The procedures in all experiments followed the tenets of the Declaration of Helsinki and were approved by the Institutional Review Board of the Tohoku University Graduate School of Medicine (2016-2-224-1). This study comprised 20 right eyes of 20 healthy subjects, recruited from volunteers at Tohoku University Hospital in Miyagi, Japan. The characteristics of these subjects are shown in Table 1. To assess the reproducibility of our novel testing protocol, all subjects visited our hospital twice from March to September 2017. The inclusion criteria were (1) age between 20 and 70 years, (2) normal findings in slit lamp and funduscopic examinations, (3) no history of ophthalmic or general disorders, (4) no use of systemic or topical medication, and (5) no prior ocular laser or incisional surgery in either eye. Exclusion criteria were (1) a regular smoking habit, (2) high myopia (i.e., refractive error less than −8 diopters or axial length longer than 26.5 mm), (3) baseline IOP > 21 mm Hg in either eye, (4) pregnancy, and (5) a history of discomfort during pressure on the eye. All subjects abstained from alcohol and caffeine for at least 6 hours before the measurements. Informed consent was obtained from the subjects before the tests. 

Spherical equivalent was measured with an autorefractor keratometer (Tonoref II; Nidek Co., Ltd, Gamagori, Japan). IOP was measured with a handheld tonometer (Icare; Tiolat Oy, Helsinki, Finland), and axial length was measured with the IOL Master (Carl Zeiss Meditec, Dublin, CA, USA). A muscarinic antagonist, 0.4% tropicamide (Mydrin M; Santen Pharmaceutical Co., Ltd, Osaka, Japan) was used for pupil dilation, as previously reported.²⁶ Blood pressure (BP) and pulse rate (PR) were conventionally measured in the brachial artery at the height of the heart (HBP-1300; Omron Colin Co., Ltd. Tokyo, Japan). Mean BP (MBP) and OPP were calculated as follows: MBP = diastolic BP (DBP) + 1/3 (systolic BP [SBP] − DBP); OPP = 2/3 MBP − IOP. 

The principles of LSFG have been described previously.^34,35 Briefly, this instrument consists of a fundus camera equipped with a diode laser and an ordinary charge-coupled device camera. LSFG measures the motion of the speckle pattern created by illuminating the blood cells in the blood vessels with laser light. The main measurement parameter of LSFG is mean blur rate (MBR). LSFG acquires MBR images of the fundus continuously at the rate of 30 frames per second over a 4-second period, and the supplied software then averages these images to produce a composite map of ocular BF. This study used the LSFG-NAVI device (Softcare Co., Ltd., Fukutsu, Japan). Outside the ONH, approximately 90% of the hemodynamic speckle pattern originates in the choroid.^28,29 Within the ONH, the pattern originates in the large vessel and tissue (capillary) areas of the ONH, which the software divides automatically and measures separately.³⁶ The focus of this study was on tissue-area MBR, because it has been reported to be a good indicator of BF in the deep ONH (i.e., the area of the lamina cribrosa).^37,38 LSFG-NAVI can simultaneously capture MBR images in both the choroid and ONH tissue, as a result of its wide capture area that does not require pharmacological pupil dilation.³⁹ We set a circle around the ONH and then set a same-sized circle around the temporal side of the ONH, to mark the choroid region. Furthermore, as all captured MBR images are synchronized to each cardiac cycle, LSFG can be used to calculate many pulse waveform parameters, including skew, blow-out score (BOS), blow-out time (BOT), rising rate (RR), falling rate (FR), flow acceleration index (FAI), acceleration time index (ATI), and resistivity index (RI). Schematic explanations and the formulas used to calculate these variables are described in detail elsewhere.^30,35 In this study, we compared these parameters at baseline and during the test protocol, and expressed the difference as a ratio. We used these ratios to investigate BF dynamics during the protocol in the choroid and in the ONH tissue. 

This study used a newly developed tubular component attached to the LSFG device, the schematic illustration of which was provided in Figure 1A (left panel). The subject's face was firmly attached to the LSFG apparatus with a band (Fig. 1A: green arrow) and then the tube was moved back and forth (Fig. 1A: yellow double-headed arrow) by turning a dial (Fig. 1A: yellow arrow). Moving the tube forward onto the eyelid of the subject, with the eye open, created pressure, thereby raising IOP. We used a handheld tonometer inserted into the tube (Fig. 1A: red arrow) to measure IOP. Then, LSFG measurement was performed similarly, through the tube (Fig. 1A: blue arrow), as quickly as possible (in a few seconds). The protocol had four phases (Fig. 1B): baseline, +10 mm Hg, +20 mm Hg (i.e., IOP was elevated approximately 10 and 20 mm Hg above baseline), and recovery. Baseline clinical characteristics were measured before testing. The average of three baseline LSFG measurements was used for the analysis. The subjects rested for 2 minutes between the +10 and +20 mm Hg phases, and again between the +20 mm Hg and recovery phases. During each phase, LSFG waveform parameters, BP, PR, and IOP were recorded. Changes in these values during the protocol were calculated as the ratio against the baseline (i.e., % change OPP, % change MBR, etc.). The protocol was repeated five times. Comparisons of LSFG waveform parameter alterations during the protocol in the choroid and ONH tissue used averages of the change ratios in the five sets. Additionally, % change OPP in each phase of all five sets was averaged to determine the relationship between % change OPP and % change MBR in the choroid and in the ONH tissue. 

To confirm the reproducibility of measurements obtained during our test protocol, we repeated the test five times, obtaining five sets of measurements: two in an initial visit and three in a second visit (Fig. 1C). Volunteers took 5-minute rests between sets 1 and 2 and between sets 3 and 4. Furthermore, the examiner was changed between sets 3 and 4. After set 4, tropicamide eye drops were instilled into the right eyes of the subjects. After a 15-minute rest, we confirmed pupil dilation and obtained set 5. After excluding the baseline measurements, the coefficients of variation (CVs) for intrasession reproducibility, intersession reproducibility, and interexaminer reproducibility were calculated based on, respectively, % change MBR during sets 1 and 2, sets 1 and 3, and sets 3 and 4. Percent change MBR during sets 4 and 5 was used to assess the significance of differences before and after mydriasis. 

To confirm that the increase in IOP was successfully maintained during LSFG measurement, we performed a preliminary experiment on five eyes of five healthy subjects (mean age: 27.2 ± 3.8 years old; male to female ratio = 2:3). First, we measured the time between the confirmation of increased IOP and the end of LSFG measurement, and then, in a separate experiment, measured the increases in IOP 10 and 20 seconds after the first confirmation of increased IOP. 

All data are shown as the mean ± standard deviation. CVs were used to assess the reproducibility of the BF measurements. The Wilcoxon signed rank test was used to confirm that the increase in IOP was stably maintained 10 and 20 seconds after the initial increase. Analyses of variance (ANOVA) were used to determine the significance of differences in IOP and systemic variables during the provocation. Two-way ANOVA and a post-hoc Dunnett's test were used to analyze the significance of differences in dynamic changes in waveform parameters. Steiger's test, a piecewise linear regression analysis and Davies' test was performed with R software⁴⁰ (version 3.2.5) to determine the linear or nonlinear relationship between OPP and MBR in the choroid and ONH tissue. A Deming regression analysis was also performed, with Sigmaplot 13.0 (Hulinks Inc., Tokyo, Japan). Other statistical analyses were performed with JMP 13.1.0 (SAS Institute Japan, Inc., Tokyo, Japan). The significance level was set at P < 0.05. 

A preliminary experiment showed that the time between the initial measurement of increased IOP and the end of LSFG measurement was 8.2 ± 1.2 and 8.6 ± 1.9 seconds during the +10 and +20 mm Hg phases, respectively. In both phases, increased IOP was stably maintained 10 and 20 seconds after the IOP increase (P = 0.38–0.63, Wilcoxon signed rank test) (Supplementary Fig. S1). 

No adverse events were observed during the provocation and no subjects dropped out of the study. Intrasession reproducibility, intersession reproducibility, and interexaminer reliability were 8.1% ± 6.6%, 8.7% ± 7.0%, and 7.1% ± 6.0%, respectively, for % change MBR in the choroid, and 6.1% ± 4.6%, 7.1% ± 5.7%, 6.4% ± 5.8%, respectively, for % change MBR in the ONH tissue. As shown in Figure 2, mydriasis did not affect BF measurements in either region (P = 0.57–0.96, 2-way ANOVA test). 

As shown in Table 2, IOP rose and OPP decreased significantly during the provocation (P < 0.01), while no systemic variables changed significantly (P = 0.64–0.99). By contrast, as shown in Figure 3, many waveform parameters changed significantly during IOP elevation and then recovered to their baseline level, while ATI and RR did not change during the protocol. Skew, BOS, BOT, and FR decreased during the protocol, and RI increased; these changes were consistent in the choroid and ONH tissue (P = 0.39–0.95, 2-way ANOVA). On the other hand, changes in MBR and FAI were dissimilar in the two regions (P = 0.04). MBR in the choroid decreased significantly during both the +10- and +20-mm Hg phases, while MBR in the ONH tissue decreased significantly only in the +20-mm Hg phase (P < 0.01, Dunnett's test). FAI in the choroid did not change during the protocol, while FAI in the ONH tissue increased significantly (+10-mm Hg phase: P = 0.02; +20-mm Hg phase: P < 0.01, respectively). Figure 4 shows the average waveform of these characteristics in the two regions during the protocol. 

A linear regression analysis showed that % change MBR had stronger positive and negative correlations with % change IOP and % change OPP in the choroid than in the ONH tissue (r = −0.72, P < 0.01 vs. r = −0.48, P < 0.01; r = 0.80, P < 0.01 vs. r = 0.59, P < 0.01, respectively). Deming regression analysis showed that this tendency was similar for the relationship between % change OPP and % change MBR (Supplementary Fig. S2). On the other hand, % change IOP and % change OPP were more strongly correlated with % change FAI in the ONH tissue than in the choroid (r = 0.52, P < 0.01 vs. r = 0.30, P = 0.02; r = −0.53, P < 0.01 vs. r = −0.32, P = 0.01, respectively; Supplementary Fig. S3; Fig. 5). In a further analysis, the correlation coefficient between % change OPP and % change MBR was significantly higher in the choroid than in the ONH tissue (P < 0.01, Steiger's test). Furthermore, % change MBR in the choroid and in the ONH tissue were plotted against % change OPP and fitted to a piecewise (two-segmental) linear regression (r = 0.83, P < 0.01 and r = 0.67, P < 0.01, respectively). The breakpoint of the two-segmented regression line was statistically significant in the ONH tissue and was not significant in the choroid (breakpoint = 53.5%, P = 0.02 and breakpoint = 99.9%, P = 0.09, respectively). The breakpoint in the ONH tissue (53.5%) indicated the point for % change OPP at which MBR in the ONH tissue started to decline (i.e., it defined the lower limit of autoregulation). 

This study investigated a novel LSFG-based testing protocol and used it to examine the effects of IOP elevation on ocular BF dynamics in healthy volunteers. The protocol was easy to use, safe, reproducible and was not affected by pupil dilation. LSFG-measured MBR decreased during the IOP elevation phase of the protocol, before recovering to baseline. Moreover, the choroid and ONH tissue had different hemodynamic characteristics during IOP elevation, including MBR and FAI responses and the relationship between % change OPP and % change MBR. Thus, the choroid and ONH tissue may have different BF autoregulation capacities. Our novel, LSFG-based protocol may allow the future clinical assessment of autoregulatory capacity during IOP elevation. 

Most importantly, our method may be safer and more reproducible than conventional methods, which apply a suction cup directly to the sclera. This requires local anesthesia and is invasive: in a previous report, 2 of 14 subjects could not tolerate this method and dropped out.⁴¹ Here, no subjects dropped out and no anesthesia was used. Additionally, mydriasis did not affect our results, a valuable advantage when studying diseases that do not tolerate dilation, such as angle-closure glaucoma. Our method also had high reproducibility, with all CV values lower than 10%, likely due to using LSFG and a simple, manual procedure. Furthermore, our protocol did not affect systemic variables, indicating that only elevated IOP and decreased OPP acted to lower MBR. This is important, because even if decreased BP and increased IOP cause similar decreases in OPP, their effect on ocular BF is dissimilar.¹⁷ Independent control of IOP thus simplifies data interpretation, as does a stable PR, which affects ocular circulation.^33,42 Thus, we have successfully established a new, convenient method to investigate ocular BF during IOP provocation. 

Specific findings of this study include constant ATI and decreased skew and FR during IOP elevation, suggesting that IOP elevation slows the decrease in BF after the waveform peak (Fig. 4). We also found that MBR, BOS, and BOT decreased, while RI increased (i.e., the pulse waveform became flattened and the difference between maximum and minimum BF increased relative to average BF). This suggests that circulation was unstable and insufficient, agreeing with previous color Doppler imaging (CDI)-derived findings that retrobulbar central retinal artery BF velocity decreases and RI increases during IOP elevation.¹⁵ Increased RI, which can indicate vascular resistance, may occur during IOP elevation because of pressure on the ocular vessels causing decreased vascular capacity.⁴³ Usefully, LSFG has more waveform parameters than CDI and can observe the intraocular microcirculation, key advantages in investigating the pulse waveform of intraocular hemodynamics during IOP elevation. 

Another interesting finding of this study was that the choroid and ONH tissue have unique MBR and FAI responses to IOP elevation. MBR decreases are milder in the ONH tissue, possibly due to increased FAI. Furthermore, the two regions had different relationships between % change OPP and % change MBR: a linear slope was better fitted to the choroid, while only the ONH tissue showed a statistically significant breakpoint in a piecewise linear regression. This confirms previous findings that autoregulation exists in the ONH, but is either lacking or weak in the choroid.^11,44 Choroidal and ONH tissue both originate in the short posterior ciliary artery, suggesting that a local autoregulatory system might exist in the deep region of the ONH. FAI represents the maximum change every 1/30 seconds in a rising curve and is clinically useful.^45,46 Our study shows that increased FAI during IOP elevation might also represent local autoregulatory capacity in the ONH tissue. Previous studies on autoregulation, using different protocols and species, reported widely varying lower limits (i.e., % change OPP): 40% to 80%.^{19,41,47–49} Our finding that the % change OPP breakpoint was 53.5% is within this range and is reasonable. Therefore, to prevent ischemic damage, in addition to mechanical stress in the ONH, clinicians should be alert to this breakpoint for % change OPP-represented autoregulation, especially in glaucoma care. 

Finally, it is promising that the quickness of our testing protocol enabled the assessment of dynamic autoregulation in human subjects. Dynamic autoregulation acts within seconds to keep local BF constant despite acute provocations and OPP reduction. In contrast, static autoregulation is a steady-state effect that acts over several minutes.^19,50,51 Glaucoma does not affect static autoregulatory capacity in animals or human subjects,^18,19 but does affect dynamic autoregulation in animal models.⁵⁰ Thus, investigating the human dynamic autoregulatory capacity may shed new light on glaucoma pathogenesis. Furthermore, local autoregulation relies on myogenic mechanisms, that is, smooth muscular or vessel endothelial responses, and metabolic mechanisms, that is, glial or neuronal responses.⁵² The myogenic mechanism acts over several minutes, while the metabolic mechanism requires only seconds. Thus, our findings suggest that neuronal- or glial-mediated mechanisms exist in the ONH tissue, but are lacking or weak in the choroid. 

Study limitations comprised, first, the inclusion of only young, healthy Japanese volunteers; moreover, the burden of provocation testing limited us to only 20 subjects because of ethical considerations.^15,18,41 Thus, larger studies with more varied subjects are necessary to confirm our findings. Second, although LSFG's reliability matches the microsphere and hydrogen gas clearance methods in animals, its reliability in deeper ONH scans is still unconfirmed.^37,38 Third, it is possible that a slight IOP decrease during the short (<9 seconds) gap between IOP and LSFG measurements affected our results. However, a preliminary experiment showed that IOP did not decrease significantly for up to 20 seconds, indicating that this effect was minimal. Nevertheless, features such as a device-coupled IOP-monitoring sensor might improve our protocol. Fourth, for ethical reasons, we induced very short-term IOP elevation, affecting only dynamic autoregulation. Studying chronic high-IOP diseases, such as glaucoma, will require an animal-model version of our protocol. Nevertheless, protocols for assessing dynamic autoregulation in human subjects will gain in importance. We consider that our protocol should aid the study of diseases involving dysfunctional autoregulation, such as glaucoma, diabetic retinopathy, and age-related macular degeneration. 

In conclusion, this study established an easy, safe, and reproducible protocol to study the effects of IOP elevation, based on a newly designed tubular device and LSFG. This technique showed that increased FAI caused a milder decrease in MBR in the ONH tissue than in the choroid. Furthermore, the relationship between % change OPP and % change MBR differed in these regions, suggesting that the ONH tissue has a local autoregulation system. Thus, our novel LSFG protocol should allow the assessment of dynamic autoregulation capacity in the ONH tissue. 

The authors thank Tim Hilts for reviewing the manuscript, and technical support by Shiori Suzuki and Minami Yoshida. 

Supported by the JSPS KAKENHI Grants-in-Aid for Scientific Research (B) (TN; 26293372), for Exploratory Research (TN; 26670751), and by JST Center for Revitalization Promotion. 

Disclosure: N. Kiyota, None; Y. Shiga, None; K. Ichinohasama, None; M. Yasuda, None; N. Aizawa, None; K. Omodaka, None; N. Honda, None; H. Kunikata, None; T. Nakazawa, None