Research

  • Current Research
  • Previous Research

Overview

Context-Specific Adaptation of Gravity-Dependent Vestibular Reflex Responses

Principal Investigator:
Mark J. Shelhamer, Sc.D.

Organization:
Johns Hopkins University School of Medicine

NASA Taskbook Entry


Technical Summary

When we move about in the environment, we constantly make use of reflexive motor adjustments in order to maintain posture and balance in reaction to disturbances. Two such motor activities are the movements of the head and the eyes. Impairment of these motor reflexes can lead to disorientation and reduced performance in sensorimotor tasks (such as piloting of spacecraft). Therefore, the adaptive abilities of these systems are important to prevent mishaps during changes in environmental conditions (e.g. gravito-inertial force, GIF).

In the absence of a normal earth gravity field, the dynamics of head stabilization, and the interpretation of vestibular signals that sense gravity and linear acceleration, are subject to change. Transitions between different GIF environments - as during different phases of space flight - provide an extreme test of the adaptive mechanisms that maintain these reflexes. During extended space flight, crew members may live in artificial gravity and make transitions to weightlessness, planetary exploration, and return to Earth. If they learn sensorimotor skills such as piloting in the normal gravity of Earth, will they be able to perform them adequately in the weightless or the artificial gravity environment? More generally, can people have two different sets of vestibular reflexes, which they are able to switch between rapidly? Are there procedures that could help to transfer (or to inhibit) training from one situation to another? These are the main types of questions addressed by our work. The overall goal of this project is the study of context-specific vestibular and oculomotor reflexes. Special emphasis is placed on the use of GIF as a context cue for switching between adapted reflexes. The general approach is to adapt a specific motor response (saccades, VOR, VCR) in one way (e.g. increase in gain) in one GIF condition, and another way (e.g. decrease in gain) in another GIF condition, and then see if the GIF condition itself (the context cue) can recall the previously learned adapted responses.

The knowledge gained from our studies will help us to design adaptation strategies (pre-flight and in-flight) to assist flight crews in making transitions between different gravitoinertial force situations, and can provide design data for spacecraft facilities (artificial gravity, exercise centrifuge) by delineating the limits of human adaptive capabilities.

A prerequisite to the use of context-specific adaptation procedures as a countermeasure is to identify those responses that need to change in a context-specific manner during space flight. There are many physiological changes that occur during flight, but not all of them are adaptive in the sense of bringing performance back to the normal pre-flight level. One must also think in terms of possible detrimental effects during long-duration flight. If reflexes become inappropriately calibrated during extended flight, then this "incorrect" calibration may generalize to the planetary gravity phase. There is certainly evidence that this occurs, as evidenced by difficulties with posture and locomotion immediately after shuttle flight.

Some reflex responses that develop in flight are inappropriate for planetary gravitational fields, while perfectly acceptable for 0-G. An example is the putative reinterpretation, in space, of all otolith stimulation as translation (rather than tilt). As there is no true tilt (change in orientation with respect to gravito-inertial vector) in space, the absence of tilt responses is acceptable there. However, when this response configuration generalizes to a planetary gravity environment (whether Earth 1-G or Mars 0.38-G), it is inappropriate. Thus we might tailor context-specific adaptation to maintain acceptable planetary responses while in flight, in association with artificial gravity or another stimulus arrangement that simulates one requiring tilt responses.

As an example, two responses which may be amenable to context-specific adaptation follow:

  • If artificial gravity is used during long-term flight, it may be desirable to maintain the appropriate sensorimotor calibrations for the rotating (artificial gravity) and non-rotating (weightless or planetary) environments. Normally, exposure to a rotating environment induces disorientation and inappropriate reflex components, due to the action of cross-coupled (Coriolis) rotational stimuli on the vestibular system. Adaptation would involve training subjects to have the appropriate reflex calibrations while rotating and not rotating, and to switch between the two sets immediately when switching environments.
  • Adaptation to the space environment may involve reinterpreting otolith (linear acceleration) stimulation as arising from translation, rather than from combinations of translations and tilts with respect to gravity as on earth (otolith tilt-translation reinterpretation: OTTR). One pre-flight adaptation strategy would be to reduce the tilt component of motor responses to mimic the flight environment (while alternately presenting stimuli to retain the tilt component, to enhance the eventual return to a gravity environment).
Related issues addressed in our experiments include determining 1) effective adaptation schedules, 2) if gravity can be an effective context cue and for which responses, and 3) the role of the cerebellum. The role of the cerebellum in these adaptations is important as a point of fundamental knowledge, but it has serious practical import as well. If, as seems likely, the cerebellum is adversely affected by space flight, then its ability to implement adaptation-based countermeasures is suspect. We would be remiss to propose countermeasures based on adaptation of vestibular responses without assessing the involvement of the cerebellum.

Outline of Individual Sub-Projects
Various experiments investigate the behavioral properties, neurophysiological bases, and anatomical substrate of context-specific learning mechanisms. We use otolith (gravity) signals as the contextual cue for switching between adapted states of the saccadic system, the angular and linear vestibulo-ocular reflexes, and the VCR. (By LVOR we mean the oculomotor response - horizontal, vertical, and torsional - to linear translation of the head and body.)

Context-specific saccade adaptation - We have evidence for context-specificity in human saccades. Two sets of parabolic flight experiments examined the use of instantaneous gravity level (alternating 0-G and 1.8-G) as a context cue for adapted saccadic eye movements. Saccades (rapid eye motions that move the eyes between targets) can be adaptively altered by presenting a target, then moving that target to a new location before the eyes can get to its first location. After several trials, an adaptive sensorimotor mapping takes place, so that the eyes move directly to the new target location when presented with the original target. Ground experiments at Johns Hopkins successfully used vertical eye position, horizontal eye position, head roll tilt, and upright/supine posture as context cues, so that saccades are increased in size in one context (when subjects look upward, or tilt their heads to the right, or are seated upright), and decreased in size in the other context (when subjects look down, or tilt their heads to the left, or are supine). Data from parabolic flight indicate that g-level also can serve as an effective context cue.

Context-specific LVOR adaptation - We demonstrated the ability to use a gravity cue (head orientation) as a context for switching between two different adapted versions of the linear VOR. The gain of the LVOR can be adaptively changed by having the subject view a visual field that moves with him or her on the sled (driving the gain down, since no eye movements in response to head/body translation are required to stabilize the visual field) or view a visual field that moves opposite to sled motion (driving the gain up). We have been able to induce changes in gain that are associated with head roll tilts (context cues) in different directions.

Properties of AVOR and LVOR in squirrel monkey - In the squirrel monkey, we completed baseline investigations of the dynamics of the AVOR with high frequencies and accelerations, revealing interesting nonlinearities which must be understood before adaptive effects can be investigated. Monkey LVOR adaptation studies were also performed, demonstrating adaptive increases and decreases. Torsional eye movement responses to the linear translations did not change significantly after adaptation, suggesting that the translational and tilt components of the LVOR (horizontal and torsional eye movements, respectively) may not be closely coupled. This has implications for paradigms designed to adaptively change tilt-translation interpretation.

Pursuit and the LVOR in humans and in rhesus money - Results in animals indicate that the LVOR is abolished after flocculectomy, and it is greatly impaired in humans with vestibular deficits as well. Pursuit deficits mirror these changes in the LVOR. This suggests that pursuit and the translational LVOR are tightly linked. A separate set of experiments has demonstrated context-specific adaptation of pursuit gain in humans and monkeys. These two results together may form the basis for a powerful strategy to adapt the otolith-mediated translational LVOR.

Properties and adaptation of head-neck reflexes - Experiments at Baylor College of Medicine on adaptation of the VCR also show evidence of context-specificity. These experiments have established baseline properties of the response along different axes, in terms of mathematical models. Adaptation to an artificial increase in inertia of the head has been demonstrated, as manifest by a decrease in head oscillation during body perturbations. The appropriate adapted response was stored by the head-neck control system even after subsequent re-adaptation back to normal inertia: the system responded appropriately to each inertial load to keep head oscillations at the same level. This capacity to switch between two sets of system parameters persists for at least 35 days after the initial adaptation: the appropriate head damping occurred immediately for both normal and increased inertia loads, showing that two sets of damping parameters can exist simultaneously and be switched in and out as needed.

Adaptation to a rotating environment - Short-radius centrifugation (a form of artificial gravity) is a promising potential countermeasure to long-term weightlessness. Unfortunately, it has a number of side effects related to the unexpected effects of head movements in the rotating environment. Transitions between the artificial gravity (rotating) and weightless (non-rotating) environments will likely cause additional problems. Experiments at MIT are investigating the extent to which these side-effects can be overcome through adaptation. Head movements during centrifugation induce discomfort, non-compensatory vestibulo-ocular reflexes, and illusions of body tilt. Significant adaptation occurred following a series of experimental sessions of head turns during rotation in the light, such that these detrimental effects were reduced.

Key Findings and Implications

  • Saccadic eye movements can be adapted in a context-specific manner, using a number of different context cues. The more relevant the context cue is to the response being adapted, the more effective it seems to be in context-switching (e.g. horizontal eye position is a more effective cue for horizontal saccade adaptation than is vertical eye position).
  • The magnitude of GIF (during parabolic flight) can be used as a context cue for switching between adapted saccade states. There is evidence for retention of this adaptation after 8 months. The lunar and Martian g levels can recall adaptations imposed during 0-G. (This and the above result satisfy a modified version of aim 1 of the original proposal; the original aim involved adaptation of the AVOR, but saccade adaptation is more easily accomplished in parabolic flight, and there is evidence that saccade accuracy may be adversely affected during flight, due to alterations in static torsional eye position. The essential component of the aim - use of g level as a context cue - was achieved.)
  • Compensatory eye movements made in response to translational (LVOR) can also be made context-specific, using the orientation of gravity with respect to the head (head tilt) as a context cue. For inter-aural translations, head roll is a more effective context cue than is head pitch. This is analogous to the situation with saccade adaptation: the closer the context cue is to the response being adapted, the more effective it is. (This satisfies a modified aim two of the proposal. It was proposed to adapt phase rather than gain, but gain adaptation has turned out to be easily accomplished, and has more countermeasure relevance.)
  • Sensorimotor adaptation to head movements during short-radius centrifugation (23 rpm, 1-G at the feet) occurs, as quantified by measures of inappropriate vertical eye movements, motion sickness, and illusory tilt. Three 10-minute adaptation sessions produced adaptation that was retained (at reduced level) a week later. Adaptation to head movements to one side did not generalize to head movements in other directions. Full adaptation did not take place; while motion sickness disappears after 10 adaptation sessions, vertical nystagmus and illusory tilt do not. Context-specificity of the adaptation is apparent since subjects did not experience motion illusions when off the centrifuge between test sessions. (This satisfies aim three of the original proposal. Subjects can acquire adaptation to short-radius centrifugation, and move between rotating and normal environments without detriment.)
  • Properties of the head-neck control system (VCR) in three dimensions (roll, pitch, yaw) can be adequately modeled by a relatively simple, second-order linear system, plus a single dead-zone nonlinearity. Adaptation of this system to changes in head inertia can be induced. This adaptation can be made dual-state, such that the appropriate neural control mechanisms for head stabilization change modes immediately upon a change in head inertia. (This satisfies aims four and five: modeling of the vestibular contribution to head stabilization has been accomplished, short-term adaptation has been demonstrated, and some measure of context-specific adaptation to immediate and repeated changes in head inertia has been shown.)
  • Bilateral removal of the flocculus and paraflocculus in rhesus monkey produced almost complete loss of the horizontal LVOR (even after the angular VOR had recovered). Likewise, human cerebellar patients have comparable defects in pursuit and the LVOR, while the AVOR appears to be controlled independently. This suggests that the vestibulocerebellum plays a critical role in the generation of the LVOR, and that there is a tight relationship between the generation of the LVOR and smooth pursuit. This has implications for countermeasures that are based on adapting translation versus tilt responses mediated by the otoliths. (This satisfies multiple aspects of aim six: the role of the vestibulocerebellum in the LVOR, and the role of pursuit in the generation of the LVOR.)
  • A separate experiment showed systematic variations in the axis of eye rotation at different vertical elevations, during pursuit, AVOR, and LVOR. Axis tilts for pursuit and LVOR were almost identical, and different from that for the AVOR, again showing a close relationship between neural processing for pursuit and the LVOR. (This satisfies the remaining portion of aim six: assessment of axis of rotation in pursuit, LVOR, and AVOR.)
  • Context-specific adaptation of smooth pursuit eye movements has been demonstrated in both humans and rhesus monkeys. Using vertical eye position as a context cue, the initial acceleration of the eyes, when presented with a moving target, can be made to decrease with the eyes elevated, and to increase with the eyes depressed. This has implications for context-specific adaptation of some types of otolith-mediated responses, which seem to be at least partly expressed through the pursuit system (see above). (This partially addresses aim eight, which was intended to determine the role of the vestibulocerebellum in context-specific LVOR adaptation. Although the original aim was not dealt with directly, progress was made in the general area by determining the role of the cerebellum in pursuit and the LVOR, and by demonstrating context-specific adaptation of pursuit.)
  • LVOR gain adaptation was induced in squirrel monkeys, and was specific to the frequency used for adaptation. Following adaptation of LVOR gain, there was no significant change in the torsional eye movements to head tilt, suggesting that the responses to head tilt and head translation are not tightly coupled. (This is the initial stage of aim seven, meant to determine the role of the vestibulocerebellum in the adaptive control of the gain and phase of the LVOR. Other parts of aim seven are still under investigation, and have had to await the development of equipment and procedures for LVOR adaptation in the monkey, and lesioning of same.)
As pointed out above, although much of aim eight (role of the vestibulocerebellum in context-specific LVOR adaptation) was not addressed directly, a number of related experiments in humans (not all of which were originally proposed) have made a significant contribution to the overall goal of this aim. In particular, elucidation of the role of the cerebellum in pursuit and the LVOR, demonstration of the close connection between these two responses, and the production of context-specific adaptation of both the LVOR and pursuit, all contribute to understanding the role of the cerebellum in these adaptive processes. This made some of the specific proposed monkey experiments relatively less important. The animal work continues to have relevance, however, in that it will allow more extensive testing over a range of stimulus parameters, and localization of those cerebellar pathways which contribute to adaptation.

Additional Implications, Relationship to NASA Critical Path Issues
Neurovestibular problems have been identified and listed on the NASA "Critical Path Roadmap" for serious problems that could affect a mission to Mars. Some indication of the range of problems and their severity is found in the May 1997 "Final Report of the NASA Task Force on Countermeasures," which states: "Based on the experience of both the cosmonauts and the astronauts, it is apparent that the ability to egress suddenly will be limited unless effective countermeasures for the loss of neuromuscular performance are identified and adhered to rigidly during prolonged spaceflights." Specific problems listed in the report include changes in eye-head coordination, decrements in postural control, sensory illusions such as otolith tilt-translation reinterpretation, and "flashbacks" between 1-G and 0-G states with associated motor dysfunction. Concerns were raised for the effects of these problems on vehicle control and unassisted egress. The issue of "flashbacks" is especially interesting relative to our work, as it indicates the simultaneous existence of two adapted states (one for 0-G and another for 1-G). Knowledge about how to avoid such inadvertent flashbacks, as well as how to make use of contextually-gated dual-state adaptation, is the central aim of all of our studies in this project.

An especially useful aspect of our parabolic flight experiments is that we fly in consecutive years. With the same subject tested each year, we can assess how much the 0-G responses have been maintained throughout the intervening period of 1-G exposure. This is particularly germane to Mars missions, when gravity-based responses which may have been trained before flight may have to be recalled in the Martian gravity environment many months later.

Not only do our studies provide valuable information for the development of countermeasures, they will also provide basic information on adaptive neurovestibular processes. This is especially true of experiments dealing with the role of the cerebellum in motor control, and signal processing of otolith information for the generation of reflex responses in different environments.

One specific clinical implication of these studies is in the area of vestibular rehabilitation (and physical rehabilitation in general). Rehabilitation exercises are generally learned and carried out under supervision in a clinical setting. There is the possibility that inadvertent contextual cues in this setting will be associated with improved performance while in the clinical setting, which will not transfer completely to settings of normal daily living. In this respect, it is useful to know what context cues are most effective, what types of responses can be made context-specific, and how to avoid such context-specificity when it is detrimental (i.e. when generalization is desired).


This project's funding ended in 2000