Neurolab Launches the Decade of the Brain into Space
Sandra Aamodt
Sandra Aamodt is an assistant editor with Nature Neuroscience.
April will see the launch of Neurolab, a set of experiments aboard the
space shuttle that will investigate the effects of zero gravity on the nervous
system.
The space shuttle Columbia, scheduled to launch Neurolab into orbit on
16 April. The mission will last 16 days, and will carry 26 projects designed
investigate the effect of zero gravity on the nervous system. The crew will
perform physiological and behavioural experiments on a wide range of species,
from crickets to rats to human astronauts themselves.
As Nature Neuroscience goes to press, nine astronauts are scheduled
to take rats, mice, fish, snails and crickets into orbit aboard the space shuttle Columbia on 16 April. The mission's goal is to
increase understanding of the mechanisms responsible for neurological and
behavioral changes under microgravity. The shuttle will remain in orbit for
16 days, and six 'mission specialists' (with eight advanced biology or medicine
degrees among them) will carry out a total of 26 separate projects. In addition
to performing the animal experiments, some of the astronauts are subjects
themselves, for experiments on human physiological and behavioral adaptations
to microgravity . The Neurolab mission is NASA's contribution to the Decade
of the Brain, and it results from an unusual collaboration between NASA and
the US National Institutes of Health (NIH). The experiments were reviewed
and selected by NIH, from a total of 175 applications that were submitted
in response to the original announcement of the mission back in 1993. Since
then, NASA, as well as NIH, NSF, the European Space Agency and other national
funding bodies from France, Canada and Japan, have sponsored the several years
of ground-based research that has led to the final flight selection. Some
of the experiments are intended to examine human physiology in space, partly
with a view to preparing the way for longer space flights in the future. Others
address more basic scientific questions; the principle criteria for selection
were scientific merit and feasibility. What follows is a brief account of
three projects that illustrate the range of topics being studied.
Bruce McNaughton and colleagues at the University of Arizona are taking
advantage of the microgravity environment to test a theory of how visual and
vestibular inputs interact to calibrate cognitive maps of spatial location.
In a familiar environment, a rat's location is signaled by ensemble coding
in the place cells of the hippocampus. Each place cell fires most strongly
at a preferred spatial location, and the population response of all place
cells defines the rat's position in a two-dimensional map of the space. This
spatial map is anchored to visual landmarks, as moving the set of landmarks
causes the preferred locations of the place cells to change.
Although the hippocampus does not receive any direct projections from the
vestibular system, behavioral experiments show that vestibular stimuli affect
spatial maps. In complete darkness, rats can return to the nest by a direct
path, even if they have reached their current position by a complicated series
of turns. If the test area (including the nest) is rotated at a rate that
exceeds the threshold for the vestibular system, the rat can still find its
way back. But if the rotation is below the threshold of detection, the rat
behaves as if no movement had occurred, and loses its way . These findings
imply that rats use vestibular inputs to keep track of their location when
visual cues are unavailable.
The source of this vestibular information may be a group of neurons called
head direction or H cells, which are tuned to the orientation of the rat's
head in space. The H cells are found in many regions of the brain, including
the subicular complex, cingulate and parietal cortex, lateral dorsal thalamus
and superior colliculus. Like place cells, H-cell tuning is controlled by
vision in a familiar environment, but H cells will maintain their absolute
direction sense if the transition to an unfamiliar environment is smooth.
If rats are disoriented or visual landmarks are unreliable, H cells adopt
random preferences. Although H cells remain tuned in darkness, their preferences
tend to drift over time.
McNaughton and colleagues have postulated that H-cell preferences do not
drift in the light because active head cells form associations with cells
sensitive to the spatial location of visual stimuli. This process would recalibrate
the H cells to familiar visual landmarks. The hypothesis predicts that after
rotation of the rat's environment, landmarks and all, the visual input should
control H cell tuning in a familiar environment, but the vestibular input
should dominate in an unfamiliar environment. Three experiments in normal
gravity confirmed this prediction.
Another way to demonstrate that stable spatial maps require associations
between H cells and place cells involves recording from both types of cell
in an environment with a single visual landmark, a white card on one wall.
The card determines the preferred locations of place cells − unless
the rat is disoriented every time it enters the environment. In that case,
both H cells and place cells act as they would in the dark. Because disorientation
causes H cells to adopt an arbitrary orientation, the rat presumably perceives
the white card as moving from trial to trial and therefore rejects it as a
visual landmark. This interpretation is strengthened by a control in which
rats initially are exposed to the white card without disorientation. The initial
training allows the card to gain control of the place-cell fields, which persists
even in later trials with disorientation.
The Escher maze used by McNaughton and colleagues to study the effect of
gravity on the brain's representation of space. Rats are trained to walk around
the maze, using the dark velcro strips to enhance their grip. In this three-dimensional
arrangement, three turns lead the rat back to its starting point, whereas
in a flattened maze four would be needed. In zero gravity, however, the vestibular
cues that would normally distinguish the two arrangements are absent; the
experiment will use a multiple recording electrode to test how spatial representations
in the hippocampus are affected by the conflict between vestibular and visual
cues.
The microgravity environment of Neurolab offers a unique opportunity to
find out what happens when both visual cues and vestibular inputs fail to
match the cognitive spatial map. McNaughton's plan is to record from an array
of 30-40 place cells in the hippocampus while the rat performs the "Escher maze" shown in the picture. Rats have been trained to
walk along the maze while hanging onto both edges of the track, to obtain
a brain stimulation reward. Previous experiments indicate that H cells should
perceive only the 90° yaw turns, and not the 90° forward pitches, as changes
of direction. In this three-dimensional maze, though, the rat returns to its
starting position after three (rather than the usual four) turns, suggesting
that H cells will lag by 90° after each full circuit. This will lead to a
false input to the hippocampal place cells, such that the hippocampal spatial
coordinates will not be fully updated as the animal makes its turns; instead,
the map coordinate will indicate that the rat is on third base when it has
in fact already returned to the home plate. This continuing lag as the rat
walks around the maze should prevent the formation of consistent associations
between place cells and visual landmarks. Possible outcomes of the experiment
include complete suppression of place cell activity, as occurs when rats cannot
move, control of place cells by visual landmarks, or dissociation of place
cell firing from the rat's spatial location. Any of these results will increase
our understanding of how cognitive spatial maps are constructed and maintained.
One Neurolab experiment on human subjects will test how gravity is incorporated
into internal cognitive models of external object motion. This project has
been coodinated by the Centre National d'etude Spatiale in France, and the
primary investigators are Alain Berthoz at CNRS-College de France in Paris,
Francesco Lacquaniti at Clinica Santa Lucia in Rome and Joseph McIntyre, who
has a joint appointment at both institutions. Berthoz describes his experiment
as a dynamic version of what the French refer to as "L'expérience du garçon
de café". If a waiter who is carrying a bottle of wine on his tray lifts the
bottle himself, the tray will not move upward in spite of the load release;
but if a client suddenly lifts the bottle unexpectedly, the tray will rise.
The brain of the garçon de café has been able to anticipate (in the case of
active lifting) the consequences of the load release using his internal model
of the mechanics involved. These static representations of the force of gravity
are formed by about one year of age; the microgravity experiments will test
whether the brain has a similar model of the acceleration due to gravity,
and whether these are susceptible to experience-dependent modification in
the zero-gravity environment of the space shuttle.
The task is to use an outstretched hand to catch a tennis ball dropped
with varying initial velocities. The timing of ball arrival theoretically
could be computed from the expansion of its signal on the retina by assuming
a constant velocity for the ball. In practice, although subjects do need to
see the ball for at least 330 ms to make the correct anticipatory muscle response,
they also take gravity into account when computing the ball's motion path.
The hypothesis of this study is that the brain instead assumes that gravity
will remain constant. In space, where the ball does not accelerate, the hypothesis
predicts premature anticipatory muscle responses, measured by surface electrodes
on the hand muscles and by arm movements detected by a video-computerized
technique. The hypothesis is supported by the observation that one newly arrived
cosmonaut on the Russian space station MIR, upon dropping his camera, immediately
reached "down" to grab it, missing completely as the camera continued moving
away in an unanticipated straight line.
If the constant-gravity model is supported by this experiment, the investigation
will focus on whether and how the response changes over the mission duration.
On the other hand, if subjects perform perfectly in microgravity from the
beginning, follow-up experiments during future shuttle flights are planned
to determine whether calculation of the gravitational force is based on visual,
vestibular, proprioceptive information or on efferent copy from motor commands.
These experiments will rely on a virtual environment, which allows visual
and vestibular cues to be manipulated independently. The virtual environment
generator includes a three-dimensional graphics workstation, helmet-mounted
display, head tracker and joystick. The subject also wears a harness that
pulls the body "downward" to simulate gravity. This apparatus gives the experimenters
more complete control of the sensory input than would be possible on earth.
Aside from producing temporary adaptations in the spatial computations
of adult animals, the sensory environment has profound and often permanent
effects on neural development. For example, visual experience during a critical
period is required to produce the normal pattern of connections from the thalamus
to the visual cortex, and changes in visual input can cause permanent deficits
in young animals. If neural activity is similarly important for developing
vestibular areas, then the signals provided by gravity are expected to be
critical for normal synapse formation in brain regions that process spatial
information.
Ken Kosik at Harvard and Oswald Steward at the University of Virginia are
testing this by investigating the effects of gravity on synapse development.
Young rats are raised in microgravity from postnatal day 4 to 20 days of age.
This period was chosen because synapse number, dendritic complexity and spine
number, and expression of synaptic proteins all increase dramatically in
the hippocampus during this time. Upon returning from space, the animals will
be perfused for light and electron microscopy. Because of space shuttle procedures,
the earliest time point is approximately 24 hours after landing, and the second
is at 30 days after landing to assay for long-lasting changes. Synapse development
in several brain regions will be determined with unbiased stereological sampling
to determine volume and synapse density.
Three regions have been chosen for analysis. The vestibulocerebellum receives
input from vestibular nuclei and sends an inhibitory projection to brainstem
motor pathways thought to be important for posture and balance. The parietal
cortex receives information from sensory association areas and projects to
motor cortex. As discussed above, the hippocampus signals location in space
and responds to vestibular information. Given that many other sensory systems
show experience-dependent plasticity during development, it is possible that
any or all of these regions may show alterations as a result of the exposure
to microgravity.
There is of course much uncertainty as to whether any of the Neurolab experiments
will work exactly as planned. They have been thoroughly screened for feasibility,
with extensive pilot experiments culminating in dummy runs performed by the
astronauts themselves in a mock-up of the shuttle lab. But there is plenty
that could still go wrong; McNaughton notes, for instance, that although his
rats are trained and the tetrodes are already implanted and working (two weeks
prior to the launch date), the effects of brain edema in zero gravity may
yet compromise the recordings once the rats are in space. Worse still, the
rats may decide on exposure to zero gravity that the novelty of floating round
the cabin outweighs the rewards of cooperation with the experimenters. But
these uncertainties are inescapable; it is of course the unfamiliarity of
the zero gravity environment that constitutes the justification, as well as
the risk, of all the Neurolab experiments.
