body mass index, kg/m2	24.1 (3.1)
IOP, mm Hg	14.8 (2.8)
Refractive error, spherical equivalent	−1.35 (2.05)
BCVA, logMAR chart	−0.03 (0.06)
Ratio of morning to afternoon tests	9:11

 Data are expressed as means (SDs) unless otherwise indicated.

The diameters of selected arteriole and venule segments before flicker stimulation are shown in Figure 1. The mean (SD) resting arteriole segment diameters were 114.1 (18.8), 114.2 (18.0), and 114.2 (19.2) MU in tests 1, 2, and 3, respectively. The mean (SD) resting venule segment diameters were 134.5 (12.5), 135.5 (12.2), and 135.1 (12.4) MU in tests 1, 2, and 3, respectively. No significant differences were observed between the mean resting arteriolar (P = 0.991) or venular (P = 0.246) diameters in the three tests. 

Figure 1

 

Baseline arteriolar and venular diameters before flicker stimulation in the first and subsequent tests, with 5 and 30 minutes of rest. Data are expressed as means (SDs).

The maximum percentage dilations of selected arteriole and venule segments during flicker stimulation are shown in Figure 2. The mean (SD) arteriolar dilations were 3.23% (2.06%), 2.44% (1.62%), and 3.36% (2.11%) in tests 1, 2, and 3, respectively (P < 0.001). Compared with test 1, arteriolar dilations were significantly reduced after 5 minutes (P = 0.008) but not 30 minutes (P = 0.437) of rest, with adjusted α = 0.025 for multiple comparisons. The mean (SD) corresponding values for venular dilations were 4.26% (1.28%), 3.81% (1.61%), and 4.43% (1.73%) in tests 1, 2, and 3, respectively. However, the mean venular dilations were not significantly different across the three tests (P = 0.128). 

Figure 2

 

Maximum arteriolar and venular dilation during flicker stimulation in the first and subsequent tests, with 5 and 30 minutes of rest. Data are expressed as means (SDs). **P < 0.01 versus the first test.

The reproducibility of measures between tests 1 and 3 was assessed. The ICCs (95% CIs) for resting arteriolar and venular diameters were 0.99 (0.98–1.00) and 0.97 (0.93–0.99), respectively. The ICCs (95% CIs) for arteriolar and venular dilations were 0.94 (0.87–0.98) and 0.56 (0.16–0.80), respectively. 

Age, sex, refractive error, and timing of the tests were investigated for a potential effect on retinal vessel diameters and flicker responses. None of these parameters were significantly associated with baseline vessel diameters or flicker responses or changes in these outcomes between baseline and follow-up tests. 

Discussion

To our knowledge, this is the first study to show that the responsiveness of retinal arterioles to flicker stimulation is reduced in the immediate posttest period. This effect is short lived and appears to last less than 30 minutes, which would account for the good reproducibility of the retinal flicker response reported in other studies. ^8,9 The reproducibility of our baseline test results was excellent after 30 minutes of rest with the exception of venular dilation, which had a high degree of variability. Although the mean venular dilations were reduced after 5 minutes of rest, they were not statistically different. 

No differences were observed in the baseline diameters of selected arteriolar or venular segments over our three tests. This indicates that retinal vessels rapidly return to their resting state on cessation of the flicker stimulus and suggests that our results were not confounded by residual dilation from preceding tests. Furthermore, it implies that impaired arteriolar responsiveness in the immediate posttest period may be due to desensitization of the pathways that mediate the retinal flicker response. 

One possible explanation for reduced arteriolar dilation in the immediate posttest period is a desensitization of cones, particularly the M-cones most sensitive to the green light used in our experiments. Rods were unlikely to be significant mediators of the retinal flicker response in our experiments because rod signaling is suppressed during constant bright light illumination. ¹⁶ Although cones are resistant to visual pigment bleaching at high light intensities, ¹⁶ cone sensitivity is dependent on background light intensity. ¹⁷ Light adaptation from prolonged fundus illumination might therefore have reduced cone signaling in our second test of the flicker response. Further studies are required to determine whether short-term impaired arteriolar responses involve light adaptation in cone photoreceptors. 

In the primate visual system, flickering light activates magnocellular and parvocellular neural pathways to varying degrees according to parameters such as the stimulus frequency and the degree of luminance or chromatic modulation. ^{18, 19} Magnocellular retinal ganglion cells are activated by central and surrounding light impulses and respond optimally to luminance modulation at frequencies of 10 to 20 Hz. ²⁰ In contrast, parvocellular ganglion cells receive opposing input from different wavelength-specific cones and are most sensitive to chromatic modulation at frequencies of 5 Hz or less. ^20–22 Luminance flicker with red or green light produces the greatest increase in optic nerve head blood flow at frequencies of between 10 and 15 Hz. ¹² The magnocellular pathway is probably the major driver of this response, with a lesser contribution from the parvocellular pathway. ^20,23 Chromatic flicker with minimum or no luminance change produces a maximum increase in optic nerve head blood flow and retinal vasodilation with red-green modulation at a frequency of approximately 2 Hz. ^12,15 Low-frequency equiluminant chromatic flicker is thought to be selective for the parvocellular pathway. ^12,20  

Flicker stimulation in our study was delivered by green light at a frequency of 12.5 Hz at a modulation depth of approximately 1, ¹² corresponding to pure luminance modulation. These stimulus parameters are most selective for the magnocellular pathway, with a possible lesser influence on the parvocellular pathway. The reduced arteriolar response to luminance flicker stimulation in our second test may reflect decreased neural activity, particularly through magnocellular retinal ganglion cells. Tests of ganglion cell function and cortical visual processing such as the pattern ERG and visual evoked potential, respectively, may help to identify whether reduced short-term arteriolar responses reflect an impairment at the level of retinal neurons or vascular cells. It is also worth considering whether reduced flicker light–induced retinal vasodilation in patients with diabetes may reflect dysfunctional retinal processing through the magnocellular or parvocellular neural pathways. ^3–5  

The mechanisms by which flicker light causes retinal vasodilation remain unclear, although nitric oxide (NO) has been implicated as an important signaling molecule. Systemic inhibition of NO synthase (NOS) with N ^G-monomethyl-l-arginine reduced the retinal flicker response by approximately half in healthy volunteers. ²⁴ Furthermore, flicker stimulation increases blood flow to the optic nerve head and retina via NO in the cat. ^25,26 Some researchers have interpreted these findings to mean that the retinal flicker response is a measure of endothelial function. ^4,27 However, no study has convincingly identified any NOS isoform activated during flicker stimulation or a cellular source of NO. It is possible that retinal neurons activate neuronal NOS during flicker stimulation and directly mediate vasodilation, without a major contribution from endothelial cells. In addition, evidence from rodent studies ^28,29 suggests that retinal glial cells are important mediators of the flicker response via the release of vasodilatory arachidonic acid–derived epoxyeicosatrienoic acids and prostaglandins. Clearly, more work is required to elucidate the relative contributions of retinal neurons, glia, and vascular cells to the retinal flicker response. 

Strengths of our study include our young and healthy participant sample, which included both Asian and Caucasian participants. Measurement of the retinal flicker response is generally more difficult in Asians with current technology because of their narrow-set eyes, ⁸ but this did not preclude our significant findings. In addition, the simple repeated-measures design of our study provided a precise means to investigate the impact of short-term retesting on the retinal flicker response. 

Limitations of our study include, first, a small sample size, which may have been underpowered to detect a significant reduction in venular dilation on retesting. Second, we did not specifically exclude people with myopia because of our focus on within-subject changes, although high myopia could potentially affect the flicker response. However, this is unlikely to be a major confounder of our results because of the mild myopia and excellent BCVA in our sample. Third, our equipment could only test the flicker response with green light, specific to M-cones. The addition of more light wavelengths may yield larger vasodilations but could render accurate measurements more difficult. Fourth, given that all our participants were healthy, it is unclear whether diabetes or diabetic retinopathy might affect the length of time needed to recover from flicker stimulation. Fifth, our study is unable to provide specific information on the mechanism of impaired arteriolar dilation in the immediate posttest period. 

Follow-up studies should be undertaken to investigate why retinal arteriolar dilation during diffuse luminance flicker stimulation was reduced on retesting with the same parameters after 5 minutes of rest. For example, the influence of light adaptation could be investigated by testing the flicker response after a period of exposure to specific background luminance levels. Conventional ERG parameters could be investigated before and after tests of the flicker response. Higher processing of visual signals could be investigated; if retinal arteriolar dilation was reduced because of desensitization of local neural pathways, we would expect to see a similar reduction in the visual evoked potential amplitude. Given that our flicker parameters were specific to the magnocellular pathway, the effect of low-frequency equiluminant chromatic flicker directed at the parvocellular pathway should also be investigated in future studies. As described above, vessel diameters are best measured under green light, so chromatic flicker should use alternating green-red or green-blue light, while the luminance level is held constant. 

Our finding of reduced arteriolar responsiveness in the immediate posttest period has important implications for the measurement of the retinal flicker response in clinical studies. If this response is to be used as a marker of retinal function in patients with diabetes, then other sources of variation need to be minimized. Researchers using the DVA machine (IMEDOS Systems UGI) should allow at least 30 minutes between consecutive tests, where indicated, and minimize the duration of fundus illumination. 

Acknowledgments

Supported by Postgraduate Medical Scholarship 1038701 (JEN) and Senior Research Fellowship 1045280 (ELL) from the Australian National Health and Medical Research Council. The Centre for Eye Research Australia receives Operational Infrastructure Support from the Victorian Government. 

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