Based on our previous studies, we anticipate a threshold of about 0.3 g, where there is a transition from ocular alignment that prevails in 1 g to that which is normal in 0 g (Karmali et al., J Vestibular Res 16:117-125, 2006). A subsequent model suggests a slightly higher (but not abrupt) transition at about 0.6 g (Beaton et al., Frontiers Syst Neurosci 9, 2015); thus, we predict a switching threshold in the range of 0.3 to 0.6 g.

Based on the specific aims above, the immediate goals are: - Develop and refine instrumentation and procedures for parabolic flight. This includes implementation of the experiment software on a virtual-reality (VR) headset and performance validation. - Perform experiments in parabolic flight. - Perform baseline lab experiments for comparison. This is an ongoing task, which will provide information on normal responses and variability, by which to interpret the results seen in flight. - Analyze and interpret data from parabolic flight.

Accomplishments

Our alignment testing was implemented on a Meta Quest 2 virtual-reality headset. An Xbox hand controller is used by the subject to interact with the headset; modifications were made to the controllers to remove unnecessary (and distracting) controls in order to simplify operation in flight. Extensive validation, verification, and familiarization were performed before the flight. (A single parabolic flight with Zero G Corporation, to verify the operation of the headset in altered gravity, was twice scheduled and canceled.)

Flights took place in the summer of 2023, in Bordeaux, France. Logistics, planning, and documentation for parabolic flights were carried out in association with Novespace, which provided test subjects. These subjects were shared with Gary Strangman’s project. Subjects had previous parabolic flight experience. This partly mimics the spaceflight situation that is relevant to this project: astronauts on their way to the Moon (Mars), who have adapted to 0g during transit, and who are then subjected to altered g during and after lunar (Martian) landing. Twelve subjects were tested at four g levels in parabolic flight: 0, 0.25, 0.50, 0.75. The first flight was dedicated solely to 0g parabolas. In the other three flights, the three other g-levels were presented, in a different order on each flight Baseline measures were made at 1g. Vertical and torsional alignment were obtained for each subject at each g level.

There is an abrupt (rather than smooth) transition in alignment as g level increases. This alignment occurs, in some cases, between 0 and 0.25g. It is unclear if this occurs at a g level less than or greater than 0.16g (lunar). Statistical testing verified these indications of a low-g transition (between 0g and 0.25g) – not a smooth change – as a function of g level. This was seen in six subjects in vertical alignment and ten subjects in torsional alignment. Six TAN subjects and three VAN subjects satisfy a statistical criterion for this low-g transition, which was performed as follows: • remove outliers at each g level (more than three scaled median absolute deviations from the median) • test for no difference between medians at 0.25 and 0.5 g (Kruskal-Wallis, p>0.1) • – AND – • test for difference in medians between 0g and grouped 0.25 and 0.5 g (Kruskal-Wallis, p<0.1).

In some cases, there may be an additional abrupt transition in alignment between the higher g levels (0.5 and 0.75) and 1g. This implies that partial-g is different from both 0g and 1g, with a low-g transition and a high-g transition. This may reflect the novelty of hypo-g relative to lifelong experience at 1g. A significant concern was the stability of the visual images on the VR display under altered gravity levels. We thought that this might be adversely impacted if the system uses inertial sensors (accelerometers) to maintain the stability of the rendered visual scene as the head moves. Efforts to resolve this were aided by discussions with Laurence Harris and Pascal Serrarens, who have experience with the Quest platform in parabolic flight. Attempts to perform a validation flight with Zero G Corporation were unsuccessful. In flight, there was, in fact, initial drift of the visual image during the first few seconds of each transition into a new g level. This resolved quickly, with the image returning to the desired screen location and remaining stable. Operationally this was accommodated by having subjects ignore the initial few seconds of each transition until the image was stable.

Based on our previous studies, we anticipate a threshold of about 0.3 g, where there is a transition from ocular alignment that prevails in 1 g to that which is normal in 0 g (Karmali et al., J Vestibular Res 16:117-125, 2006). A subsequent model suggests a slightly higher (but not abrupt) transition at about 0.6 g (Beaton et al., Frontiers Syst Neurosci 9, 2015); thus, we predict a switching threshold in the range of 0.3 to 0.6 g.

All NASA Institutional Review Board (IRB) approvals have been obtained. The protocol had been on hold with the NASA IRB per guidance from NASA Human Research Program (HRP), until the NASA protocol template and consent form were finalized. This approval was granted in July 2022. A Reliance Acknowledgement is in place, whereby the Johns Hopkins University (JHU) IRB cedes authority to the NASA IRB.

Logistics, planning, and documentation have been carried out for Novespace. Test subjects, provided by Novespace, will be shared with Gary Strangman’s project. These subjects have parabolic-flight experience. This will partly mimic the spaceflight situation that is relevant to this project: astronauts on their way to the moon (Mars), who have adapted to 0g during transit, and who are then subjected to altered g during and after lunar (Martian) landing.

Most notably, our alignment testing has been implemented on a Meta Quest 2 virtual-reality headset. An Xbox hand controller is used by the subject to interact with the headset; modifications have been made to the controllers to remove unnecessary (and distracting) controls in order to simplify operation in flight. Extensive validation, verification, and familiarization are now proceeding. A single parabolic flight with Zero-G Corporation is planned for May, in order to verify the operation of the headset in altered gravity (planning for this flight has been beset by scheduling and maintenance issues). (A significant concern is the stability of the visual images on the VR display under altered gravity levels. This might be adversely impacted if the system uses inertial sensors (accelerometers) to maintain stability of the rendered visual scene as the head moves. Efforts to resolve this have been aided by discussions with Laurence Harris and Pascal Serrarens, who have experience with the Quest platform in parabolic flight.)

Based on our previous studies, we anticipate a threshold of about 0.3 g, where there is a transition from ocular alignment that prevails in 1 g to that which is normal in 0 g (Karmali et al., J Vestibular Res 16:117-125, 2006). A subsequent model suggests a slightly higher (but not abrupt) transition at about 0.6 g (Beaton et al., Frontiers Syst Neurosci 9, 2015); thus, we predict a switching threshold in the range of 0.3 to 0.6 g.

An HRP / Principal Investigator (PI) meeting on 17 March 2022 clarified the likely flight campaign and number of subjects. Consideration is now being given to the desired subject population. Ideally, for this specific project, subjects who are completely naïve to parabolic flight (no previous 0g experience) would be preferred. These subjects would provide information on the initial, un-adapted, “raw” response of the neurovestibular system to various g levels. However, this is unlikely to be feasible, given the available subject pool. On further consideration, a set of subjects with previous parabolic flight (0g) experience may actually better mimic the spaceflight situation that is relevant to this project: astronauts on their way to the Moon / Mars, who have adapted to 0g during transit, and who are then subjected to altered g during and after lunar / Martian landing. The scientific return might not be as “pure” as with inexperienced subjects, but the operational relevance will likely be higher.

Most of the work in the previous grant year has, as before, been dedicated to refining our apparatus and procedures (Beaton et al., 2017) for use in parabolic flight. One concern is that the tablet-computer version of the experiment requires complete darkness, in order for the subject to not see the edges of the computer screen, or other cues, for ocular alignment. In previous parabolic flights we used large shrouds that enclosed the subject’s head along with the computer. This is viable but unwieldy, and can be uncomfortable. Hence, a team of students has been investigating two alternatives:

1. Optical magnification with a headset, so that the tablet-computer screen fills the entire field of vision, and peripheral cues are masked or distorted (and hence rendered unusable for ocular alignment). 2. Virtual reality (VR) implementation on an Oculus Quest headset, which has been completed and undergone initial testing in a clinical setting.

Since mid-2021, we have decided that the magnifier implementation is infeasible and scientifically inadvisable due to high response variance and differences between this and the tablet-computer implementation. Thus, we have turned our attention to the VR implementation.

The following concerns still exist regarding VR: • Need for improved user interface (starting/stopping trials, entering user ID). • Putting stimulus line segments closer to each other. • Verifying compatibility with parabolic flight (radio frequency interference / RFI, electromagnetic interference / EMI). • Checking for interference between multiple devices used in close proximity.

The VR implementation is very appealing, and we have been comparing it to the tablet-computer version in the lab and clinic. We have implemented our ocular alignment (OA) testing routine (VAN/TAN software) on such a unit, which is undergoing testing at JHU. Sample results were obtained by comparing the tablet and VR versions of TAN (torsional alignment), for five subjects. Variability is generally larger for VR, and mean values are also different. A more useful comparison is to see if the two implementations show the same trends across different test conditions (upright, sitting, supine), and this testing is underway. Work is also underway in refining the calibration of the visual display in the VR system, so that meaningful comparisons can be made with the tablet version. (It is a simple matter to measure distances and line offsets on the tablet computer, and determine angular misalignments via trigonometry. The optics of the VR system introduce distortions that are complicated and not completely characterized in the open literature.)

A significant concern is stability of the visual images on the VR display under altered gravity levels. This might be adversely impacted if the system uses inertial sensors (accelerometers) to maintain stability of the rendered visual scene as the head moves. We have been exploring the available technical information on this device, and have discussed the matter with Joe McIntyre, a European Space Agency (ESA) consultant, who was helpful, but could not say definitively if the system would function properly in flight, since that depends on the specific software implementation and hardware settings.

It may be that the only way to know for sure if this will work in altered gravity is to actually test it under those conditions. For this reason, we are pursuing various options to obtain one or two parabolic flights (presumably with Zero G Corp) in the US, early in the next grant year.

Based on our previous studies, we anticipate a threshold of about 0.3 g, where there is a transition from ocular alignment that prevails in 1 g to that which is normal in 0 g (Karmali et al., J Vestibular Res 16:117-125, 2006). A subsequent model suggests a slightly higher (but not abrupt) transition at about 0.6 g (Beaton et al., Frontiers Syst Neurosci 9, 2015); thus, we predict a switching threshold in the range of 0.3 to 0.6 g.

This project will provide information on binocular alignment as a measure of otolith asymmetry – more specifically as a measure of the neural compensation for asymmetry, which changes as a function of g level. This low-level function is easily and rapidly measured, and has been validated in vestibular patients and parabolic flight. The project draws on related Human Research Program (HRP) initiatives: Sensorimotor Assessment and Rehabilitation Apparatus (NNX10AO19G, 2010-2014), and Assessment of Otolith Function and Asymmetry as a Corollary to Critical Sensorimotor Performance in Missions of Various Durations (80NSSC19K0487, 2019-2027).

During g-level changes there are changes in torsional eye position, often markedly asymmetric. This change in torsional alignment is due to loss of compensation for otolith asymmetry in unusual g environments; normally the nervous system compensates for asymmetries in otolith properties, but in other than 1 g this compensation is inappropriate and yields torsional misalignment. Sudden changes in g level can also lead to differences in the vertical positions of the eyes. This is also thought to be a consequence of an asymmetry between the otolith organs, and has been demonstrated in our parabolic flight and laboratory studies.

In our assessment procedure, two line segments are presented on a computer display. The subject sees one segment with each eye, and adjusts them so that they appear to be aligned with each other (vertically or torsionally). Any actual misalignment of the segments reflects vertical or torsional skew of the eyes, which is a manifestation of uncompensated otolith asymmetry.

Background

Otolith asymmetry

During the g-level changes of parabolic flight there are changes in torsional eye position (Cheung et al. 1994). These changes can be markedly asymmetric (Markham & Diamond 1993, Markham et al. 2000) and are on the order of 1 deg. This change in torsional alignment may be due to loss of compensation for otolith asymmetry in unusual g environments; on Earth, the nervous system presumably compensates for natural asymmetries (e.g., unequal otoconial mass) in otolith properties (von Baumgarten & Thumler 1979), but in other than 1 g this compensation is inappropriate and produces torsional misalignment. A similar disconjugate change has been found during space flight (Diamond & Markham 1998), persisting throughout flights up to 180 days and for many days after flight. Torsional offsets seen in parabolic flight have been proposed as a predictive test for space motion sickness (Diamond & Markham 1991, Markham & Diamond 1993). Motion sickness in parabolic flight has likewise been correlated with differences in counterrolling with tilts to the right and left (Lackner et al. 1987); this is intriguing because it implies a link between motion sickness susceptibility in parabolic flight and in space flight, while other studies have not been able to establish this connection (Oman et al. 1986). More recently, a connection between such vestibular asymmetry and terrestrial motion sickness has also been postulated (Neupane et al. 2018).

Central neural compensation for such asymmetry becomes inappropriate in gravity fields other than 1 g, leading to potentially disruptive changes in ocular alignment (Karmali 2007). A mathematical model of the central compensating mechanisms for otolith asymmetry has been developed based on our findings in parabolic flight (Beaton et al. 2015a).

Vertical ocular alignment

Sudden changes in g level can also lead to small differences in the vertical positions of the two eyes, which can result in double vision (diplopia). This is also thought to be a consequence of asymmetry between the otolith organs on each side of the head, and has been demonstrated repeatedly in our parabolic flight and laboratory studies (Karmali et al. 2006, Karmali 2007). The effect is small but distracting, and would likely occur simultaneous with maximum piloting workload. We have quantified this g-related skew and resulting diplopia, and their adaptation, in parabolic flight (Beaton et al. 2017a).

Hypothesis

Based on our previous studies we anticipate a threshold of about 0.3 g between the response (binocular alignment) present in 1 g and that present in 0g, for a given subject (Karmali et al. 2006). A mathematical model suggests a slightly higher (not as abrupt) transition at about 0.6 g (Beaton et al. 2015a), Thus we predict a switching threshold in the range of 0.3 to 0.6 g.

Specific Aims

1. Evaluate modifications that permit use of our tablet-computer system in ambient lighting (current version requires darkness to eliminate binocular alignment cues).

2. Assess binocular alignment (vertical and torsional skew) in different g levels of parabolic flight.

3. Determine the nature of the function that relates the sensorimotor response of binocular alignment to instantaneous g level: continuous (with what slope) or piecewise-linear (with a threshold).

References

KH Beaton, M Schubert, M Shelhamer (2017a) Assessment of vestibulo-ocular function without measuring eye movements. J Neurosci Methods 283:1-6.

KH Beaton, MJ Shelhamer, DC Roberts, MC Schubert (2017b) A rapid quantification of binocular misalignment without recording eye movements: vertical and torsional alignment nulling. J Neurosci Methods 283:7-14.

KH Beaton, WC Huffman, MC Schubert (2015a). Binocular misalignments elicited by altered gravity provide evidence for nonlinear central compensation. Frontiers Sys Neurosci 9.

BS Cheung, KE Money, IP Howard IP (1994) Human gaze instability during brief exposure to reduced gravity. J Vestibular Res 4:17-27.

SG Diamond, CH Markham (1991) Prediction of space motion sickness susceptibility by disconjugate eye torsion in parabolic flight. Aviat Space Environ Med 62:210-205.

SG Diamond, CH Markham (1998) The effect of space missions on gravity-responsive torsional eye movements. J Vestibular Res 8:217-231.

F Karmali (2007) Vertical eye misalignments during pitch rotation and vertical translation: evidence for bilateral asymmetries and plasticity in the otolith-ocular reflex. Ph.D. Thesis, Johns Hopkins University, Department of Biomedical Engineering.

F Karmali, S Ramat, M Shelhamer (2006) Vertical skew due to changes in gravitoinertial force: a possible consequence of otolith asymmetry. J Vestibular Res 16:117-125.

JR Lackner, A Graybiel, WH Johnson, KE Money (1987) Asymmetric otolith function and increased susceptibility to motion sickness during exposure to variations in gravitoinertial acceleration level. Aviat Space Environ Med 58:652-657.

CH Markham, SG Diamond (1993) A predictive test for space motion sickness. J Vestibular Res 3:289-295.

CH Markham, SG Diamond, DF Stoller (2000) Parabolic flight reveals independent binocular control of otolith-induced eye torsion. Arch Ital Biol 138:73-86.

AK Neupane, K Gururaj, SK Sinha (2018) Higher asymmetry ratio and refixation saccades in individuals with motion sickness. J Am Acad Audiol 29:175-186.

CM Oman, BK Lichtenberg, KE Money, RK McCoy (1986) M.I.T./Canadian vestibular experiments on the Spacelab-1 mission: 4. Space motion sickness: symptoms, stimuli, and predictability. Exp Brain Res 64:316-334.

RJ von Baumgarten, R Thumler (1979) A model for vestibular function in altered gravitational states. Life Sci Space Res 17:161-170.

Progress since July 2020

This project is in the definition phase. Parabolic flights are tentatively planned for the fall of 2021. The study protocol has been submitted to the NASA Institutional Review Board (IRB) and is under review. A Reliance Agreement has been agreed to by Johns Hopkins University (JHU), to cede authority to the NASA IRB.

Most of the work in the initial grant period has been dedicated to refining our apparatus and procedures (Beaton et al. 2017b) for use in parabolic flight. One concern is that the tablet-computer version of the experiment requires complete darkness, in order for the subject to not see the edges of the computer screen, or other cues, for ocular alignment. In previous parabolic flights we used large shrouds that enclosed the subject’s head along with the computer. This is viable but unwieldy, and can be uncomfortable. Hence, a team of students has been investigating two alternatives:

1. Optical magnification with a headset, so that the tablet-computer screen fills the entire field of vision, and peripheral cues are masked or distorted (and hence rendered unusable for ocular alignment).

2. Virtual reality (VR) implementation on an Oculus Quest, which has been completed and undergone initial testing in a clinical setting.

Analysis of the benefits and disadvantages of each implementation reveals the following:

Magnifiers - Pros: Compatible with existing VANTAN tablet technology; Data saved automatically; Adjustable magnification levels; Cheap, lightweight; Wireless; Less training required (don’t need to learn VR control system); Can set up easily and quickly

Magnifiers - Cons: Have to hold tablet or velcro to wall; Some discrepancies between tests with prototype lenses vs. regular glasses; Varying degree of comfort; Somewhat unstable

VR - Pros: Software already developed, ready to use; Headset is comfortable and has only minor light leakage; Software is fast and intuitive

VR - Cons: Small light leakage ; May need to be plugged into laptop during trial to save data ; Software automatically exits after trial, needs to be restarted; Lines are not connected to each other; Headset can be uncomfortable with glasses; Takes time to set up; Multiple hardware components

A comparison of vertical-alignment results (VAN) between the original implementation (tablet-computer in darkness) and the magnifier implementation, for four subjects, shows that there are discrepancies, including much larger variability with the magnifier system.

Work in the next grant period will further investigate these two implementations. We are beginning to favor the VR implementation, which would require the following:

• Reduction of light leakage; • Improvement of user interface (starting/stopping trials, entering user ID); • Moving stimulus line segments closer to each other; • Verifying compatibility with parabolic flight; • Checking for interference between multiple devices used in close proximity.

Based on our previous studies, we anticipate a threshold of about 0.3 g, where there is a transition from ocular alignment that prevails in 1 g to that which is normal in 0 g (Karmali et al., J Vestibular Res 16:117-125, 2006). A subsequent model suggests a slightly higher (but not abrupt) transition at about 0.6 g (Beaton et al., Frontiers Syst Neurosci 9, 2015); thus, we predict a switching threshold in the range of 0.3 to 0.6 g.