Short-term Structural and Microvascular Ocular Changes Induced by Hypergravity Exposure During Human Centrifuge Training

Abstract

Purpose

To analyze the short-term effects of hypergravity on ocular parameters, particularly retinal and peripapillary microvasculature changes, in participants undergoing human centrifuge training.

Methods

This prospective, observational study enrolled healthy trainees who participated in centrifuge training at the National Army GangShan Aviation Training Center, Kaohsiung City, Taiwan, from August to September 2023. Ocular data were collected at four time points: 24 hours before training, immediately after, and 15 and 30 minutes after exposure to gravitational force along the head-to-foot axis. Assessments included non-contact tonometry, optical biometry, and optical coherence tomography angiography. Nonparametric statistical methods were used for data analysis.

Results

Nineteen participants (12 male, seven female) with a mean age of 27.89 ± 4.4 years were included. Intraocular pressure, corneal curvature, and pupil diameter remained unchanged after centrifugation. However, axial length decreased immediately after centrifugation (25.79 ± 1.54 mm vs. 25.77 ± 1.57 mm; P = 0.012), then rebounded at 15 minutes (25.81 ± 1.53 mm) and 30 minutes (25.81 ± 1.54 mm; both P = 0.005). Central corneal thickness increased and remained elevated for 30 minutes. Retinal and peripapillary retinal nerve fiber layer thicknesses significantly increased after training (P < 0.05). Meanwhile, the parafoveal and perifoveal vessel density of the right eye showed a decreasing trend immediately after hypergravity exposure, followed by a rebound, although the change was not statistically significant.

Conclusions

Short-term exposure to hypergravity induces transient yet measurable alterations in ocular parameters, particularly retinal thickness and vascular density. These novel findings suggest potential mechanisms underlying hypergravity-associated visual impairment and highlight the importance of monitoring ocular health in individuals exposed to high-G environments.

Vision impairment experienced by an aircrew during flight could potentially lead to disastrous consequences. The acceleration experienced during flight can result in a positive gravitational force along the head-to-foot axis (+Gz), which leads to the pooling of blood in the lower extremities and reduced blood perfusion to the brain. Individuals, such as jet fighter pilots and crew members, face a heightened risk of experiencing high gravity (g)-induced symptoms, ranging from vision blurriness and gray-out to black-out, incapacitation, and in extreme cases, g-induced loss of consciousness (g-LOC).^1,2 In a survey conducted among Chinese aircrew, more than 50% of them reported experiencing visual symptoms related to +Gz forces at least once.³

The implementation of centrifuge high-g training not only improved the aircrew's understanding and tolerance to g stress but also facilitated aeromedical research.⁴ Current medical and scientific interest in gravitational stress is primarily centered on the cardiovascular and nervous systems.^5–7

Visual impairments under high g-forces are believed to result from brain ischemia and subsequent retinal hypoperfusion, given that the retina and optic nerve are extensions of the central nervous system. Although Duane⁸ described a three-stage ophthalmoscopic retinal vascular response during high-g exposure, quantitative assessments of structural and vascular ocular changes under hypergravity remain limited.^9–11 Moreover, although optical coherence tomography (OCT) and OCT angiography (OCTA) have been increasingly used in spaceflight research, no prior studies have quantitatively evaluated posterior ocular changes in hypergravity environments.^12,13

In this study, we used a human centrifuge to simulate the actual conditions experienced by high-performance aircraft pilots, subjecting individuals to short-duration exposure to hypergravity. The objective of this study was to assess the impact of hypergravity on ocular structural and vascular parameters.

Methods

The prospective, observational study adhered to the Declaration of Helsinki and was approved by the Institutional Review Board of the Tri-Service General Hospital in Taiwan (identifiers, A202305114). Before the study, informed consent was obtained from each participant after a detailed explanation.

The participants comprised healthy trainees who underwent centrifuge training at the National Army GangShan Aviation Training Center in Kaohsiung City, Taiwan, from August to September, 2023. Personal data including date of birth, sex, height, weight, body fat, and medical and ocular histories were recorded.

Individuals with systemic diseases, ocular disorders, or surgical history, except for ocular refractive surgery or current medication use, were excluded from the study. Only OCTA images with scan quality ≥6/10 were included. Data with motion artifacts or segmentation errors were excluded. Trainees were instructed not to wear contact lenses during examinations to avoid measurement interference.

Centrifuge Training Profile

After receiving lectures on g awareness and anti-g straining maneuver (AGSM), trainees wore anti-g suits. They entered the centrifuge gondola and were securely fastened to the aircraft seats, maintaining an upright position with a 13° backward tilt (Fig. 1A). The centrifuge training consisted of two sessions. During the first session, acceleration commenced at 1.4 g and gradually increased by 0.1 g per second until reaching the upper limit at 9 g (Fig. 1B). Notably, anti-g suits were not inflated during this phase. Trainees were instructed to grasp a switch and focus their gaze on a central red light and a yellow light bar at the periphery. The g-force value at which the trainees lost complete visibility of the peripheral yellow light bar, with and without the AGSM, signified the thresholds for their relaxing and straining g tolerances.¹⁴

Figure 1.

(A) The human centrifuge comprises a rotating arm measuring 6.1 m connected to a central pivot, with a gondola situated at the opposite end. The trainee occupies a seat with a 13° backward tilt, oriented to face the direction of gondola rotation. (B) During acceleration, the gondola rotates, generating a gravitational vector orthogonal to the gondola's floor. This results in the trainee inside the gondola experiencing head-to-foot direction acceleration force (+Gz), leading to the accumulation of blood and body fluids in the lower extremities and impairing cerebral blood flow.

After a brief intermission, the second session commences, involving acceleration to 6 g and maintaining this level for a duration of 15 seconds. During this session, the anti-g suit was inflated when the centrifugal force exceeded 2 g.

Ocular Examinations

An optical biometer (Lenstar LS900; Haag-Streit AG, Koniz, Switzerland) was used to collect various parameters, including keratometry, astigmatism, axial length (AXL), central corneal thickness (CCT), anterior chamber depth (ACD), and pupilometry. Each examination entailed recording measurements thrice, which were subsequently averaged for documentation. IOP was measured thrice for mean estimation using an air puff tonometer (HNT-1; Huvits, Anyang-si, Republic of Korea).

We assessed macular and optic disc retinal thicknesses, along with vessel parameters through OCTA, performed using the RTVue XR system (phase 7.0 Version 2018.11.63; Optovue Inc., Fremont, CA, USA). Both eyes were examined, with imaging of the macula and optic nerve consistently initiated in the right eye.

We used three-dimensional orthogonal registration technology to enhance image quality and minimize motion artifacts. We obtained macular OCTA 6 × 6 mm and optic nerve OCTA 4.5 × 4.5 mm for each eye. In the context of macular OCTA, computations were performed for whole retinal thickness (from the internal limiting membrane to the retinal pigment epithelium), vessel density (VD) in the superficial capillary plexus (ranging from the inner limiting membrane to the inner plexiform layer), deep capillary plexus (ranging from outer border of superficial capillary plexus to the outer border of outer plexiform layer) and the area of the foveal avascular zone. The measurement regions, automatically generated by Optovue software, encompassed the fovea (Ø = 1 mm), parafoveal (Ø = 3 mm) and perifoveal (Ø = 6 mm) circular areas. We used the radial peripapillary capillary (RPC) mode for optic nerve head (ONH) OCTA analysis. The thickness of the retinal nerve fiber layer (RNFL) was measured at circumpapillary region and VD was measured at the whole en face image, inside disc, and peripapillary (cpVD).¹⁵ Each scan data point, including scan quality, aberrant segmentation, and motion artifacts, was assessed by two independent reviewers (CCH and KHC). Details of the OCTA imaging acquisition are provided in the Supplementary Methods S1. To mitigate diurnal variation effects, ocular examinations were conducted 24 hours before centrifuge training and repeated immediately after, as well as at 15 minutes and 30 minutes after training.

All participants underwent heart rate (HR), systolic and diastolic upper arm blood pressure measurements (Nissei DSK-1031J; Nissei, Gunma, Japan), and percentage of body fat, leg muscle, and leg fat measurements using a body composition instrument (InBody 270; Upwards Biosystems, Ltd., Taipei, Taiwan) before the OCTA examination. The mean arterial pressure (MAP) was calculated as 1/3 systolic blood pressure (SBP) + 2/3 diastolic blood pressure (DBP). The mean ocular perfusion pressure (OPP) was calculated as 2/3 of the MAP−IOP.¹⁶

Statistical Analysis

We used SPSS version 26.0 (SPSS, Inc., Chicago, IL, USA) for statistical analysis, setting significance at P < 0.05. Descriptive statistics show means with standard deviations, and categorical data are presented as values and proportions. Because of the relatively small sample size (n = 19) and the results of Shapiro-Wilk normality tests indicating that several variables did not follow a normal distribution (e.g., AXL, MAP, SBP), we uniformly applied nonparametric tests to ensure robust and conservative statistical inference. The Mann–Whitney U test compared continuous variables between independent groups, whereas the Wilcoxon signed-rank test evaluated differences within the same group before and after centrifugation. These nonparametric tests do not rely on the magnitude of mean differences but rather assess the consistency and direction of ranked values across paired or independent subjects. Spearman's rank correlation coefficients assessed associations between ranked variables. Furthermore, multiple linear regression analysis estimated the relationship between demographic data and trainees’ relaxing g tolerance.

To complement statistical significance testing, we calculated effect sizes (Cohen's d) for ocular-related parameters across time points (see Supplementary Methods S2). Interpretation of d followed Cohen's conventional thresholds, with 0.2, 0.5, and 0.8 representing small, medium, and large effects, respectively.¹⁷

Results

Nineteen trainees were enrolled in this study. The mean age was 27.89 ± 4.4 (range 22–39) years, and seven of them were female.

Sixteen macular OCTA scans (three baseline, three immediate, five at 15 minutes, five at 30 minutes) from seven participants, and six disc scans (one baseline, one immediate, two at 15 minutes, two at 30 minutes) from three participants, were excluded because of inadequate image quality for analysis.

All participants were Asian with no prior experience of centrifuge training. Table 1 summarizes their characteristics. Male and female trainees differed significantly in height, weight, and body fat, but not in age, leg muscle percentage, or relaxing g tolerance.

Table 1.

Characteristics of the Enrolled Trainees

	Male (n = 12), Mean (SD)	Female (n = 7), Mean (SD)	P Value*
Age (years)	28.75 (4.94)	26.43 (3.10)	0.666
Height (cm)	170.88 (2.94)	162 (6.06)	0.002
Weight (kg)	66.04 (9.27)	55.54 (9.82)	0.022
Body mass index (kg/m2)	23.34 (1.96)	21.11 (3.41)	0.151
Body fat (%)	21.77 (5.05)	28.16 (4.91)	0.021
Leg muscle (%)	98.24 (6.23)	98.04 (6.19)	0.821
Leg fat (%)	135.23 (42.32)	116.04 (48.06)	0.298
Relaxing g (g)	5 (0.62)	5.17 (1.18)	0.610

^*Mean (SD) Mann-Whitney U test

On average, the overall group exhibited relaxing and straining +Gz tolerances of 5.06 ± 0.84 and 8.28 ± 0.87, respectively. During testing, all trainees successfully completed the second session (+6 Gz for 15 seconds); however, two of them experienced a g-LOC during the straining g test. Notably, no trainee had absolute values of more than six spherical or three cylindrical diopters. One had ocular refractive surgery three years prior. Spearman's rank correlation coefficient and multiple linear regression analyses indicate no significant correlations between relaxing g tolerance and subject factors (Supplementary Tables S1, S2).

After human centrifuge training, we observed significant increases in HR, DBP, MAP, and OPP immediately after centrifugation, which remained significant for 30 minutes after centrifugation. Notably, we did not observe any significant change in ocular-related parameters, including IOP, corneal curvature, and pupil diameter before and after gravitational stress. (Fig. 2; Table 2)

Figure 2.

Line graphs of parameters before and after positive acceleration. (A) Heart rate; (B) mean arterial pressure and ocular perfusion pressure; (C) intraocular pressure; (D) axial length; (E) central corneal thickness; and (F) anterior chamber depth. Asterisk (P < 0.05) indicates that the differences in the parameter between the time point and the baseline are statistically significant.

Table 2.

Parameters Before and After Positive Acceleration

		Immediately After		After 15 Minutes		After 30 Minutes
	Baseline Mean ± SD	Mean ± SD	P Value	Mean ± SD	P Value	Mean ± SD	P Value
SBP (mm Hg)	119.68 ± 13.19	130.63 ± 19.94	0.070	117.12 ± 13.86	0.218	116.68 ± 13.38	0.324
DBP (mm Hg)	70.26 ± 10.2	80.38 ± 9.39*	0.033	79.71 ± 9.96*	0.022	78.95 ± 9.29*	0.006
AST (OD) (D)	1.12 ± 0.52	1.14 ± 0.54	0.727	1.13 ± 0.58	0.904	1.13 ± 0.54	0.936
AST (OS) (D)	1.19 ± 0.56	1.26 ± 0.63	0.760	1.19 ± 0.69	0.879	1.25 ± 0.68	0.177
PD (OD) (mm)	4.11 ± 0.6	4.1 ± 0.66	0.962	4.14 ± 0.8	0.983	4.02 ± 0.65	0.571
PD (OS) (mm)	3.89 ± 0.6	3.94 ± 0.72	0.705	3.99 ± 0.73	0.481	4 ± 0.76	0.959

AST, astigmatism; PD, pupillary diameter.

^*P < 0.05 (Wilcoxon signed-rank test).

The AXL of the right eye decreased immediately after centrifugation (25.79 ± 1.54 mm vs. 25.77 ± 1.57 mm; P = 0.012) but increased at 15 minutes (25.81 ± 1.53 mm; P = 0.005) and 30 minutes (25.81 ± 1.54 mm; P = 0.006) after exposure to hypergravity. For the left eyes, although a statistically significant increase in AXL was noted only at 30 minutes (25.75 ± 1.41 mm vs. 25.78 ± 1.40 mm; P = 0.023) after centrifugation, the trend was consistent with the right eyes.

We also observed a consistent trend in the changes in CCT between the right and left eyes. CCT significantly increased immediately (533.32 ± 47.25 µm vs. 544.94 ± 47.72 µm; P = 0.008) and persisted 30 minutes (538.11 ± 46.36 µm; P = 0.038) after exposure to hypergravity (only data of right eye was shown). ACD of the right and the left eyes increased significantly at 30 minutes (3.1 ± 0.37 mm vs. 3.14 ± 0.41 mm; P = 0.047) and immediately (3.1 ± 0.32 mm vs. 3.13 ± 0.32 mm; P = 0.009) after centrifugation, respectively (Fig. 2).

Effect size analysis revealed moderate to large effects among variables with significant p-values. Notably, AXL changes in the right eye exhibited consistently large effect sizes—Cohen's d of 3.400, 3.354, and 4.775 at 0, 15, and 30 minutes, respectively. Despite small numerical differences, these values reflect consistent within-subject changes over time (Supplementary Table S3).

Comparison of Foveal, Parafoveal, and Perifoveal Retinal Full Thickness (FT) Before and After Centrifuge Training

The retinal thickness of whole image significantly increased at 15 and 30 minutes after centrifuge training in both the right and left eyes compared with that of the baseline pretraining data (P < 0.05). Furthermore, the FT in the parafoveal and perifoveal regions exhibited a significant increase immediately after acceleration, persisting for 30 minutes in the left eye. In the right eye, FT significantly increased at 15 and 30 minutes after centrifuge training. Notably, no significant differences in FT in the foveal region were observed in either eye after +Gz exposure (Fig. 3; Table 3). There was also no correlation between MAP/OPP with the retinal thickness (Supplementary Tables S4, S5).

Figure 3.

Changes in retinal thickness at (A) the whole image, (B) parafovea, (C) perifovea, and (D) peripapillary RNFL, and superficial retinal microvascular density at (E) the whole image, (F) parafovea, (G) perifovea, and (H) peripapillary region before and after positive acceleration. Asterisk (P < 0.05) indicates a statistically significant difference compared with baseline.

Table 3.

Changes in Foveal Retinal Thickness and Superficial Retinal Microvascular Density Before and After Positive Acceleration

		Immediately After		After 15 Minutes		After 30 Minutes
	BaselineMean ± SD	Mean ± SD	P Value	Mean ± SD	P Value	Mean ± SD	P Value
Thickness (µm)
OD	256.95 ± 21.19	256.56 ± 21.7	0.699	259.19 ± 24.15	0.276	256 ± 22.63	0.616
OS	258.3 ± 21.13	257.41 ± 20	0.905	259.53 ± 22.92	0.954	256.89 ± 21.81	0.412
Microvascular density (%)
OD	22.49 ± 5.29	22.04 ± 5.49	0.286	24.46 ± 5.89	0.485	23.8 ± 5.82	0.601
OS	22.72 ± 5.86	22.16 ± 5.08	0.653	23.92 ± 5.55	0.449	24.17 ± 5.89	0.257

^*P < 0.05 (Wilcoxon signed-rank test).

Comparison of VD at Foveal, Parafoveal, and Perifoveal Region and Grid-Based Layout Before and After Centrifuge

At baseline, we observed no significant differences in VD between the right and left eyes in the whole image, fovea, parafovea, perifovea region. No significant changes in VD were observed in the foveal region of either eye following +Gz exposure (Table 3). However, temporal changes in superficial VD across the whole image, parafoveal, and perifoveal regions varied between eyes. In the right eye, although the change did not reach statistical significance, VD decreased immediately after acceleration, increased at 15 minutes, and returned to baseline by 30 minutes. In contrast, the left eye exhibited a nonsignificant initial increase, followed by a significant peak at 15 minutes and a subsequent return to baseline (Fig. 3). Figure 4 displays a representative image illustrating the macular superficial microvascular density.

Figure 4.

Representative OCTA images of the superficial vessel density in the macular. (A) Baseline; (B) immediately after; (C) 15 minutes; and (D) 30 minutes.

Regarding deep VD, although most of the data did not achieve statistical significance, the evolving patterns in both the left and right eyes were generally consistent. They mirrored the change trend observed in the superficial VD of the right eye, characterized by an immediate decrease after acceleration, followed by an increase after 15 minutes, and eventually returning to baseline values after 30 minutes (Supplementary Table S6).

Except for the fovea region, there was no statistically significant correlation identified between retinal thickness and the superficial and deep VD of the corresponding area (Supplementary Table S7). The correlation between MAP/OPP with the VD of macular was shown in Supplementary Tables S8 and S9. The foveal avascular zone area did not significantly change after centrifugation (P > 0.05) in either eye. (Supplementary Table S10).

Circumpapillary RNFL Thickness and VD Detected in ONH OCTA

We observed significant increases in the mean thickness of the peripapillary RNFL at 15 min in the right eyes, and both immediately and at 15 min for the left eyes, subsequent to centrifuge training, as compared to the baseline measurements (P < 0.05) (Fig. 3). The optic disc VD values on RPC mode exhibited no statistically significant differences at baseline or across the three subsequent follow-up time points for both the right and left eyes except for the inside disc VD (P < 0.05 at 15 minutes) (Table 4).

Table 4.

Changes in the Vessel Density on Radial Peripapillary Capillary Mode

	RPC Vessel Density/%
	Baseline	Immediately After	After 15 Minutes	After 30 Minutes
Whole image (OD)	49.68 ± 2.01	48.82 ± 2.47	49.57 ± 2.12	49.86 ± 1.87
Whole image (OS)	48.74 ± 2.92	48.21 ± 2.25	49.25 ± 2.28	47.89 ± 2.28
Inside disc (OD)	52.32 ± 6.47	51.93 ± 5.57	53.12 ± 6.02*	52.62 ± 8.5
Inside disc (OS)	53 ± 4.29	54.61 ± 3.7	54.66 ± 3.35*	52.98 ± 5.28
Peripapillary (OD)	51.7 ± 2.02	50.8 ± 3.1	51.6 ± 3.02	51.88 ± 2.18
Peripapillary (OS)	51.28 ± 3.83	51.03 ± 2.9	52.11 ± 3.61	50.92 ± 3.42

RPC, radial peripapillary capillary.

^*P < 0.05 (Wilcoxon signed-rank test).

No significant correlation was found between the thickness of peripapillary RNFL and RPC cpVD, except for an immediate post-training period in the left eyes (Spearman's ρ = −0.721, P = 0.012) (Supplementary Table S7). Additionally, no statistically significant correlation was found between MAP/OPP and RPC cpVD (Supplementary Tables S8, S9).

Discussion

This study investigated the short-term ocular changes to hypergravity in a cohort of young, healthy participants, simulating the acceleration profiles of high-performance military aircraft. Although +Gz tolerance can vary with different centrifuge training protocols, the average relaxing +Gz tolerance observed was comparable to, or exceeded, values reported in previous studies.^18,19 Furthermore, both monotonic correlation and multivariate regression analyses revealed no significant predictors of g tolerance, aligning with earlier research findings.^20–23

In response to reduced baroreceptor input during +Gz exposure, the autonomic nervous system triggers sympathetic activation, increasing HR, stroke volume, and peripheral resistance to maintain cerebral perfusion.^24,25 AGSM, anti-g suit, and elevated catecholamines further amplify blood pressure.^26,27 Accordingly, our study observed significant increases in HR, DBP, and MAP lasting 30 minutes after centrifugation.

IOP elevation under microgravity and head-down tilt has been documented.^28–30 However, studies on IOP under positive g-forces are limited and inconsistent, often involving small samples and low g levels.^31–33 No IOP reduction has been reported under such conditions. In our study, IOP remained stable post-centrifugation, consistent with Eiken et al.,³³ suggesting that IOP is highly resilient to gravitational stress. Under microgravity, the IOP elevation is attributed to the headward movement of bodily fluids and elevated episcleral venous pressure. However, the hydrostatic column effect did not apply under hypergravity in our study. These inconclusive data suggest that IOP under high G forces may be influenced by a combination of factors, such as gravitomechanical compression, muscle contracture, and varies depending on the study design.

Consistent with a previous study, our data demonstrated no significant changes in corneal curvature after high +Gz exposure.^11,34 However, in contrast to a previous study indicating pupil dilation after positive acceleration,^11,35 our findings revealed no significant difference in pupil diameter before and after centrifugation. The variability in individual responses regarding changes in pupil diameter may be attributed to factors such as environmental stimulation, emotions, fatigue, and eye accommodation, which are challenging to tightly control during study.

In this study, AXL initially decreased and then increased at 15 and 30 minutes after acceleration. The early reduction may be attributed to mechanical compression of the globe by surrounding muscles and adipose tissue under high +Gz, as the eye deforms within the rigid orbit. Subsequent elongation may result from the restoration of tissue elasticity.³⁶ Similarly, Delori et al.³⁷ reported transient shortening followed by an overshoot beyond baseline of the anteroposterior diameter in enucleated eyes following high-speed impact. Additionally, previously observed transient choroidal thinning after gravitational stress may further contribute to the biphasic AXL alterations.³⁸

Our data showed a 1% to 2% increase in CCT and a significant increase in ACD after acceleration. We hypothesize that under hypergravity, the lens shifts posteriorly, moving the iris diaphragm closer to the lens. The increased ACD suggests greater aqueous volume, whereas the slight CCT rise may result from hydrostatic gradients driving fluid into the stroma. These findings align with Tsai et al.¹¹ who reported a 10% CCT increase after 9 Gz for 15 seconds. Differences in g-force intensity and duration may explain the variation in the degree of change.

Effect size analysis further supports the presence of physiologically meaningful transient changes. Moderate to large d-values for AXL and CCT indicate consistent within-subject alterations likely linked to hypergravity, reinforcing the findings despite the limited sample size.

Prolonged microgravity exposure is known to cause spaceflight-associated neuro-ocular syndrome, including optic disc edema, hyperopic refractive changes, posterior globe flattening, and thickening of the choroid and RNFL.³⁹ Based on this, we initially hypothesized that hypergravity might lead to RNFL thinning; however, contrary to our expectations, OCT revealed significant increases in FT across all analyzed regions except the fovea, with these changes persisting up to 30 minutes post-centrifugation. Peripapillary RNFL thickening was also observed. Although individual retinal layers were not assessed, it is inferred that the observed thickening of the retina primarily stems from RNFL and GCC expansion. Because the fovea lacks these layers, it showed no measurable change. We infer that the increases in FT and RNFL thickness may be related to inner retinal cell edema resulting from transient retinal ischemia induced by acceleration. This response is similar to that observed in the acute phase of retinal artery occlusion, where thickening of the inner retinal layers is commonly seen.⁴⁰ Our finding contrasts with a previous study reporting reduced choroidal thickness without changes in retinal or RNFL thickness after 6Gz for 30 seconds,³⁸ possibly because of differences in training protocols.

Acceleration exposure affects not only structure but also hemodynamics, particularly vascular stress; however, studies on hypergravity impact on retinal and optic disc perfusion remain limited.⁴¹ OCTA is a noninvasive and efficient tool that enables quantitative analysis of retinal structure and blood flow, while also allowing identification of distinct capillary plexuses, including the superficial and deep layers.⁴² In our study, although not statistically significant, superficial VD decreased in the right eye and increased in the left eye immediately after centrifugation. This discrepancy may be attributed to the time gap between the OCT assessments of the two eyes. The right eye was consistently examined first, resulting in a seven- to eight-minute delay for left eye measurements. This temporal offset may introduce bias, suggesting that right eye data more accurately represent the immediate effects of centrifugation.

Our hypothesis was supported by a previous study that conducted real-time fundus examinations using an ophthalmoscope while the participants experienced +Gz forces.⁸ This study revealed the collapse of the retinal arterioles and their loss of pulsation a few seconds before a man subjectively experienced a blackout. Notably, arteriolar pulsation returned rapidly, accompanied by venous distention, after the cessation of peak g-force. Based on these findings, it can be inferred that retinal VD decreased during acceleration exposure and promptly recovered after the removal of high g-forces. In our study, the increased VD observed after centrifuge training relative to baseline may reflect a compensatory mechanism in response to transient retinal hypoperfusion, potentially facilitated by an increase in OPP. However, these interpretations should be approached with caution, as the observed VD trends did not consistently achieve statistical significance.

Although most deep VD changes were not statistically significant, their temporal pattern in both eyes mirrored the superficial VD variation trend of the right eye—an initial decline followed by recovery. Remarkably, this consistency persisted despite a seven- to eight-minute delay in left-eye imaging, suggesting prolonged hemodynamic alteration in the deep plexus. This aligns with prior studies indicating that the deep capillary plexus, located in a watershed zone, is more susceptible to ischemic stress and slower to recover.⁴⁰ Thus, although the superficial plexus responds more rapidly to gravitational stress, the deep plexus appears more vulnerable to sustained ischemia and delayed vascular normalization.

Additionally, research has shown that the distinct interactions between glial cells and each capillary layer, along with differences in oxygen tension, may reflect the unique anatomical and functional roles of these layers.^43,44 These findings reinforce the notion that vascular changes are layer-specific and time-dependent, possibly explaining the VD asynchrony after gravitational stress.

Although MAP, OPP, and retinal thickness increased after high g-force exposure, retinal thickness changes were not significantly correlated with MAP, OPP, or superficial/deep VD. Similarly, both superficial and deep VD was not correlated with MAP or OPP. This is likely attributable to the complex interplay of both known and unknown systemic compensatory mechanisms that regulate hemodynamics during centrifuge training, warranting further investigation.

No differences in the RPC VD measurements of the whole image and peripapillary regions before and after exposure to gravitational acceleration. Notably, the OCTA measurements displayed a more pronounced impact in the macular region than in the peripapillary region. These findings suggest that the central retinal artery, which is responsible for supplying retinal capillaries, may be more susceptible to the effects of gravitational acceleration than the posterior ciliary circulation, which predominantly contributes to the RPC network. The present data are also consistent with previous research showing that ONH head VD is not significantly affected by the change in mean OPP but is influenced by retinal autoregulation.^45,46

This study had several limitations. The first limitation is the small sample size. Second, participants were younger and less experienced with g-force exposure and AGSM than professional pilots. As a result, ocular parameters affected by +Gz may differ in more experienced pilots. Third, because of mechanical and safety constraints, only pre- and post-centrifuge data were collected, which may not fully reflect real-time changes during +Gz exposure. Finally, choroidal blood flow and structure were not assessed, as enhanced depth imaging OCT was unavailable. The choroid, lacking autoregulation, is vital for outer retinal perfusion and is sensitive to changes in OPP. Previous studies have reported significant post-centrifugal reductions in choroidal thickness.^38,47 Further investigation into choroidal vascular changes could improve understanding of high-gravity effects on ocular perfusion.

Conclusions

After brief exposure to +Gz forces, some ophthalmic parameters showed resilience to gravitational stress, while others experienced temporary changes. Our study offers insight into the short-term effects on retinal thickness and VD in bi-laminar segmentation. Further research is needed to explore the underlying mechanisms of the ocular changes to hypergravity.