Vestibular Experiments in Space

Bernard Cohen; Sergei B Yakushin; Gay R Holstein; Mingjia Dai; David L Tomko; Anatole M Badakva; Inessa B Kozlovskaya

doi:10.1016/s1569-2574(05)10005-7

. Author manuscript; available in PMC: 2026 Mar 10.

Published in final edited form as: Adv Space Biol Med. 2005;10:105–164. doi: 10.1016/s1569-2574(05)10005-7

Vestibular Experiments in Space

Bernard Cohen ¹, Sergei B Yakushin ¹, Gay R Holstein ¹, Mingjia Dai ¹, David L Tomko ², Anatole M Badakva ³, Inessa B Kozlovskaya ³

PMCID: PMC12969068 NIHMSID: NIHMS2145269 PMID: 16101106

Introduction

Life on Earth is conditioned by the constant downward pull of gravity, which can be considered as an equivalent constant upward linear acceleration (Einstein, 1911). During lateral, fore-aft, and vertical translations as well as during centripetal accelerations generated by turning around distant axes, these linear accelerations summate with the linear acceleration of gravity to form a resultant vector, which we term gravito-inertial acceleration (GIA). These linear accelerations increase the magnitude of the GIA and tilt it relative to the head. The otolith organs and body tilt receptors sense the resultant changes in the GIA and generate compensatory and orienting eye movements through the linear vestibulo-ocular reflex (lVOR) to direct and stabilize vision. Angular head movements are sensed by the semicircular canals, which produce compensatory eye movements over the angular vestibulo-ocular reflex (aVOR). The vestibular system also responds to linear and angular accelerations through vestibulo-spinal reflexes to produce body postural movements that stabilize the body when stationary and support it when one is in motion. Additionally, the vestibular system responds to changes in head and body position relative to the GIA to maintain blood pressure and enhance respiration. Thus, the vestibular system plays a critical role in sensing and responding to linear and angular accelerations in a gravitational environment.

In microgravity, the linear acceleration of gravity is reduced to negligible levels, and a standing posture is no longer maintained. Instead, subjects swim in Space rather than walk, and they no longer experience the vertical linear translations that accompany locomotion. Additionally, gravity no longer contributes to the resultant change in the GIA when the head is translated. Thus, the GIA is always in the direction of the translational and centripetal accelerations due to movement. Angular acceleration, which is in a head coordinate frame, is not markedly altered in microgravity. Because of the ubiquitous nature of gravity on Earth and the difficulty of performing experiments on orbit, we know relatively little about how the nervous system is altered by the prolonged absence of gravity.

From the 1970s, the Russian Space Agency made orbital space flight available to INTERCOSMOS scientists from many countries for animal experimentation, primarily on rats. This included Czechoslovakia, Poland, Hungary, Germany, France, and the United States. Between 1983 and 1995, monkeys and rats flew for approximately 1–2 week periods on a Vostok Space Capsule in experiments supported by the Russian Space Agency and the National Aeronautics and Space Aadministration (NASA). Rodents and invertebrates also flew on NASA Space Shuttle missions, particularly on the Neurolab Mission (STS-90) in 1998. The purpose of this chapter is to provide a summary of these experiments. Experiments on rodents and other animal species in the NASA Space Shuttle flights and the Russian COSMOS/BION flights are summarized in the first section. The second section summarizes inflight experiments performed by the Russian Space Agency on the aVOR and on single unit activity in the vestibular nuclei and flocculus of monkeys. Many of these data appear here for the first time. In the third section, we summarize NASA-supported research on rhesus monkeys on the linear and angular vestibuloocular reflex (lVOR and aVOR). The intent was to provide a summary of experimental data on which future experimentation in Space can be expanded. In addition to revealing some new and interesting scientific results, it is hoped that it will become obvious that basic studies, involving subhuman primates as well as rodents should be essential components of any future space-related research. The results of the earliest flights are not included in this chapter, but can be found in references at the end of the bibliography.

SECTION 1: CELLULAR RESPONSES TO ALTERED GRAVITATIONAL ENVIRONMENTS IN ADULT ANIMALS

A number of studies have examined the impact of altered gravity exposure on vestibular system structures at the cellular level. As a whole, they clearly document the modifiability of cellular and subcellular constituents of vestibular pathways in response to altered gravitational stimuli.

Peripheral vestibular system

Some evidence suggests that exposure of adult as well as developing animals to altered gravitational environments can trigger adaptive responses in weight-lending structures. For example, an increase in otoconial mass has been reported in the utricles of adult animals of several species (rats, frogs) following 7 days of exposure to microgravity (Vinnikov et al., 1980; Ross et al., 1985; Ross, 1987; Lychakov et al., 1989). Conversely, the saccular otolithic membrane volume is reduced in rats exposed to 2.3 g or 4.15 g hypergravity through centrifugation (Lim et al., 1974), although otoconial morphology (examined by scanning electron microscopy) and otoconial synthesis (assessed by mRNA expression of the otoconial matrix protein osteopontin) appear to be normal after 2 h to 7 days of hypergravity exposure (Uno et al., 2000).

Increases in otoconial mass have also been observed in the saccular otoliths of newt larvae (Weiderhold et al., 1997) and late-stage embryos of swordtail fish (Weiderhold et al., 2000) reared in Space, although such increases are not observed in older fish. As in adult animals, the saccular otolithic membrane volume is reduced in cichlid fish reared on a centrifuge (Anken et al., 1998), and the volume of statoconia in marine mollusk larvae reared at 2–5 g is diminished in a graded manner as compared with 1 g controls (Pedrozo and Wiederhold, 1994). Similarly, there is a reduction in the number of large otoconia that are present in the anterior peripheral portion of the utricle in rats raised in a 2 g environment for 60 days than in 1 g controls (Krasnov, 1991). Taken together, these results suggest that otoconial mass adapts to fluctuations in the gravitational stimulus, perhaps to maintain a consistent force on the maculae.

A number of studies have addressed the question of morphologic alterations in the sensory hair cells of adult mammalian otolith organs in response to altered gravitational conditions. These studies indicate that short-term (e.g., 7 day) exposure to space flight does not cause otolith receptor cell degeneration in rats (Ross et al., 1985), although the type I hair cells reportedly exhibit abnormal cytologic features such as increased chromatin, and enlarged perinuclear and intercellular spaces (Krasnov, 1987, 1991). However, more prolonged exposure to microgravity clearly impacts hair cell synaptology. In general, these studies indicate that type II hair cells in particular, but also type I cells evidence adaptive structural modifications in response to alterations in the gravitational environment (Ross, 1993, 1994, 2000). Quantitative ultrastructural analysis of utricular hair cells from adult rats sacrificed immediately after landing from a 9-day space flight demonstrated statistically significant increases in the number of ribbon synapses (Ross, 1994). In addition, the number of synapse pairs and clusters markedly increased in the type II hair cells. Nine days after landing, which was the mission duration, synapse counts remained elevated in the flight rats, but were also elevated in the type II hair cells of ground control animals, suggesting that these receptor cells are particularly vulnerable to stress-induced morphologic changes. In a subsequent study (Ross and Tomko, 1998), the mean number of synapses on type II hair cells doubled, and those on type I cells increased by over 40% by the 13th day of a 14-day space flight. Immediately postflight, these synapse counts were reduced to 67% and 13% increases, respectively, over control values. Comparable studies of vestibular hair cells have been conducted utilizing 2–4 g hypergravity exposure for periods of time ranging from 14–30 days (Lim et al., 1974; Lychakov et al., 1988; Ross, 1993). In general, these experiments support Ross’ centrifugation study (Ross, 1993) demonstrating that type I hair cell synapse number does not change in response to hypergravity, but there is a decrement of approximately 40% in type II hair cell synapses. Clearly, the otolith organs of adult mammals have the potential for morphologic reorganization at several sites in response to altered gravity conditions (Ross, 1997).

Since studies of molecular changes in peripheral vestibular cells in response to fluctuations in the gravity stimulus have been initiated relatively recently, generalizations about molecular alterations are premature. One recent study investigated glutamate receptor mRNA expression using RT-PCR on cells of the vestibular ganglion, as well as the vestibular nuclei, and vestibulocerebellum of rats exposed to hypergravity for 2 h–7 days (Uno et al., 2002). In the vestibular periphery, this experiment demonstrated that synthesis of GluR2 (an AMPA glutamate receptor subunit) receptors in vestibular ganglion cells is reduced in rats exposed to 7 days of hypergravity.

Behavioral, physiological and biochemical correlates of vestibular cellular responses

A spinning movement during swimming, termed “looping,” was initially reported in some (but not all) fish at the transition from 1 g to microgravity during parabolic flight (von Baumgarten et al., 1972), and subsequently described in larval cichlid fish and Xenopus laevis immediately following prolonged 3 g centrifugation (Rahmann et al., 1992). Looping has also been reported in killifish in microgravity (von Baumgarten et al., 1975). Fish that evidence looping behavior exhibit a significantly higher asymmetry in otolith size (Anken et al., 1998) and weight (Beier et al., 2002), in comparison with nonlooping siblings. In addition, although the total number of sensory and supporting cells in the utricular maculae is the same, the cell density is significantly lower in looping than nonlooping fish, due to the atypical presence of enlarged epithelial cells (Bauerle et al., 2004).

Adult mammals may be more profoundly affected by changes in the gravitational environment than are fish. Upon landing following 9 days of space flight, adult rats reportedly show substantially reduced locomotion (Ross, 1994). They maintain a posture with their abdomens and tails flat against the cage floor, with their limbs and digits extended. After 9 days at 1 g, flight rats display normal body posture and movement. Similarly, after 14 days of exposure to 2 g through centrifugation, adult rats show profound deficits in righting responses, swimming and balance (Fox et al., 1992). Recovery of normal orientation during swimming requires 4–24 h at 1 g, whereas the righting reflex does not return to normal for 5–7 days (Ross and Tomko, 1998).

Neurophysiological studies have further documented an immediate post-flight alteration in vestibular activity (Boyle et al., 2001). Within the first 16 h post-space flight, utricular afferents in the oyster toadfish were hypersensitive to translational acceleration, but had no change in directional selectivity. The afferent sensitivity returned to baseline levels by approximately 30 h postflight. Since anamniotes such as the toadfish have only type II hair cells, the reported increase in ribbon synapse number (Ross, 1997) could be one explanation for the afferent hypersensitivity.

Central vestibular system

Several studies have addressed the impact of altered gravitational environments on cells of the central vestibular system. Most of these studies have examined the vestibular nuclei and/or vestibulo-cerebellum, and have utilized either a morphological approach, or a marker for cellular activity. Early studies of the cerebellar nodulus of adult rats exposed to 18 days of space flight revealed ultrastructural alterations in Purkinje cell dendrites and mossy fiber terminals (Krasnov and Dyachkova, 1986, 1990). In these studies, the major modification described in the Purkinje cells was a widening of the synaptic cleft at contacts with Purkinje cell dendrites. In mossy fiber terminals, structural alterations included densely packed synaptic vesicles, with unusual clustering of such vesicles near the presynaptic membranes of axodendritic synapses, increased electron density of pre- and postsynaptic membranes, and enlargement of the synaptic gap. Similar changes are observed in nodular mossy fiber terminals from rats flown on 5–7 day missions and sacrificed within 8 h of landing. Moreover, ultrastructural alterations in the primary somatosensory cortex of these flight animals included a profound decrement in the number of axodendritic synapses, and an increase in the number of axon terminals showing “light” degeneration, as well as signs of “superexcitation” including an increase in the number of axon terminals showing dark degeneration. In concert with this finding, increased numbers of synaptic contacts have been demonstrated in neonatal swordtail fish after 16 days in microgravity, specifically in the nucleus magnocellularis of the primary vestibular brainstem region, area octavolateralis (Ibsch et al., 2000).

More recently, the ultrastructure of the otolith-recipient zones of the cerebellar nodulus has been analyzed in tissue from flight and cage-control rats sacrificed in microgravity after 24 h of space flight (Holstein et al., 1999). Qualitative observations of this tissue indicate that several structural alterations occur in the neuropil, and in the Purkinje cell cytoplasm, of the nodulus of rats exposed to 24 h of space flight. These anatomical alterations are not apparent in the cage control animals. Most notably, the cisterns of smooth endoplasmic reticulum that are normally present throughout Purkinje cells are substantially enlarged and more complex in Purkinje cells of the otolith-recipient zones of the nodulus. The increased complexity of the cisterns results in the formation of long, stacked lamellar bodies that are observed throughout entire Purkinje cells, including the somata, dendrites, thorns, and axon terminals. In addition, occasional enormous mitochondria, >1 μm in cross-sectional diameter, are present in the Purkinje cell somata of flight animals. Ultrastructural indications of degeneration and synaptic reorganization are also observed in the molecular layer of the nodulus from the flight animals, but not cage controls. Since these morphologic changes are not apparent in control animals, they are not likely to be due to caging or tissue processing effects. The particular nature of the structural alterations, including the formation of lamellar bodies and the presence of degeneration, suggests that excitotoxity may play a role in the short-term neural response to space flight.

In the granular layer of the nodulus of rats raised in a 2 g environment for 60 days, 80% of the glomeruli showed altered synaptic morphology, including changes in the density of pre- and postsynaptic membranes, increased thickness of the postsynaptic density, enlargement of the synaptic cleft, increased packing density of synaptic vesicles, enlarged mitochondria, and an increase in the number of microtubules (Krasnov and Dyachkova, 1986; Krasnov, 1991). Two days after return to a 1 g environment, the ultrastructure of the nodulus resembles that of control animals. The synaptic vesicle packing density is decreased, and the number of microtubules is diminished, suggesting reversibility of the gravity-induced effects. Taken together with the microgravity findings, these results more generally indicate that the central vestibular system responds to major changes in the gravitational stimulus with similar morphological restructuring, regardless of the direction (hypo- or hyper-) of that change.

In light of the ultrastructural findings suggesting that excitotoxicity may be a factor in early neuronal responses to altered gravitational environments (Holstein et al., 1999), the recent study of glutamate receptor expression is of particular interest. This study examined glutamate receptor mRNA expression using RT-PCR on cells in the vestibular nuclei and vestibulo-cerebellum of rats exposed to hypergravity for 2 h to 7 days (Uno et al., 2002). The results indicate that mRNA expression of GluR2 (an AMPA glutamate receptor subunit) and NR1 (the obligatory subunit of the NMDA glutamate receptor) in the nodulus/uvula and NR1 expression in the medial vestibular nucleus increase after 2 h of stimulation. This expression gradually returns to baseline during the 7 days of hypergravity exposure. As noted above, mRNA expression of GluR2 receptors in vestibular ganglion cells is reduced after 7 days of stimulation. Neither the mGluR1 metabotropic receptor nor the δ₂ glutamate receptor in the flocculus and nodulus/uvula is affected by hypergravity exposure for 2 h to 7 days. It was suggested that the immediate (2 h) adaptation to hypergravity involves enhanced cerebellar inhibition of the vestibular nuclei mediated by Purkinje cell NR1 and GluR2 receptors, whereas longer term adaptation involves decreased transmission from vestibular hair cells to primary afferent neurites mediated by down-regulation of postsynaptic GluR2 receptors on the primary afferents. One interesting aspect of this study is that functional NMDA receptors are not normally present on cerebellar Purkinje cells in adult mammals. Conceivably, an altered gravitational environment presents a sufficient stimulus to trigger their expression and activity. However, the activity of enzymes involved in energy metabolism (lactate dehydrogenase and creatinine kinase) in giant Deiters’ neurons of the lateral vestibular nucleus and cells of the cerebellar nodulus in rats are not appreciably affected by 22 days of space flight (Krasnov, 1975).

Space flight-related studies utilizing markers for vestibular neuronal activity have concentrated primarily on the immediate early gene c-fos. The gene encodes for Fos protein, which reaches peak values within 2–4 h of the effective stimulus and returns to baseline within 6–8 h. Fos-related antigen (FRA) proteins, which are generated by multiple genes, are induced shortly after stimulation, but persist for days (“acute” FRAs such as FRA-1, −2, FosB, ΔFosB) or weeks (“chronic” FRAs, e.g., modified forms of ΔFosB). Fos protein is often utilized as a neural activity marker, since it can be identified with higher spatial resolution than metabolic indicators. Although the basal expression level of c-fos expression is low in the brain, increases in Fos-like immunoreactivity have been reported in vestibular structures following galvanic stimulation (Kaufman and Perachio, 1994; Mashburn et al., 1997) and centripetal acceleration (Kaufman et al., 1992), as well as space flight.

Results from space flight experiments on adult rat brain tissue indicate a trend toward increased numbers of Fos-immunopositive cells in the vestibular brainstem (particularly the medial and descending vestibular nuclei, MVN and DVN, respectively) 24 h postlaunch, and a statistically significant increase in the number of immunostained cells 24 h after return from a 17-day mission (Pompeiano et al., 2002). The number of Fos-immunostained vestibular cells were equivalent in flight animals and ground controls by 13 days postlaunch and at 13 days postlanding. The pattern of FRA protein immunolabeling was qualitatively similar to that of Fos, except at 1 day after landing, when FRA-immunolabeled cells were observed throughout the entire DVN, but only in the caudal MVN while Fos staining was reported throughout the entire MVN. In addition, Fos- and FRA-like immunoreactivity in the vestibular portions of the inferior olivary complex of rats was unchanged 24 h postlaunch (d’Ascanio et al., 2003). However, while Fos immunolabeling remained unchanged 24 h postlanding in these regions of the inferior olive, increased FRA-immunostaining was reported at that time. The authors attribute these findings to the mixed sensory signals derived from rapid fluctuations in g-force during launch. In contrast, the flight rats sacrificed in microgravity 1 or 13 days postlaunch had fewer Fos- and FRA-immunolabeled efferent vestibular neurons than ground controls (Balaban et al., 2002), although no differences were observed between flight and control rats during the postlanding re-adaptation period. The results are interpreted as indicative of general physiological and morphological changes in the cells. Interestingly, Fos- and FRA-like immunoreactivity in autonomic regions such as area postrema and nucleus tractus solitarius of these flight rats were equivalent to control tissue during exposure to microgravity, but were significantly increased 24 h after landing (Pompeiano et al., 2004).

Upregulation of genomic activity has also been demonstrated in several brainstem nuclei related to otolith pathways, particularly the dorsomedial cell column of the inferior olivary complex as well as MVN, DVN and the y-group, in head-fixed rodents exposed to 2 g centripetal acceleration (Kaufman et al., 1992). This upregulation occurs following one-axis stimulation restricted to the plane of the saccule (Mashburn et al., 1997) or the utricle (Kaufman et al., 1991). Immunohistochemistry for fos protein and for FRA proteins has also been performed on vestibular tissue from 60-day-old rats exposed to 2 g or 4 g centrifugation in the plane of the saccule (hypergravity exposure), and in 60-day-old rats born and housed at 2 g, then exposed for 90 min to 1 g (hypogravity condition) (Duflo et al., 2000). This study found enhanced Fos-related labeling of the vestibular brainstem, particularly MVN, DVN, nucleus of Roller, the y-group and the inferior olivary nucleus, only in the hypergravity condition. Although the β subnucleus of the inferior olive was not immunostained in this study, nor in Mashburn et al.’s study (1997), it did display Fos-like immunoreactivity in the experiments of Kaufman and colleagues (Kaufman et al., 1992, 1993), suggesting that utricular rather than saccular inputs activate β-subnucleus Fos expression.

Summary

It is clear that cellular and subcellular constituents of the vestibular pathways are modified in response to altered gravitational stimuli. In the vestibular periphery, opposite structural effects appear to result from hypo- and hypergravity stimulation. In the central vestibular system, neural circuits may well exhibit apparently identical structural changes in response to diverse hypo- and hyper-gravity stimuli, reflecting the more dynamic, highly regulated interactions of the central pathways. More research will be needed to resolve the inconsistencies in the current published literature.

SECTION 2: STUDIES OF THE EFFECTS OF MICROGRAVITY ON VESTIBULAR AND OCULOMOTOR FUNCTION IN THE RUSSIAN COSMOS PROJECT

Abstract

Experiments were performed while monkeys flew in space in the “Cosmos/Bion” Missions to determine the effect of microgravity on the oculomotor and vestibular systems. Eye-head coordination during gaze shifts to lateral targets (gaze fixation reaction, GFR) and multiunit activity in the medial vestibular nuclei (MVN) and cerebellar flocculus were studied in rhesus monkeys in the Bion 6 (Cosmos 1514) through Bion 11 projects. In the first few days of space flight, gaze displacement onto lateral targets became hypermetric, and the amplitude of head movements decreased. This was compensated for by increases in the gain of the angular vestibulo-ocular reflex (aVOR) that could last for the duration of the missions. Associated with this, there were increases in neuronal activity in MVN and flocculus. Sensitivities of the same populations of MVN neurons to linear acceleration, in general, increased gradually over the first 5–7 days in microgravity and then normalized over the course of the flight. These data indicate that the gain of the aVOR is increased during active lateral gaze fixations in space flight, and show that the underlying neural activity is appropriate to produce these changes.

Introduction

The primate research program “Bion” on the biosatellite “Cosmos” was planned in the 1970s to investigate vestibular dysfunction in space, with the aim that the outcome would benefit humans traveling in space. At that time the Space Adaptation Syndrome (SAS) was observed in 30–40% of cosmonauts, but there was no possibility of obtaining direct measurements of parameters related to vestibular, proprioceptive, motor, or other dysfunction during piloted flight, and most of the information on vestibulo-oculomotor dysfunction related to microgravity was obtained during postflight testing (Uganov, 1974). Moreover, changes observed after landing could be related not only to the effects of weightlessness but also to the factors that cosmonauts experience during landing (see Section 1). To obtain direct measurements of the vestibular dysfunction in microgravity, the space capsule “Vostok,” which was originally designed for single-occupancy Cosmonaut flight and which had been used for six orbital flights, was adapted to handle two primate capsules (Gazenko and Ilyin, 1987). In this review we will refer to these experiments as Cosmos or Bion flights interchangeably.

There were three main vestibular studies in the Cosmos project. One was to study eye–head coordination and activity in the vestibular nerve, the medial vestibular nuclei (MVN) and the cerebellar flocculus associated with angular head movements in the horizontal plane during gaze shifts performed by the head and eyes to lateral targets, the gaze fixation reaction (GFR). Another project studied the sensitivity of central vestibular neurons to linear head displacements along the body axis. The third set of experiments studied the activity of leg flexor and extensor muscles during foot movements at different times of adaptation to microgravity. In the present review we will only cover the first two projects since the last project has been described in detail elsewhere (Edgerton et al., 2000).

The results presented here are based on information from many sources (Shipov et al., 1986; Sirota et al., 1987, 1988a, 1989a,b, 1990a,b,c, 1991b,c; Kozlovskaya et al., 1989, 1991, 1994; Yakushin et al., 1989, 1990, 1992. However, the amount of data presented in these publications is limited, and the majority of quantitative data was taken from “Final Reports” that were submitted by investigators to the Institute of Biomedical Problem officials at the end of each project (Sirota et al., 1984, 1986, 1988b, 1991a; Badakva et al., 1993).

Methods

Male monkeys (Macaca mulatta) of 3–5 kg were used in these studies. Their names were assigned alphabetically; the first letter of the monkey’s name corresponded to the sequential letters of Cyrillic alphabet (see Table 1). Although two primates traveled in each flight, information on vestibulo-oculomotor coordination and unit activity was not always available from both of them. Thus, although twelve animals took part in six Bion flights, valid data were obtained from only seven animals during four space flights of different durations. These results form the basis for the present report.

Table 1.

Flight numbers, times of launch, and monkeys in Cosmos flights

Cosmos flight number and date	Taking off time	Project number launch date	Monkey name	Flight duration (days)
1514	7:00 GMT	BION 6	Abrek	5
12.14.83	10:00 Moscow		Bion
1667	3:21 GMT	BION 7	Vernyi	7
07.10.85	6:21 Moscow		Gordyi
1887	12:43 GMT	BION 8	Drema	13
09.29.87	15:43 Moscow		Yerosha
2044	6:28 GMT	BION 9	Jakonia #782	14
09.15.89	9:28 Moscow		Zabiaka #2483
2229	13:40 GMT	BION 10	Ivasha #6151	12
12.29.92	16:40 Moscow		Krosha #7906
^*	13:50 GMT	BION 11	Lalik #484	15
12.24.96	16:50 Moscow		Multik #357

Open in a new tab

^*

At this point, the Russian Space Agency stopped assigning flight numbers to the “Cosmos” Missions.

Eye–head coordination test

The gaze fixation reaction (GFR), which is a shift of gaze onto lateral targets using both head and eye movements, is a structural unit of daily operant behavior, and the dynamic characteristics of the GFR are similar in men and monkeys (Bizzi et al., 1971, 1972). Small gaze shifts are usually performed first by an eye saccade and later by head movements. During gaze shifts larger than 20°, eye saccades are always accompanied by head movements (Tomlinson and Bahra, 1986; Phillips et al., 1995). The test was structured so that the monkey first directed its gaze toward a fixation light in front in the primary position. The target was then repositioned laterally. In response, the animal first made an eye saccade toward the target at its new location. Generally, about 20–40 ms after the beginning of the eye saccade the head started to move toward the target. Since the eye saccade was much faster than the head movement, the eye was directed toward the target first. Thus, the head was still in motion at the end of the saccade, and the head performed only about 20–30% of the total motion.

Experimental conditions had an effect on GFR parameters: if the appearance of the lateral target was predictable, the head could start an anticipatory movement first, while the animal was still fixating the centrally located target. In other cases the head could delay the eye saccade significantly (Dichgans et al., 1973; Grigaryan et al., 1986). Thus, when the gaze jumped from the central position to a visual target located 40° laterally, the total gaze displacement was 40°. However, the contribution of the eyes and the head varied from trial to trial (Phillips et al., 1995). To keep the gaze stationary on the new target, the eyes counter-rotated to compensate for the head motion (Bizzi et al., 1972). As demonstrated in monkeys with a bilateral loss of vestibular function, the counter-rotation of the eyes after the saccade could also be anticipatory (Dichgans et al., 1973). In normal animals, however, the counter-rotation of the eyes was due to activation of the angular vestibulo-ocular reflex (aVOR) (Dichgans et al., 1973). Visual feedback after the saccade onto a newly located target also affected the ocular counterrotation and provided the source for the adaptive modification of the aVOR gain. The major focus of the research was to determine the gain (eye velocity/head velocity) of the aVOR during these active gaze shifts onto target.

Although gaze shifts from one point to another are based on visual information, visual correction can occur only after the gaze has shifted to a new position. Moreover, since visual recognition has some delay, there was no visual feedback within about 80 ms immediately after the eye saccade. Therefore, the counter-rotation of the eyes reflected the current state of the aVOR gain. Since parameters of the GFR, such as amplitude and velocity of the eye, head and gaze movements, as well as counter-rotation of the eyes after a saccade, are determined by and reflect changes in sensitivity of the vestibular and proprioceptive sensors or brainstem structures, the GFR was considered a suitable reaction to study the effects of microgravity on vestibulo-oculomotor coordination.

Experimental paradigm utilized in Cosmos experiments

Gaze fixation reaction

Monkeys were trained to look at a central target and to position their heads straight ahead in the horizontal plane for 0.8 s. Head position was used as a feedback signal to trigger the next step. Targets located ±40° laterally from the center were presented for 1 s. The lateral targets were of two configurations. When the stimulus, a “C”, was presented, the animals were required to press on the lever located in front, beneath the target plate. If the animal pressed the lever within 1 s from the time the lateral conditional target appeared, it got a reward of 0.3 ml of juice. If the animal did not press the lever within the required time, or if it pressed the lever in response to the stimulus, which was an “E”, no juice was given. Additionally, the presentation of the next central target was delayed by 7–10 s. The appearance of targets on the left and right as well as the presentation of stimuli were randomized. Each program presentation was comprised of a set of 256 stimuli with 2/3 of them being conditional. This test was performed twice a day. In general, this test was accomplished within 20–25 m but, regardless of the animals’ performance, only the first 20 min of the morning session were stored on tape for analysis (Sirota et al., 1984, 1986, 1988b, 1991a).

MVN unit responses to linear acceleration

The response of MVN neurons to vertical linear acceleration was studied in each of the Cosmos 1667 to Cosmos 2229 flights. For this, the primate chair was elevated 45 mm slowly and then dropped suddenly to its original position with the aid of a spring mechanism. The total motion of the chair was 45 mm in all flights, but there was variation in the stimulus between flights. Most flight stimuli were comprised of a slow chair elevation over 8 s and a fast drop down to the original position. Peak acceleration for the upward motion was 0.14×10⁻⁴ g. This is close to the threshold for detection of linear acceleration by the vestibular system, which is ≈0.1×10⁻³ g for humans (Guedry, 1974). Thus, any responses observed during the elevation phase could have been due to random fluctuation or other factors. The stimuli parameters for the motion down in the drop varied on Earth and in Space. On Earth, the chair moved down over 0.6 s, and the downward pull of the spring was aided by the pull of gravity, to provide a downward acceleration of 25.4×10⁻³ g. In Space the chair moved over 0.9 s, to provide a downward acceleration of 11.0×10⁻³ g.

Surgical procedure

Two surgical approaches were used in the Cosmos experiments to implant the head holders. In the traditional method, used in Cosmos 1514, two bolts were implanted on the skull to fixate the head mount. A new technique that was minimally invasive was developed for later flights (Sirota et al., 1988a), and is described in full detail elsewhere (Yakushin et al., 2000b). EOG electrodes were implanted bitemporally to record the horizontal component of eye movements as well as above and below the left eye to record vertical eye movements.

Identification of the brain structures

Neurons recorded in the vestibular nuclei had various types of activation but most of the isolated neurons were modulated in phase with head velocity and were activated by eye position or eye velocity (Miles, 1974; Fuchs and Kimm, 1975; Keller and Daniels, 1975; Chubb et al., 1984). Units recorded in the flocculus were similar to floccular units in previous studies (Lisberger and Fuchs, 1978a,b). Some of the recorded floccular cells had complex spikes that confirmed electrode locations near Purkinje cells in the cerebellar cortex. Attempts were also made to record units in the vestibular nerves in the Bion 8–11 projects, and some preliminary data are available (Kozlovskaya et al., 1989, 1991; Correia 1998). Since there was no way to confirm that the recordings were taken from the same population of fibers on different flight days, these observations are omitted from this review.

Recorded signals and calibration

Head position in the horizontal plane was recorded with a special device placed on the center of the head-mount in the Bion 6–10 flights. The device was based on a compass principle. DC magnets were placed on either side of the primate chair at the level of the top of the head, where the device was mounted. The head position sensor was calibrated before flight. The horizontal EOG was calibrated using two assumptions based on the preflight performance: first, it was assumed that the final gaze position was equal to the angular position of the lateral target; second, under normal, preflight conditions, it was assumed that the counter-rotation of the eyes was fully compensatory during the GFR, and, therefore, that gaze (eye + head) position was stable after the eye saccade until the end of the head movement. Two channels capable of recording neural activity at frequencies ranging from 0.2 to 5 kHz were used in the Cosmos 1514. Two more channels with a frequency range from 0.2 to 10 kHz were used for unit recording in Cosmos 1667 flights, while four channels of this frequency band were used in later flights.

Electrode placements into MVN and the flocculus were based on the physiological responses of the identified units. The microelectrodes were introduced through metal guide-tubes implanted in the skull through 1 mm holes drilled through the scalp and skull in stereotaxic coordinates. Coordinates to reach MVN were P2–P4, lateral 2 mm. The flocculus was approached at the level of lateral vestibular nuclei, and the microelectrodes were tilted laterally so that they would not enter the brainstem after penetrating the flocculus. Three to four guide tubes targeted each structure. The microelectrodes were made of 80 μ tungsten wires covered with epoxy.

Data analysis

The methods of data processing varied with improving technology. Records obtained in the Cosmos 1514 flight were printed on a chart recorder. Since the amplitude and velocity of many parameters had relative calibrations, it was impossible to determine delays and durations of various GFR parameters precisely. Therefore, the results from the Cosmos 1514 flight are expressed as a percent relative to the preflight values. In all other flights, analog tapes recorded during flight were digitized with eight-bit resolution with a Motorola 6800 general-purpose mainframe computer. The channel that contained the marker of central and peripheral target presentation was digitized at 5 kHz. When the presentation of a lateral target was detected, analog signals, including the horizontal EOG, head position, lever-press and marker channel were digitized at the same frequency. Data were averaged over eight data points and stored in the computer at a 1.6 ms sampling rate for future analysis. Thus, although the data were stored at 625 Hz, each data point represented an average value of eight sequential data points and, therefore, the noise level due to the digitizing process was reduced. The following parameters of GFR were analyzed: latency, duration, amplitude and peak velocity of the eye, and head and gaze movements. There were no consistent changes in latencies and they will not be considered further.

Results: Gaze fixation on lateral targets; gaze fixation reaction (GFR)

Hypermetria of gaze

A typical saccadic gaze shift to the lateral target in preflight testing before the Cosmos 1514 flight is shown in Fig. 1A. A head movement in the same direction accompanied the eye movement. In the first inflight recording on the second day of space flight, the amplitude of the saccades was larger than before the flight, and the amplitude of head movement was approximately the same. Therefore, the gaze shift was hypermetric (Fig. 1B). Presumably, since the reward was dependent on accurate fixation of the target within one second, the animal overcame the hypermetric gaze shift after the initial period, as in Fig. 1F. Overall, the amplitude of saccades during gaze shifts performed with a single saccade in this monkey increased during flight. The increase was 11 ± 3% on day 2 and gradually increased during the flight, reaching a maximum of 42 ± 7% on day 5. Concurrently, the amplitude of head movement was smaller in flight, decreasing by 34% on flight day 2, and the head movements were still 27% lower than before flight on day 3. The head movements then normalized, and were only about 8% smaller on flight days 4 and 5. This animal performed the lateral gaze shift with one saccade before flight, using multiple saccades only 4% of time (Fig. 1C). The multisaccadic gaze shifts increased to 36% on the second day of space flight. The number of gaze shifts made with multiple saccades was highest on day 2 and then decreased to 14% on flight days 4 and 5 (Fig. 1C).

Parameters of the GFR were similar in the monkey studied during the Cosmos 1887 flight. The gaze amplitude, which was 40.2 ± 6.9° before flight, became hypermetric in space, going from 46.0 ± 8.5° on the first day, to 58.3 ± 7.0° on the 5th day. The gaze amplitude was slightly smaller (51–52°) when tested on days 8 and 10. Examples of these gaze shifts are shown in Fig. 1D–F. Before flight, gaze was stable on the lateral target over the entire period of head movement (Fig. 1D, dashed line). In flight, gaze was hypermetric, mostly due to the increased amplitude of the saccades, but gaze came back onto the target after the initial overshoots (Fig. 1E, F, dashed lines).

In Cosmos 2044, the gaze movements to the 40° lateral targets before flight were 38.0 ± 0.4° and 37.0 ± 0.6° in the two monkeys (Fig. 2). On day 2 in space, the gaze amplitudes were 47.0 ± 1.3° in the monkey Jakonia (Fig. 2A) and 48.0 ± 1.9 in monkey Zabiaka (Fig. 2B). The gaze amplitudes normalized, however, within the next several days in both animals. Detailed analysis revealed substantial differences in the source of observed changes. Although both animals had hypermetric gaze on day 2, the gaze overshot the target due to larger saccades in Jakonia, while the overshoot was due to larger head movements in Zabiaka.

In the Cosmos 1667 and 2229 flights, most gaze shifts were associated with multiple saccades. The amplitude of the associated head movements was decreased (Gazenko and Ilyin, 1987; Sirota et al., 1987). Multiple saccades were not characterized quantitatively, but the gain of the compensatory aVOR was increased (see below). Thus, similar to the other flights, the lateral gaze displacements during the Cosmos 1667 and 2229 flights were also hypermetric.

In summary, gaze shifts to lateral targets in each of the six animals became hypermetric in the first few days of flight. This was observed whether the monkeys made the gaze shift in single or multiple saccades. In five of the six monkeys, the gaze overshoot was mainly due to an increase in the amplitude of the saccades, while in one animal it was due to an increase in the amplitude of the head movements (Fig. 2B). Thus, for successful visual fixation of target during head movements, the hypermetria had to be compensated for either by a corrective saccade or by a change in the gain of the aVOR (Fig. 1F).

Increase in the gain of the angular VOR (aVOR)

The gain of the aVOR was defined as the ratio of eye velocity to head velocity during counter-rotation of the eyes after the animals had made a saccade onto target. In Cosmos 2044, the instantaneous aVOR gain was studied as a function of time, starting from the end of the eye saccade over a 128 ms time period. When the eyes were stationary just after the saccade, the aVOR gain was zero. As counter-rotation began, the gain gradually rose. Under normal conditions the aVOR gain is 1.0, so the eye velocity is equal to head velocity over the counter-rotation period.

Nine superimposed gain curves obtained before flight in monkey Zabiaka are shown in Fig. 3A. The aVOR gain increased to unity within the first 30 ms after the saccade and then remained stable while the head was in motion. In most cases, the head movements ceased after 80 ms (Fig. 3E). Consequently, the aVOR gains were calculated as average values over the time interval from 32–64 ms after the initial saccade. Average instantaneous gain curves based on 30 responses before flight (Fig. 3B) were similar to those in Fig. 3A.

In flight there were substantial changes in aVOR gains. As shown in Fig. 3C, the gain of the aVOR increased to about 1.5 on the first recording of day 2 (Fig. 3C) and remained at this level until day 8 (Fig. 3D), when the increase became even larger (≈2.0). The gain was still about 1.5 on day 14 (Fig. 3E). Changes in the gains of the aVOR were similar during gaze shifts in either direction and were combined to obtain the average gain changes for this monkey (Fig. 3F, filled circles). The aVOR gain was also increased when the other animal of this flight, Jakonia, was tested for the first time in space on day 3 (Fig. 3F, filled squares). The gain increase was smaller subsequently, but was present over the entire flight.

The average gain values obtained from the instantaneous gain curves were compared to the aVOR gains obtained from taking the ratio of eye and head velocities at arbitrary points during the counter-rotation. The average aVOR gains were the same when measured with either method (Sirota et al., 1991b). Thus, it was possible to compare the aVOR gain measurements in the different Cosmos projects. Individual gain curves for each of the monkeys in this report are shown in Fig. 3F. There were individual differences, but as a group, there was a substantial increase in the aVOR gains that persisted over the entire flight (Fig. 3G).

In summary, the general conclusion is that gaze became hypermetric upon entry into space on each of the Cosmos flights and that the positional errors generated by this gaze overshoot were compensated for by an increased gain of the aVOR. Although both of these changes could have occurred independently, since they occurred in parallel, it is more likely that one was primary, while the other was an adaptive response to the primary change.

The angular acceleration that activates the semicircular canals is the same on Earth and in space. Accordingly, the gain of the passive aVOR induced by steps of velocity and by voluntary sinusoidal head oscillation in darkness were not affected in microgravity (Benson et al., 1986; Cohen et al., 1992; Clarke et al., 2000). If the hypermetric gaze was primarily due to exposure to microgravity, then the alteration in aVOR gain was a simple compensation for the positional error, driven by the visual feedback. However, there are reasons to question this explanation. First, the source of the hypermetric gaze varied among animals. In one, it was due to an increase in the amplitude of head movements (Fig. 2B), while in the others, it was due to the increase in the amplitude of the saccades. Of interest is that similar increases in gaze amplitude have been observed in both human and monkeys after water immersion (Barmin et al., 1983; Kreidich et al., 1983; Badakva et al., 2003). In monkeys the hypermetric gaze was due to an increase in saccade amplitude (Badakva et al., 2003), while in humans the amplitude of the head movements increased (Barmin et al., 1983; Kreidich et al., 1983). There was no change in the acceleration of gravity in these experiments, yet the changes were similar to those during space flight.

An alternate explanation is that the increase in the gain of the aVOR gain was the primary event. Similar changes in aVOR gain were also observed after water immersion (Barmin et al., 1983; Kreidich et al., 1983; Badakva et al., 2003), and as in the Cosmos 1667 flight, the first gain changes were observed after only two hours of immersion. Since changes in gravity were not a factor in these experiments, changes in proprioceptive inputs could have contributed to or been responsible for producing the gain changes. If the aVOR gain changes were the primary even, then parameters of the GFR should be different if the eye saccade was not accompanied by head motion. Kozlovskaya and colleagues demonstrated in cosmonauts tested before and after space flight that there are two types of the adaptive changes in GFR. One group had changes in the amplitude of the saccades and in aVOR gains, similar to those observed in the monkeys. The second group was comprised of spacecraft commanders, e.g., professional pilots. They performed gaze shifts in two stages: first by a saccade onto target and then by a head movement, which followed the saccade. In the pilots, changes in their aVOR gains were similar to those described above. Since the amplitude of the saccades during the gaze shift was accurate within ±3°, there was no visual/vestibular mismatch that could have driven the changes in the aVOR gains (Kozlovskaya et al., 1985).

An alternate mechanism could also have been responsible for the changes in the gain of the aVOR. It was recently demonstrated that adaptive changes of the aVOR gain are a function of head position with regard to gravity (Tiliket et al., 1993; Yakushin et al., 2000a, 2003a,b, 2005). Thus, “normal” aVOR gains during active head movements may only exist under the level of gravity and/or proprioception in which the aVOR gains were adapted, and a change in gain of the aVOR due to insertion into microgravity could have been the primary change that produced the changes that were observed in microgravity.

If there were changes in the gain of the aVOR, there should be associated changes in the cellular activity in the vestibular nuclei and/or flocculus, which produced the adaptive changes in aVOR gain ((Lisberger, 1994; Lisberger et al., 1994a,b), see Ito (1984) for review). Unit activity was recorded in the vestibular nuclei and flocculus during each space flight and is considered in the next section.

Neural activity in the vestibular nuclei and flocculus during space flight

Activity of a single lateral canal-related neuron in the right MVN of monkey Vernyi (COSMOS 1667) during sinusoidal head/body oscillation before flight is shown in Fig. 4A. The discharge rate increased in phase with ipsilateral head velocity (Fig. 4B). Therefore, this cell was a type I unit.¹ When this unit was well isolated from the background, the electrode was fixed at this location. Over time, this electrode recorded several neurons simultaneously. The multiunit activity recorded before and after the flight increased with ipsilateral rotation and had a similar phase relationship to head velocity (Fig. 4D) as the neuron in Fig. 4A. The electrodes over the various Cosmos flights were implanted in MVN in the region of type I units.

Fig. 4. — Activity of the lateral canal-related type I neuron recorded the right medial vestibular nuclei in the monkey Vernyi (Cosmos 1667). A, Single unit recording at the site where one of the MVN electrodes was installed. Head position down is rotation to the right (ipsilateral). B, Poststimulus histogram for the unit shown in A. Traces below show head position and head velocity during sinusoidal rotation. C–D, Integrated unit activity recorded from the same electrode before (C) and after (D) seven days of the space flight. Note that the modulation in unit activity increased peaking just before the ipsilateral (rightward) head velocity had reached maximum. The head oscillation shown in A–D was driven manually with approximate frequencies 0.45 Hz (A, B), 0.5 Hz (C) and 0.4 Hz (D).

Multiunit activity recorded from three of four of the implanted electrodes in the two animals in Cosmos 1667 was of good quality. An example of a recording during a gaze shift to the right during flight is shown in Fig. 5A. Single neurons, shown as standard pulses, could be separated from the background (Fig. 5B). The quality of separation was similar for each flight day, although it was not certain that the same neurons were analyzed throughout the flight. Before flight, the peak activation occurred approximately at the time when ipsilateral head velocity reached a maximum (Fig. 5C, Before). The unit activity did not decrease relative to the resting discharge rate (white dashed line) in association with contralateral head velocity. The profile of activation by ipsilateral head velocity was the same over all flight days. The amplitude of peak activation before flight (27±4 imp s⁻¹) was comparable to the peak activation recorded during the second hour of space flight (23±6 imp s⁻¹, day 1). This activation became larger on flight days 2–4 (50±8 imp s⁻¹, day 2) and then decreased to normal on days 5–6 (25±3 and 22±3 imp s⁻¹, respectively). Thus, the activation was twice the normal value after 26 h in space (day 2). The changes were substantial over the next 2 days and then normalized. Data from this neuron are summarized in Fig. 6F (filled symbols). There was no significant change in the activity recorded during gaze shifts to the right after 2 h of space flight (day 1), but the activity gradually increased to a maximum on day 4, and returned back to normal for the rest of the flight.

Fig. 5. — A, Sample of multiunit activity recorded in the right vestibular nuclei in monkey Vernyi (Cosmos 1667) during the gaze fixation reaction onto the target on the right. The recording was obtained from the same electrode at the same location as in Fig. 4. B, Activity of the largest neuron from a sample similar to that shown in A and converted to standard pulses. C, Poststimulus histograms of this unit recorded at different times in flight. Thirty head movements of similar profile and amplitude were identified for each day. Head motion to the right and back to the center was divided into eight bins. An additional eight bins were used to characterize the data from before and after head motion. The idealized head position and head velocity are shown under each histogram.

The profile of the integrated multiunit activity was similar when multiunit activity recorded by the same electrode was integrated (Fig. 6A–C). That is, activation before flight (Fig. 6A) was similar to that on day 7 (Fig. 6C), and the increase in unit activity was larger on day 2 (Fig. 6B). The integrated activity from the multiunit recordings over the entire flight is shown in Fig. 6D (open symbols). Activation on flight day 1 after 2 h was the same as the preflight activity. This activity significantly increased on flight day 2 and then returned to normal by days 5–6 (Fig. 6F). The difference between the single and integrated multiunit recordings was presumably due to variation in response of the neurons in the vicinity of the recording electrode.

The multiunit activity in MVN was also studied in four other animals. The two animals in Cosmos 1887 had two electrodes implanted in MVN (Fig. 6D) and the two animals in Cosmos 2044 flight (Fig. 6E) each had one electrode in MVN. The sensitivities recorded in Cosmos 1887 flight were maximal on day 3 for all four electrodes and then decreased to normal in three of the four electrodes (Fig. 6E, filled circles). Changes in the sensitivities of the MVN neurons in the monkeys of Cosmos 2044 flight were more robust. The sensitivities increased significantly on day 3 and stayed high until day 8 when they began to normalize. In the animal Jakonia, the sensitivity stayed above the normal even on day 13 (Fig. 6F, filled symbols).

Multiunit activity was recorded in the flocculus from one electrode in each of the same four animals. In all cases the change in the sensitivity was maximal on the first day of flight (Fig. 6G, H). In two animals, the activity normalized by the middle of the flight (filled symbols), while in other two, it stayed above the norm until the last day of flight (Fig. 6G, H, open symbols).

A summary of all multineuronal recordings from the Cosmos flights is shown in Fig. 6I. Overall the sensitivity of the neural activity in both the vestibular nuclei and flocculus was maximal in the earlier days of flight, with the rise in activity occurring earlier in the flocculus than the MVN neurons. There was a striking similarity between the rise in activity of the flocculus (Fig. 6I, dashed line) and MVN multineuronal activity (Fig. 6I, solid line) and the increases in aVOR sensitivity shown in Fig. 3G. The flocculus activity tended to rise earlier and persist longer than the MVN activity, but both were increased over the two weeks of flight. Thus, the changes in neural activity in MVN and flocculus during space flight mirrored the changes in aVOR sensitivity during the gaze fixation reaction. Presumably, the flocculus units were involved in the learning that changed the gain of the aVOR, while the increase in unit sensitivity in MVN to head rotation reflected the actual changes in gain in the compensatory processes.

Response of MVN units to linear acceleration in space

The first inflight recordings of central vestibular neurons during otolith stimulation in the Cosmos 1667 project utilized the same population of neurons in the right MVN of monkey Vernyi that was studied during angular rotation (Figs. 4, 5, 6A–C, F). A sample multiunit activity during otolith testing on day 1 is shown in Fig. 7A, and associated single unit activity in Fig. 7B. The stimulus cycle was divided into four phases for analysis. The spontaneous discharge rate (Phase 1) was obtained from a 1 s time period before each elevation of the chair. Unit activity during the elevation (Phase 2) and rapid drop (Drop, Phase 3) were taken as average values.² Additionally, activity was also utilized over a 1 s period immediately after the drop had terminated (Stop, Phase 4). Data were first expressed in imp s⁻¹ and then normalized to activity in Phase 1 of each day. The bottom traces (Chair) in Figs. 7A, B, show the time intervals for each phase.

The resulting histograms for each flight day are shown in Fig. 7C. The first significant changes in the drop phase occurred on the first day after 2 hours of space flight. This activity normalized on days 2 and 3, but then increased again, reaching a much higher level (Fig. 7C, D). Sensitivity of the activity for the Drop in Phase 3 for each day is shown in Fig. 7D. The first significant changes were observed in the first day of flight. Activity was near normal on flight days 2 and 3 and then increased again on days 4–5. The direction of linear acceleration was opposite during the Drop and the Stop phases, but there was no difference between the two responses on any flight day. Since the head was not fixed during the Stop, it is not known whether the head continued to pitch at the time of the Stop, which could explain this apparent disparity. Therefore, the activity associated with the Stop Phase will not be considered further in the analysis.

Integrated multiunit activity from the second animal in the same flight (Gordyi) is shown in Fig. 8A–C. Before the flight, there was a brief increase in activity during the Drop (Fig. 8A). On the second day of flight there was a significant increase in activity during the Drop (Fig. 8B). The activity increased in response to the acceleration phase and then decreased to normal during the chair deceleration. The activity during the entire period on day 7 was more irregular than the other days (Fig. 8C), and the response to the Drop was again brief. Integrated activity recorded from both animals in the Cosmos 1667 flight is shown in Fig. 8D. The sensitivity of these MVN units was not changed on day 1, but was increased in two of the three electrodes on day 2. The activity increased further up to day 5, and then normalized.

Changes in the sensitivities of the MVN neurons in the monkeys Jakonia (Fig. 8E, open circles) and Zabiaka (Fig. 8E, open squares) from Cosmos 2044 were similar to those shown in Fig. 8D. The sensitivities gradually increased reaching maximal values on day 6 and then declined, although they were still above the baseline on the last day of flight (day 13). Thus, changes in the sensitivity of the MVN neurons to vertical linear acceleration mediated by the otoliths were similar in animals from the Cosmos 1667 and Cosmos 2044 flights. The sensitivity to linear acceleration gradually increased in space, reaching a maximum at days 5–6 and then decreasing toward normal values. The peak increases in sensitivity for the MVN neurons in Fig. 6D, E in response to angular acceleration during the gaze fixation reaction were between 300 and 600%, whereas the increases in sensitivity during linear acceleration were smaller, within 200–250%. Thus, the general pattern of response was similar to that obtained from angular acceleration, but the magnitude of the response to linear acceleration was smaller. This could represent a difference in response to angular and linear accelerations or could reflect a difference in the magnitude of convergence of the otolith and canal signals onto these neurons.

The responses to linear acceleration produced by sinusoidal oscillation along the body vertical axis in the Bion 11 Project were very similar to those observed in the Bion 9 (Cosmos 2044) flight. The multiunit activity of the neurons recorded in MVN in each animal was in phase with upward acceleration in monkey #484 and with downward acceleration in monkey #357 (Badakva et al., 2000). In both cases, the sensitivity to linear acceleration increased only slightly on the first recording on Day 2, and reached a maximum on Days 4–6. Sensitivities then decreased but were still above normal up to the last day of flight.

Discussion

The data from the Cosmos flights show that the sensitivity of central vestibular neurons in MVN to both angular head rotation and vertical linear motion was affected by microgravity. In general, the sensitivity of these neurons increased in the earlier days in Space and then gradually normalized. In some instances changes were similar for several populations of neurons recorded on both sides of the brainstem from the same animal (Fig. 6D, E). In other cases, although average activity of the recorded neurons had similar changes in sensitivity during adaptation to microgravity (Fig. 6F, open symbols), the changes in sensitivity of a single unit could differ from the changes observed in the remaining population of units. A striking finding was that the observed changes in neuronal sensitivity were qualitatively similar to the changes observed in the aVOR gains and occurred over the same time course. Changes in sensitivity of the floccular neurons to the angular rotation were more uniform than the changes in MVN activity, presumably, because the flocculus was responsible for producing these adaptive changes.

Changes in sensitivity of the neuronal populations in the vestibular nuclei to the otolith stimulation were also relatively uniform and had approximately the same time course as the sensitivity to angular acceleration. The sensitivity to linear acceleration gradually increased, reaching a maximum by the end of the first week in space (Fig. 7D and Fig. 8D, E). This was different from the changes in sensitivity of the same neuronal population to angular rotation, which increased earlier in flight and was normalized by the time the otolith sensitivity had reached a maximum (Fig. 6I). This was somewhat surprising, since it would be expected that microgravity would affect otolith and not canal sensitivity. A majority of direct projections from otolith afferents go to the lateral and descending vestibular nuclei (Büettner-Ennever, 1999), but there are canal-sensitive neurons with some otolith sensitivity in MVN (Markham and Curthoys, 1972), which could serve a different purpose than maintaining balance or supporting the linear VOR. Otolith information is not necessary to induce changes in the angular VOR gain (Crane and Demer, 1999). As recently shown, however, the head position with respect to gravity in which the aVOR gain was adapted is expressed subsequently in changes in aVOR gain in every head position (Yakushin et al., 2000a, 2003a,b, 2005). Therefore, the otolith signals could serve as a gravitational context for adaptation of the aVOR gain in specific head positions, and the changes in otolith sensitivity could be a reflection of changes in this gravitational context.