MRI-based quantification of ophthalmic changes in healthy volunteers during acute 15° head-down tilt as an analogue to microgravity

Abstract

Spaceflight is known to cause ophthalmic changes in a condition known as spaceflight-associated neuro-ocular syndrome (SANS). It is hypothesized that SANS is caused by cephalad fluid shifts and potentially mild elevation of intracranial pressure (ICP) in microgravity. Head-down tilt (HDT) studies are a ground-based spaceflight analogue to create cephalad fluid shifts. Here, we developed non-invasive magnetic resonance imaging (MRI)-based techniques to quantify ophthalmic structural changes under acute 15° HDT. We specifically quantified: (i) change in optic nerve sheath (ONS) and optic nerve (ON) cross-sectional area, (ii) change in ON deviation, an indicator of ON tortuosity, (iii) change in vitreous chamber depth, and (iv) an estimated ONS Young's modulus. Under acute HDT, ONS cross-sectional area increased by 4.04 mm² (95% CI 2.88–5.21 mm², p < 0. 000), while ON cross-sectional area remained nearly unchanged (95% CI −0.12 to 0.43 mm², p = 0.271). ON deviation increased under HDT by 0.20 mm (95% CI 0.08–0.33 mm, p = 0.002). Vitreous chamber depth decreased under HDT by −0.11 mm (95% CI −0.21 to −0.03 mm, p = 0.009). ONS Young's modulus was estimated to be 85.0 kPa. We observed a significant effect of sex and BMI on ONS parameters, of interest since they are known risk factors for idiopathic intracranial hypertension. The tools developed herein will be useful for future analyses of ON changes in various conditions.

1. Background

Extended spaceflight is known to result in abnormal physiological adaptations in humans. For instance, musculoskeletal deconditioning and cardiovascular adaptations have been extensively studied and they have pathologies that are relatively well defined [1–4]. More recently, neuro-ocular functional and structural changes have been identified in a condition referred to as spaceflight-associated neuro-ocular syndrome (SANS) [5]. Crewmembers with SANS often present with choroidal folds, hyperopic shifts, posterior globe flattening and optic disc oedema. These effects can persist long after return to Earth, although some recovery has been shown to occur [6–10]. SANS increases the risk associated with six-month missions to the International Space Station (ISS), requiring mitigation before long-term space habitation or interplanetary expeditions are attempted.

SANS pathophysiology has been studied extensively yet is not well understood [11]. A prevailing theory suggests that cephalad fluid shifts occurring in microgravity result in mild but chronically elevated intracranial pressure (ICP) that elicits adaptations of ophthalmic structures, including distension of the bulbar region of the optic nerve sheath (ONS) [12,13]. Significant increases in ICP may also result in a decrease in the trans-laminar pressure gradient, contributing to hyperopic shift and optic disc oedema [14,15]. This hypothesis has been supported by elevated lumbar puncture opening pressures post-flight [6], and some similarities in clinical signs and symptoms of SANS and idiopathic intracranial hypertension (IIH), such as papilloedema and posterior globe flattening [7]. However, some common symptoms of IIH have not been documented in crewmembers, including chronic headache, transient visual obscurations, tinnitus and diplopia [16]. The discrepancies between SANS and IIH suggest that the aetiology of SANS cannot be explained by elevated ICP alone and is likely multifactorial, requiring further evaluation [11].

To help evaluate the role of ICP in SANS, head-down tilt (HDT) studies have been routinely implemented to simulate the headward fluid redistributions experienced in microgravity [17]. However, microgravity has multiple physiological effects, and HDT is only a partial analogue of the more complex situation that occurs in microgravity. There is evidence suggesting that HDT overestimates the fluid shifts experienced under microgravity. For example, Laurie et al. [9] found an over-representation of optic disc oedema in subjects who participated in strict 6° HDT for 30 days (45%), when compared with the prevalence in astronauts (15%). HDT studies can vary substantially in the tilt angle, duration, ambient environmental controls, use of anaesthesia and ICP measurement methodology (table 1) [20,21,23].

Table 1.

Summary of prior HDT and astronaut studies. Comparison table of related studies that includes details on the study name, study subjects, experimental conditions, the inclusion of ICP measurements, use of anaesthesia and key findings.

study	subjects	experimental conditions	ICP measurements	anaesthesia	key findings
present study	10 males 8 females	supine 30 min, −15° HDT	none	none	increased ONS area during HDT decreased vitreous chamber depth during HDT increased ON deviation during HDT
effect of gravity and microgravity on intracranial pressure Lawley et al. [18]	5 males 3 females	supine position upright seated position acute (5 min) −6° HDT prolonged (24 h) −6° HDT parabolic flight (20 s bouts of microgravity)	invasive	none	seated ICP < supine ICP zero G supine ICP < IG supine ICP IG supine ICP < acute HDT ICP IG supine ICP ≅ prolonged HDT ICP
postural influence on intracranial and cerebral perfusion pressure in ambulatory neurosurgical patients Petersen et al. [19]	4 males 5 females	5 min, −20° HDT 5 min, −10° HDT 5 min, 0° supine 5 min, 10° head-up-tilt 5 min, 20° head-up-tilt 5 min, 90° upright standing	invasive	local	−20° HDT ICP = 19 ± 4.7 mmHg 90° standing ICP = −2.4 ± 4.2 mmHg ICP ∝ 1/ tilt angle
the pressure difference between eye and brain changes with posture Eklund et al. [20]	3 males 8 females	supine upright sitting 9° HDT	invasive	none	upright seated IOP–ICP = 19.8 ± 4.6 supine IOP–ICP = 12.3 ± 2.2 mmHg 9° HDT IOP–ICP = 6.6 ± 2.5 mmHg
effects of head-down tilt on cerebral blood flow in humans and rabbits Yasamusa et al. [21]	32 rabbits	1 h −45° HDT 1 h, −75° HDT	none	α-chloralose	unlike in humans, cerebral blood flow in rabbits did not change during 1 h HDT the authors attribute this to difference in species, angle, anaesthesia, and/or the techniques used to measure blood flow
optic disc oedema, globe flattening, choroidal folds and hyperopic shift observed in astronauts after long-duration spaceflight Mader et al. [6]	7 astronauts	case study of astronauts with ophthalmic abnormalities	invasive post-flight lumbar puncture in 4 astronauts	none	astronauts presented with mildly elevated (21–28.5 cm H2O) opening pressures
quantitative magnetic resonance image assessment of the optic nerve and surrounding sheath after spaceflight Rohr et al. [22]	10 astronauts	long-duration spaceflight (∼6 months)	none	none	ON area, ONS area and ON deviation unchanged after spaceflight

The present study adds to previous work by using a novel image processing technique to quantify changes to ophthalmic structures beyond the ONS under acute application of 15° HDT. Finite-element biomechanical modelling was previously applied to the same set of subjects to estimate the stiffness of the ONS [23]. This study proposes a less accurate but very easy-to-implement way of estimating ONS stiffness from MRIs. We hypothesized that features of elevated ICP would manifest under HDT in the form of ONS distension and decreased vitreous chamber depth. Changes in other parameters such as optic nerve (ON) cross-sectional area, ON deviation and gaze angles were also assessed. The findings in this study can be compared to those from astronaut studies to identify similarities and differences between HDT and microgravity and will help further our understanding of the effects of ICP on ophthalmic structures.

2. Methods

2.1. MRI protocol

High-resolution MRI scans were collected from healthy volunteers on a 3 T MRI scanner (Prisma, Siemens Healthineers, Malvern, PA, USA), using a 20-channel receiving coil. The study was approved by the Georgia Institute of Technology, Emory University, the University of Idaho and NASA institutional review boards (IRBs) and adhered to the tenets of the Declaration of Helsinki. Written informed consent was obtained from all participants under IRB-approved protocol and all subjects were de-identified before data transfer to the University of Idaho for analysis. Identical scans were previously used in a study by Lee et al. [23]. Each subject was scanned in a supine position after 30 min of acclimation. Supine scanning was followed by a 30 min acclimation period in 15° HDT position before a second scan was obtained. Fifteen degrees were chosen to ensure a significant elevation in ICP, thus increasing the chance that observable changes in the ONS would occur, required for the ONS Young's modulus calculation. To minimize motion artefacts due to changes in gaze angles during MRI scans, subjects were instructed to focus on a visual target central to their field of view (FOV). For the quantification of ON and ONS cross-sectional areas, vitreous chamber depth and ONS Young's modulus, T2-weighted oblique coronal spin-echo images were collected with 600 µm slice thickness and spacing, and 253 µm in-plane isotropic pixel size (FOV 65 cm²) (figure 1a). Additional sequence parameters included 170° flip-angle, 820 ms repetition time and 118 ms echo time. For the quantification of ON deviation and gaze angles, T1-weighted sagittal images were collected with 900 µm slice thickness and spacing, and 488 µm isotropic pixel size (FOV 250 × 250 pixels). Additional sequence parameters included a 9° flip-angle, 1900 ms repetition time and 2.32 ms echo time.

Figure 1.

ON and ONS cross-sectional area parameters. Methods for analysis of ON and ONS areas showing (a) example coronal MRI with ON and ONS labelled. (c) Supine and (d) HDT contours resulting from largest ONS area change at 3 mm location posterior to ONH (subject 16R). (e) Three-dimensional representation of ON/ONS trajectory starting at the ONH showing contours at each slice of MRI volumetric data (subject 16R).

2.2. Optic nerve and optic nerve sheath cross-sectional area, vitreous chamber depth and optic nerve sheath Young's modulus estimation

Three-dimensional geometries of the ONS and ON were generated using a semi-automated method previously described by Rohr et al. [22]. In brief, the average background pixel intensity for each scan was selected with a slice-by-slice mask relative to the peak frequency in the pixel intensity histogram, with the intensity ranging from 0 to 4095 (12-bit image). A global threshold was computed and used to contour each MRI slice and contours were filtered based on an isoperimetric difference quotient and point count threshold (figure 1b,c). Linear interpolation between contours (600 µm slice spacing) was applied to obtain the inner and outer contour of the bulbar subarachnoid space located 3 mm posterior to the ONH along the ON trajectory. This 3 mm posterior outer and inner contour was used to quantify the ONS and ON cross-sectional areas (figure 1d). This contour was chosen to be consistent with standard ultrasound procedures in which ONS diameter is measured at 3 mm posterior to the ONH, an area of maximum contrast [24].

We measured the change in vitreous chamber depth under HDT as a quantitative indirect indicator of globe flattening. Specifically, we manually measured the distance between the ONH and lens centre from T1-weighted MRIs. A negative change in vitreous chamber depth under HDT represented shortening of the vitreous chamber. The lens centre and ONH locations were identified in 3D based on multiplanar visual inspection. The lens centre was identified as the centroid of the lens and the ONH was identified as the interface between the centreline of the ON and the outermost boundary of the posterior globe (figure 2). The ONH point was located in the same way for both the vitreous chamber depth measurement and ON/ONS measurements but the exact locations were specific to each scan.

Figure 2.

ON deviation analysis. The method for determining ON deviation is shown, including the manually selected ON centreline trajectory (red) and the straight-line trajectory (yellow) that connects the ONH (green) and a location on the ON that is 20 mm posterior to the ONH (blue). ON deviation was measured as the maximum orthogonal distance between these two trajectories (white).

ONS Young's modulus (E) was estimated at a location 3 mm posterior to the ONH using the Law of Laplace assuming a thick-walled cylinder, namely $E=3r_i^2{r}_o(\mathrm{\Delta }P/2\mathrm{\Delta }{r}_o)/(r_o^2-r_i^2)$, with inner (r_i) and outer radii (r_o) of the ONS measured in the supine position, and pressure change (ΔP) estimated at the optic axial plane as previously described by Lee et al. [23]. Dura thickness (t) was specified as 0.4 mm [25]; thus, the outer radius was defined as r_o = r_i + 0.4 mm [23].

2.3. Optic nerve deviation and gaze angle quantification

The maximum orthogonal distance between the curved ON trajectory and a straight-line trajectory was used to determine ON deviation and is based on the methods described by Rohr et al. [22] (figure 2). In brief, greater than 20 mm of the centreline trajectory of the ON was manually selected with approximately 1–2 mm point spacing using multiplanar reconstruction (MPR) visualization of the T1-weighted MRI sequence described above. The trajectory was up-sampled using spline curve fitting and truncated at 20 mm posterior to the ONH. A trajectory length of 20 mm was used because it includes the region of ON kinking previously reported to be present in astronauts [7]. Manual selections were performed by two operators and then cross-inspected by each operator to confirm data point selection. Operators were blind to subject identity but were not blind to body posture. Inter- and intra-operator reliability was previously found to be acceptable for these methods [22].

Lateral and vertical gaze angles for the left and right eye were quantified with reference to the Frankfort horizontal plane (figure 3a) [26] to determine whether changes in ophthalmic structures were attributable to changes in gaze between supine and HDT postures. Positive and negative lateral gaze values were defined as projections directed in the left and right anatomical FOV. Positive and negative vertical gaze values were similarly defined as projections directed in the FOV superior and inferior to the Frankfort horizontal plane (figure 3b).

Figure 3.

Frankfort plane and gaze angle calculations. Lateral and vertical gaze angles were measured with respect to the Frankfort horizontal plane of the skull. (a) High-resolution T1-weighted sagittal scan oriented to horizontal and coronal planes to show left (LP) and right porion (RP) and left orbitale (LO) point selections and the resulting Frankfort horizontal plane (FHP) in red. (b) Lateral and vertical gaze vectors measured against the FHP and midsagittal planes (MP).

2.4. Statistics

A linear mixed-effects model that accounts for repeated measurements from the same individual was developed

$$\begin{array}{rl}{y}_i& ={\beta }_0+{\beta }_1{x}_{1i}+{\beta }_2{x}_{2i}+{\beta }_3{x}_{3i}+{\beta }_4{x}_{4i}+\hspace{0.17em}{\beta }_5{x}_{5i}+\hspace{0.17em}\hspace{0.17em}{z}_{0i}+z_{1i}{x}_{2i}\\ & +{ϵ}_i,\end{array}$$

where y_i is the measurement of a parameter of interest, β₀ is the baseline (left eye of a male subject in the supine position), x_1i is the posture group of HDT, x_2i denotes the right eye, x_3i, x_4i and x_5i are the age, body mass index (BMI) and sex (female) of the ith subject, respectively. While coefficient β represents fixed-effect sizes, coefficient z represents the random effects and follows a multivariate normal distribution with the mean of 0 and a symmetric variance–covariance matrix

$$\left(\begin{array}{c}{z}_{0i}\\ z_{1i}\end{array}\right)N\left(\left(\begin{array}{c}0\\ 0\end{array}\right),\left(\begin{array}{cc}{\sigma }_0^2& {\sigma }_{01}\\ {\sigma }_{01}& {\sigma }_1^2\end{array}\right)\right).$$

The ‘fitlme’ function in Matlab (v. R2019a Mathworks Corp., Natick, MA, USA) was used to estimate the coefficients and variances in this linear mixed-effects model and test the hypotheses. This model treats the subjects as random with repeated measurements (on the left and right eye), which allows us to account for the variability across subjects and dependence between the two eyes for each subject. It takes posture group, eye, age, BMI and sex as fixed effects, with the corresponding coefficient indicating the effect size. We further tested whether the true effect size was significantly different from zero, which implies that there was a statistically significant difference in the associated parameter between the two posture groups or between the two eyes, or that there was a statistically significant impact of age, BMI or sex on the parameter of interest.

Using this linear mixed-effects model, we calculated p-values for the following six parameters: ON area, ONS area, ON deviation, lateral gaze angle, vertical gaze angle and vitreous chamber depth. Some of these p-values were dependent due to dependence among several parameters of interest. We ignored such dependence and accounted for multiple comparison with the Bonferroni correction, in order to derive more conservative results. To apply the Bonferroni correction, we adjusted the threshold for p-values to be α/m, where α is the experimentwise type I error rate and m is the number of p-values under consideration. With six parameters and five fixed-effect sizes per parameter, we have m = 30, and we considered an α = 0.10.

3. Results

MRIs were collected at supine and 15° HDT for 18 subjects (10 males, eight females) with average height, weight and age of 1.63 ± 0.37 m, 67.5 ± 18.3 kg and 29 ± 12 years, respectively. A summary of all parameter results can be found in table 2. One subject was omitted from the ON and ONS cross-sectional area analysis and two subjects were omitted from the remaining analyses due to motion artefacts in the images.

Table 2.

Summary of quantified changes occurring under HDT. Parameters were developed and quantified to assess changes occurring to ophthalmic structures under HDT. The first column shows the parameter of interest, the second shows the mean measurements in the supine position, the third shows the mean measurements in the HDT position, the fourth shows the mean change and 95% confidence interval (CI), the fifth shows statistical significance and the sixth shows the number of subjects (N) and eyes assessed. ON, optic nerve; ONS, optic nerve sheath.

parameter	supine mean	HDT mean	mean change	95% CI	p-value	N, eyes
ON cross-sectional area (mm2)	9.41	9.49	0.15	−0.12 to 0.43	0.271	17, 33
ONS cross-sectional area (mm2)	29.15	32.33	4.04	2.88–5.21	0.000	17, 33
ON deviation (mm)	1.55	1.75	0.20	0.08–0.33	0.002	16, 32
lateral gaze angle (°)	−2.46	−1.39	0.89	−0.39 to 2.19	0.169	16, 32
vertical gaze angle (°)	−5.75	−5.41	0.10	1.85–2.04	0.921	16, 32
vitreous chamber depth (mm)	19.72	19.55	−0.11	−0.21 to −0.03	0.009	16, 32

3.1. Optic nerve and optic nerve sheath cross-sectional areas and optic nerve sheath Young's modulus estimation

There was a moderate linear correlation (R² = 0.64) of ON cross-sectional area between supine and HDT postures (figure 4a). The change in the cross-sectional area of the ON due to postural change was insignificant, with an average change of 0.15 mm² (95% CI −0.12 to 0.43 mm², p = 0.271). ONS areas tended to increase during HDT for the majority of subjects excluding a few cases with minimal change (figure 5d). The average change in ONS cross-sectional area was 4.04 mm² (95% CI 2.88 to 5.21 mm², p < 0.000). There was strong linear correlation (R² = 0.81) in ONS area between supine and HDT postures.

Figure 4.

Concordance plots for ON and ONS area parameters. Concordance plot of (a) ON and (b) ONS cross-sectional area at 3 mm posterior to the ONH under supine and HDT conditions. Linear regression between both conditions is shown in red. Blue circles and red triangles represent left and right eyes, respectively. Subject numbers are displayed next to their respective data points.

Figure 5.

Boxplots of ON deviation, vitreous chamber depth, ON area and ONS area during supine and HDT postures. Boxplots of (a) ON deviation, (b) vitreous chamber depth, (c) ON area and (d) ONS area showing how parameters of individual eyes changed from the supine to HDT posture.

Young's modulus for the ONS at a location 3 mm posterior to the ONH was estimated for 16 subjects (N = 29 eyes). Subjects were omitted when ONS area was unchanged or decreased during HDT, since the estimate for modulus becomes undefined in such situations. The mean and standard deviation of ONS Young's modulus was 85.0 ± 115.7 kPa.

3.2. Optic nerve deviation, gaze angles and vitreous chamber depth

ON deviation, gaze angles and vitreous chamber depth were quantified in 16 of the 18 subjects (N = 32 eyes). ON deviation increased by an average of 0.20 mm when moving to HDT posture (95% CI 0.08 to 0.33 mm, p = 0.002) (figure 5a). A strong linear correlation (R² = 0.86) between ON deviation in supine and HDT was observed (figure 6a). Changes in lateral and vertical gaze angles (figure 3b) were insignificant, with an average change in vertical gaze angle of 0.10° (95% CI −1.85° to 2.04°, p = 0.91) and an average change in lateral gaze angle of 0.89° (95% CI −0.39° to 2.19°, p = 0.169). Vitreous chamber depth showed an average change of −0.11 mm (95% CI −0.21 to −0.03 mm, p = 0.009) during HDT (figure 5b). There was a strong linear correlation between supine and HDT measurements (R² = 0.95) and subjects with the largest supine vitreous chamber depths had the largest decreases under HDT (figure 6b).

Figure 6.

Concordance plots of ON deviation and vitreous chamber depth parameters. Concordance plot of (a) ON deviation and (b) vitreous chamber depth under supine and HDT conditions. Linear regression between both conditions is shown in red. Blue circles and red triangles represent left and right eyes, respectively. Subject numbers are displayed next to their respective data points.

3.3. Demographic effects and eye unilaterality

All parameters were assessed for covariates including BMI, age, sex and eye to determine whether any of the observed changes could be partially attributed to demographic differences or unilaterality. With the Bonferroni correction at α = 0.10, BMI accounted for a significant decrease in ON area, with an effect size of −0.14 mm² (p = 0.030) and a significant decrease in ON deviation with an effect size of −0.15 mm (p = 0.005). Females had a significant decrease in ONS area with an effect size of −10.98 mm² (p < 0.005) and a significant decrease in vitreous chamber depth with an effect size of −1.39 mm (p = 0.003) when compared with males. A statistically significant difference in the lateral gaze angle between the left and right eyes of 28.45° (p < 0.000) was present as a result of subjects focusing at a medial point in their FOV. All other covariates were insignificant including differences between the left and right eyes.

4. Discussion

Non-invasive automated and manual techniques were developed and applied to MRIs to quantify changes in ophthalmic structures during acute HDT and to estimate Young's modulus of the ONS. Three-dimensional reconstructions of the ON and ONS were generated to measure their cross-sectional areas at a location 3 mm posterior to the ONH. Changes in ONS area under due to HDT were used to estimate the mechanical properties of the ONS. The curved trajectory of the ON was manually defined to identify ON shape changes occurring to the ON under HDT. Vitreous chamber depth was measured as the distance between the manually selected lens centre and the ONH locations. The ONS cross-sectional area was found to increase under HDT while the ON cross-sectional area remained unchanged. ON deviation was found to increase, an indication of increased tortuosity in HDT. While changes in lateral and vertical gaze angles were observed, no significant trends were identified between supine and HDT. A decrease in vitreous chamber depth suggests that acute application of HDT may contribute to alterations in the geometry of the optic globe. A thick-walled cylinder assumption of the law of Laplace was applied to estimate an average ONS Young's modulus of 85.0 kPa. The findings reported here can be compared to structural changes observed in astronauts to better understand the role of CSF redistribution in the development of SANS.

The ONS is known to be sensitive to changes in ICP due to communication between the orbital subarachnoid space surrounding the ON and cranial subarachnoid space [27]. Thus, the ONS distension observed within the bulbar subarachnoid region suggests that CSF pressure was elevated during acute 15° HDT, as expected. The ONS distension observed in most subjects shows that the ONS can expand in response to elevated ICP; however, because of the nonlinear mechanical behaviour of soft tissues, we expect that further increases in ICP would result in reduced or even negligible changes in ONS area [28]. It is possible that subjects who had little to no change in ONS cross-sectional area may have already had stiffened ONS tissues (e.g. due to elevated ICP) when in the supine position, so that ONS expansion was more difficult to detect by MRI. It is also possible that fluid communication within the subarachnoid space of the ON is different in different persons and that the trabeculae may affect pressure transmission along this path. However, we do not know the true reason why these subjects showed no change. The absence of changes in the ON cross-sectional area suggests that the ON is resistant to transmural fluid transport during acute pressure changes, such as due to acute HDT as used in this study. To our knowledge, this is the first study to measure ON area during HDT; however, ON area was also found to remain constant after long-duration spaceflight using identical methods to those reported here [22].

ONS Young's modulus has been estimated using a variety of methods but almost exclusively ex vivo. However, Lee et al. [23] used finite-element modelling to estimate the in vivo stiffness of the ONS in the same subjects reported here, and found a mean Young's modulus of 47 kPa, fairly similar to Young's modulus we calculated in the subarachnoid region of the ONS. Ex vivo measurements of porcine ONS stiffness reported by Wang et al. [29] are orders of magnitude greater (8.57 versus 0.085 MPa). This difference may be attributed to their use of uniaxial tensile testing in the axial direction that ignored the anisotropic nature of ONS tissue. The ONS is thought to be much more resistant to axial deformation than the circumferential deformation that was measured in this study [30]. It should be noted that the effects of the arachnoid trabecula within the subarachnoid space were not considered when estimating Young's modulus of the ONS [31]. Further, using the Law of Laplace for a thick-walled cylinder makes a number of assumptions that are not strictly satisfied, as discussed in detail by Lee et al. [23]. However, it is a simple calculation and may be useful for obtaining approximate tissue properties in vivo.

Tortuosity and kinking of the ON has been observed in astronauts after long-duration spaceflight through the subjective analysis of post-flight MRI images [7]. However, the degree of change in ON tortuosity was not measured because baseline MRI images were not collected in that study. More recent studies that include baseline comparisons did not identify significant changes in ON tortuosity [32] or ON deviation [22] after spaceflight. By contrast, the present study identified significant increases in ON deviation during HDT exposure, which may impose abnormal mechanical stresses on the ON and ONS. To our knowledge, this is the only study to quantify ON deviation in HDT subjects. The discrepancies in ON deviation findings between acute HDT and long-duration spaceflight show that acute HDT does not truly simulate all aspects of microgravity.

A linear mixed-effects model was used to account for the potential impact of demographics and eye unilaterality on the observed structural changes. Age was not found to be a contributing factor to any of the observed changes; however, the range of ages in this study was small. Females showed much greater change in ONS area and vitreous chamber depth when compared with males, suggesting females may be more susceptible to the effects of CSF redistribution. While BMI was higher in males than females in this study, it was only associated with changes in ON area and ON deviation and was not associated with changes in ONS area. These observations are particularly interesting and merit further study, since IIH is much more common in females with high BMI [33].

HDT studies have been used to simulate the fluid redistribution experienced in microgravity [34]. One such study invasively measured ICP in rhesus monkeys under 6° HDT and found that ICP was almost immediately elevated but then slowly reduced over the course of 15 min [34]. This is consistent with a human study where ICP was also found to return to supine levels during prolonged 6° HDT (table 1) [18]. The 30 min HDT duration in the present study may have been sufficient to result in a significant reduction in ICP before scans were taken, although the observed ONS distension suggests ICP was still elevated well beyond supine levels. Tilt angle is another parameter that can affect ICP with many studies showing increasing ICP with greater tilt angles [19,35]. This gives confidence that the 15° tilt angle used in our study resulted in significant ICP elevation [36]. In a study by Marshall-Goebel et al., healthy subjects were placed in three different HDT positions (6°, 12° and 18°) for 3.5 h after which ICP and intraocular pressure (IOP) was non-invasively measured. ICP was found to increase significantly during 18° HDT (p < 0.001), while IOP increased significantly during 12° (p < 0.001) and 18° (p < 0.001) HDT [15]. The increase in IOP indicates a propensity for IOP to counteract globe flattening at lower HDT angles after an adjustment period.

While HDT conditions have shown to simulate headward fluid redistribution experienced in microgravity, the magnitude of redistribution under HDT may be greater. The peripapillary total retinal thickness identified in healthy subjects after 30 days of strict 6° HDT increased to a greater degree than astronauts after a similar flight duration [9]. Furthermore, ONS distension was not identified after long-duration spaceflight when the methods used here were applied to astronauts [22]. The significant ONS distension we identified in HDT subjects potentially supports that HDT could result in a higher ICP than would be expected in microgravity.

The methods reported here involve several limitations. Manual selection of the ONH as a reference point for the purpose of locating the ON position 3 mm posterior to the ONH inherently introduces operator error. Similar error may be introduced in manual selection of the lens centre and ON centreline path. However, we considered this error to be acceptable based on an assessment of inter and intra-operator reliability of these methods performed by Rohr et al. [22], which found intraclass correlation coefficients of 0.67 and 0.68 for inter and intra-operator reliability, respectively. Lee et al. [23] provided some preliminary results and commentary on repeatability for a single subject using the methods applied here. However, an upcoming longitudinal study with multiple healthy control subjects will test the repeatability of the method and natural anatomic variation over time. The flexion of the neck during HDT was not controlled, which may have unanticipated physiological effects and which may have affected the hydrostatic pressure calculations, in turn introducing systemic error into the study. Weaker correlation between the supine and HDT ON measurements compared to the ONS measurements suggests that the automated method is not as sensitive to detecting changes in the ON area. Finally, it was not possible to directly measure the exact magnitude of ICP elevation during HDT.

Quantitative MRI-based assessment of the ON and ONS before and during HDT could help our understanding of SANS and the potential role of ICP in the physiological response of ophthalmic structures. Here, we report no change in ON area, but significant increases in ON deviation and ONS area as a result of HDT. These findings suggest that CSF pressure within the bulbar subarachnoid space was elevated during acute 15° HDT. Further research is warranted to quantitatively assess ON parameters in varying HDT durations and angles.

Ethics

MRI data collection for this study were approved by the Emory University (IRB000092791) and NASA (PRO2603) institutional review boards and adhered to the tenets of the Declaration of Helsinki. Written informed consent was obtained from all participants under IRB-approved protocol and all subjects were de-identified before data transfer to the University of Idaho for analysis.

Data accessibility

Data for this study have been provided in an .xls format in the electronic supplementary material for this manuscript.

Authors' contributions

S.H.S. carried out manual assessments, participated in data analysis and drafted the manuscript; A.M.S. developed automated algorithms, A.Q.F. and S.H.S. carried out statistical analysis and critical revision; C.R.E. conceived the study and critically revised the manuscript; J.N.O. was responsible for MRI acquisition; B.A.M., A.S. and G.C.N. carried out critical revision; B.A.M. designed and coordinated the study and carried out critical revision. All authors gave final approval for publication and agree to be held accountable for the work performed therein.

Competing interests

B.A.M. has received grant support from Genentech, Minnetronix, Biogen, Voyager Therapeutics and Alcyone Lifesciences. B.A.M. and S.H.S. are fulltime employees of Alcyone Therapeutics and B.A.M. is a scientific advisory board member for the Chiari and Syringomyelia Foundation and has served as a consultant to SwanBio Therapeutics, Roche, Praxis Medicines, Voyager Therapeutics, Medtrad Biosystems, Behavior Imaging, Cerebral Therapeutics, Minnetronix, Genentech and CereVasc.

Funding

This study was funded by NASA grant nos. 80NSSC20K0920, NNX16AT06G and 80NSSC19K1298; NASA Idaho Space Grant Consortium grant no. NNX10AM75H; National Institute of Neurological Disorders and Stroke grant no. 1R01NS11128301; the Georgia Research Alliance and the NASA Human Research Program.