Based on previous research findings, adaptive changes in sensorimotor, cardiovascular and muscle function during the microgravity transit phase of exploration missions will be maladaptive during transition to a new gravitational state. The level of functional deficit will be most profound during the acquisition of gravity loads and immediately after landing when the demands for crew intervention for emergency operations will be greatest. While we know the deficits will be greater with longer duration transits, early functional data related to operational tasks are currently unknown and therefore are required to provide an evidence base for characterizing programmatic risks and variability among crewmembers. Furthermore, it is important to assess functional capability of the astronaut or cosmonaut after landing on a planetary surface for the purpose of timely intervention or rehabilitation. Related to intervention and rehabilitation is the question of our ability to predict, based on preflight measurements and testing, the adaptability of crewmembers. Clearly there is considerable variability, both preflight and postflight in test results that correspond to postflight functional impairments (Clément and Reschke 2008). Specifically, can preflight performance on a variety of functional tests be characterized across crewmembers in ways that will enable us to predict not only the magnitude of the immediate postflight impairment but the potential duration of the impairment? Currently in both the Russian and U.S. space programs rehabilitation following space flight has been a one-size fits all approach. For example, current clinical balance assessments done in both Russia and the U.S. following spaceflight do not help identify the underlying postural control deficits responsible for poor functional balance, stance or gait. While this strategy may be somewhat beneficial, improvements are possible. By identifying the specific disorders, an appropriate and correct intervention can be designed and implemented to prevent injury or other serious problems when ground personnel cannot assist the flight crews.
Characterizing the recovery of sensorimotor performance in returning astronauts (First Award Fellowship)
Marissa Rosenberg, Ph.D.
NASA Johnson Space Center
Specific Aim 1a: Use kinematic measurements and task performance from the field test (FT) and other data sets, that I will assist in obtaining, to define a recovery time constant.
I will perform a retrospective analysis of the data from the FT and postural stability data derived from the dynamic postural control that currently exist in the Neuroscience Laboratory, and then to apply techniques from this analysis to both preflight and newly acquired postflight data from the FT. It is believed that recovery of functional performance will follow an exponential model: μ(t)-μ0=Ke-Θt , where μ(t) is the mean or median performance as a function of time, μ0 is the (theoretical) performance immediately after landing, K and Θ are variables that characterize the shape of the exponential curve. The time constant is the reciprocal of Θ and in order to characterize this parameter, frequent measurements close to landing are necessary. There is a strong possibility that recovery may be more complicated that a single exponential function. If this is the case additional analysis will be undertaken with the assistance from the Biostatistical Laboratory at JSC. While the analysis will be applied to the results of the entire FT and posture data sets, primary consideration will first be given to the FT Tandem Walk and Sensory Organization Test (SOT) with head movements.
Specific Aim 1b: Develop automated analysis tools using the kinematic data feeds to quantitatively assess decrements in balance, posture, and overall performance in real time. It is vital to know when recovery of postural transitions recover to a safe level, enhancing movement required to leave the landing vehicle and make the crew aware of potential limitations to their immediate performance. My research in this area is based on developing classifiers associated with simple postures (e.g., sit-to-stand, quiet stance, Romberg postures from standard to sharpened, etc.). This will be accomplished using the output from a selected and minimal number of inertial sensors (i.e., head, center of mass locations). Two specific experimental approaches will be used: (1) retrospective analysis from crewmember data acquired pre- and postflight and (2) data acquired from laboratory subjects while using either galvanic stimulation of the vestibular system or reversing prisms as the primary vestibular disturbance variables. The proprioceptive system can also be engaged using foam rubber as the support surface. From these data performance classifiers will be identified and used to construct a decision tree that will support recovery performance factors. It is anticipated that fluctuations in the data feed from the sensors could impact the input to variables within the decision tree. I believe that the problem of inaccurate and fluctuating classification during postural transitions can be overcome using time based filtering (Brusey et.al, 2009). Several filters can be evaluated to help in this process: a naïve voting scheme, an exponentially weighted voting scheme and a Bayes filter.
Specific Aim 2: Use trial-to-trial variability and successful task execution to establish a method of identifying fast vs. slow adaptors, and apply these criteria as a guide for determining potential training methods. Determining a crewmembers ability to adapt and then applying adaptive strategies capable of ensuring appropriate sensorimotor transitions from one gravitational state to another has been a significant problem. Adaptation, whether the individual is an astronaut learning new motor strategies required for operational success in different gravitoinertial environments or someone recovering from a life changing sensorimotor event, individuals will apply different strategies of adaptation when confronted with a new task. There are several measurements being acquired under the umbrella of FT that are suitable for investigating adaptation. The criterion for the selection of a specific test is repeated trials within a single test session. Specifically, the ‘Tandem Walk’ and ‘Jump Down’ tests qualify. Adaptation or adaptability within these tests is regarded as the modification of movement from trial-to-trial based on error feedback. I hypothesize that several parameters are associated with behavior during the adaptive phase of movements specific to a particular task. First, the goal to act, or the plan of movement to complete a task remains the same. Second, change occurs with repetition or practice of the task. Third, once adapted, the modified behavior, or adapted state, is not regressive (cannot retrieve the prior or previous state without de-adapting the new behavior with practice). Fourth, individuals will modify behavior from trial to trial at different rates, revealing behaviors that identify fast vs. slow adaptors. To accomplish this task I will first identify specific behaviors associated with the two tasks identified above using the available kinematic data provided (retrospectively using existing crewmember data and data obtained real-time). With this information I will evaluate trial-to trial changes in the planned movement (i.e., the first movement, either a step or jump) across individuals and compare changes in behavior on all subsequent trials by using a model developed by Emken et al. (2007). This model uses the minimization of kinematic error and effort to explain the adaptive process across individuals. An alternative solution would be to model the behavior (tandem walk and jump-down trials) and trial-to trial changes as kinematic error cancellation. This specific aim will permit the answer to questions about training and improving the rate of individual adaptation: (1) can intra and inter-trial correlations be related to rates of acquired adaptation, (2) are there parameters that assist in training slow adaptors to be fast adaptors?
Specific Aim 3: Evaluate the degree of landing motion sickness (LS) and its relationship with dynamic visual acuity (DVA) and to use stroboscopic vision as a potential countermeasure to reduce or mitigate both LS and detrimental changes in dynamic visual acuity. It is my intention to use the motion sickness data collected as a part of the FT and relate, through simple cross correlation, that data to the scores obtained during the DVA testing. This retrospective test is based on the concept that both LS and DVA are primarily driven by vestibulo-ocular (VOR) gain and that VOR gain is related to retinal slip. Progressive changes in both LS and DVA should reflect the recovery process across the results of all postflight behavior.
Motion sickness has a significant impact on the operational objectives that must be accomplished to assure mission success. At this time the primary countermeasure for LS requires the administration of promethazine via the oral route, intramuscular injection, or suppositories. Promethazine is not a benign drug and is frequently administered along with fluid loading after landing for those individuals who are the most debilitated (any movement results in continued vomiting). If there is a positive correlation between LS and DVA, I will request to use stroboscopic illumination (via the stroboscopic glasses developed at JSC) as a potential countermeasure (Reschke et al. 2006; Reschke et al. 2004). These glasses would not be employed until after the third test period on the day of landing so as to not interfere with the adaptation process being pursued under the banner of the JSC FT.
Aside from minimizing risks for the future of human space exploration, the results of this proposal will also benefit clinical patients suffering from vertigo or other balance disorders as well as the elderly, that often suffer from impaired balance. The adaptation that astronauts experience during gravitational transitions mirror many symptoms of vertigo, suggesting that it could be a useful model. One third of patients reporting dizziness have normal vestibular test results (Gordon et al. 1996), suggesting that improved clinical testing could benefit patient care. About 35% of Americans over age 40 have vestibular dysfunction, making them 2.6 times more likely to fall (Agrawal et al. 2009), possibly because of increased sway (Overstall et al. 1997) which is indicative of reduced sensorimotor precision. Individuals over age 50 with normal vestibular tests have increased motion thresholds, suggesting reduced vestibular precision (Roditi et al. 2012). Perceptual roll tilt thresholds are significantly lowered in patients diagnosed with vestibular migraine (Lewis et al. 2011), a disorder that does not have other quantitative diagnostic characteristics, suggesting that enhanced precision is directly correlated with perceptual complaints (e.g., vertigo, dizziness). These correlations between imprecision and functional deficits, without causation, highlight the need for further research.
In addition, many people suffer from terrestrial motion sickness. Current treatment for motion sickness involves pharmacological countermeasures, which impair judgment and reaction times. Developing a non-pharmacological treatment for motion sickness would allow for people with high-risk jobs to maintain peak performance without the hindrance of anti-motion sickness drugs.