Exp Brain Res (1993) 96:328-334
Experimental
Otolith responses in man during parabolic flight
J.T. Marcus 1, A. Kuipers 2, G.F. Smoorenburg 1,3
i TNO Institute for Human Factors IZF, Kampweg 5, 3769 DE Soesterberg, The Netherlands
2 Netherlands Aerospace Medical Centre, Soesterberg, The Netherlands
3 Department of Otorhinolaryngology, University Hospital Utrecht, Utrecht, The Netherlands
Received: 28 February 1992 / Accepted: 14 May 1993
Summary. The influence of the varying gravito-inertial
(Gz) force during parabolic flight on human otolith func-
tion was investigated experimentally. It was hypothesised
that a varying Gz force profile initiates an otolith-ocular
response that manifests itself in modulation of optokinet-
ic nystagmus slow-phase eye velocity (OKN-SPV). Six
subjects were seated in the ESA-Caravelle, facing perpen-
dicular to the aircraft's longitudinal axis. The G z profile
was subsequently 1.8 G z pull-up, 0 G z microgravity, and
1.8 G z recovery, each phase lasting about 20 s. Vertical
eye movements were recorded with electro-nystagmogra-
phy throughout the parabolic manoeuvre. Conditions
were: (1) visual fixation, (2) darkness and (3) optokinetic
stimulation of 50 deg/s in an upward or downward direc-
tion, projected on a cylindrical screen at 0.6 m viewing
distance. No consistent nystagmus or gaze shift was mea-
sured in darkness. With optokinetic stimulation, howev-
er, A N O V A revealed downward enhancement of OKN-
SPV by 5~ in 1.8 G z hypergravity, as compared with the
0 Gz condition and the 1 Gz condition. It is concluded
that an otolith-ocular pathway modulates optokinetic
eye movements in parabolic flight.
Key words: Parabolic flight - Gravito-inertial force
Microgravity - Vestibulo-ocular responses - Man
Introduction
Previous research
The present study was undertaken as part of a research
programme on the influence of linear accelerations on the
human vestibular system, especially on the otoliths.
Gravito-inertial otolith input is defined as
G = Specific force = resulting force per unit mass
Correspondence to: J.T. Marcus
BrainResearch
9 Springer-Verlag1993
which is expressed in g units, g denoting the specific force
of the gravitational field on earth, which is 9.8 m/s 2. The
subscript z refers to the subject's longitudinal head-to-
foot axis. 1 Gz and 1.8 Gz are abbreviations for Gz = 1 g
and G z = 1.8 g.
In a centrifuge we observed that a 3 G z gravito-inertial
force induced a sustained vertical nystagmus with slow
phase down (Marcus and Van Holten, 1990). In the cen-
trifuge, however, fast Gz transitions are always inherently
coupled with high angular accelerations in the subject's
sagittal plane. These angular accelerations are not repre-
sentative of normal flight patterns and induce responses
of the semicircular canals confounded with otolith re-
sponses.
The parabolic flight manoeuvre provides conditions
of fast Gz transitions, without a dominant vestibular
canal stimulation. In this manoeuvre, even minor canal
responses can be separated from otolith responses by po-
sitioning the subject such that the angular acceleration
and the potential Gz influence are effective in mutually
orthogonal planes. In the centrifuge, the amplitude of
vestibular nystagmus slow phase velocity (SPV) has been
found to increase by 5 deg/s per g unit increase from the
1 G z gravitational force (Marcus and Van Holten 1990).
In parabolic flight, the G z value ranges between 0 g and
1.8 g; the maximum vestibular SPV amplitude evoked by
parabolas might thus be expected to be 4 deg/s.
Optokinetic nystagmus as a tool for measuring otolith
responses
Additional optokinetic stimulation might provide a t o o l
for obtaining a more sensitive indicator of dynamic
otolith function. When the visual field moves with a ve-
locity greater than about 20~ the tracking eye velocity
will lag behind the stimulus. This means that the gain of
optokinetic nystagmus (OKN), defined as the ratio of eye
velocity to stimulus velocity, will be smaller than unity
(Van de Berg and Collewijn 1988; Murasugi and Howard
 329

1989). In that case OKN gain may be sensitive to other

inputs, such as an input from the otolith system. 130Kt
Using a linear track, Buizza et al. (1980) showed that
the influence of linear acceleration on the otoliths reveals 32000
itself as a modulation of the OKN. Clement et al. (1989)
studied OKN stimulation in man during parabolic flight, 29000
and observed a downward shift in the beating field upon aircraft
altitude(ft)
entering microgravity. An otolithic effect on human 300Kt
OKN was also manifested in the effects of static tilt on 25000 300 K ~ ~ , :
level pull-up [recovery
OKN (Clement and Lathan 1991). In monkeys, Raphan I
/----
and Cohen (1988) showed that otolith responses to grav-
ity can become manifest in the alignment of OKN to- G~lg) I !_#?Z'-__
wards the gravity vector. I
I
50 I
I
predicted
eye velocityup
Hypothesised effect in parabolic flight (deg.s 1)
L I

The vertical upbeat nystagmus induced by + Gz, as ob- I I

served in the centrifuge, might appear in the form of a 1 i

change in optokinetically induced eye movement veloci-
ty. The driving force of the otoliths on eye movement
might thus be demonstrated by simultaneous application
of vertical optokinetic stimulation during a complete
parabolic flight manoeuvre. It is hypothesised that the
optokinetic nystagmus slow-phase velocity (OKN-SPV)
will be modulated by the G z load varying between 0 and
1.8 Gz. The predicted modulation is presented in Fig. 1.
I 20s 25s 20s
start time
optokinetic
stimulusup
Fig. 1. The parabolic aircraft trajectory (upper trace) is plotted over
time (from ESA documentation, modified). 1 Kt = 1.852 k/h. The
resultant G force vector, shown in the middle trace, is continuously
aligned with the aircraft's z-axis. The predicted modulation of verti-
cal optokinetic eye velocity implies a downward drive to A1.8 in
1.8 Gz hypergravity and an upward drive to Ao in 0 G microgravity.
The amplitude A1.8 is the mean OKN-SPV value from the 1.8 Gz
pull up and the 1.8 Gz recovery phases

Materials and m e t h o d s

Parabolic manoeuvres

Parabolas were flown with the ESA "zero G" Caravelle, based at
the Centre d'Essay en Vol in Bretigny, France. Microgravity in the
aircraft is obtained by entering a free-fall curve. This curve is a
parabola because it spans a small range over which the earth's
gravitational field can be treated as uniform. During parabolic
flight, the aircraft pulls up from straight and level flight, entering a o
1.8 G~ hypergravity phase lasting 20 s. At 50 deg nose-up the pilot
starts the microgravity phase. This moment is indicated as "injec-
tion", because at this point the aircraft is "injected" into a purely
parabolic flight path. After about 20 s, at 50 deg nose-down, the
aircraft enters the recovery phase with, again, 1.8 Gz hypergravity
during 20 s.

Orientation of the resultant force vector

This vector is defined within the aircraft's orthogonal reference

frame x (aft-forward), y (lateral) and z (up-down). During pull-up
and recovery, the pilot keeps the resultant force vector oriented in
the z-direction by careful manipulation of the aircraft's forward
acceleration, velocity and altitude. The G z value at each subject's
location was continuously recorded with a z-axis accelerometer
type Entran Egax-5, range _+5 G, frequency range 0-100 Hz.
The z-orientation of the G vector was verified by continuous
measurement. The a z vector was registered on our data-recorder
simultaneously with the eye movement data. In addition, the x, y
and z vector components were continuously measured throughout
the parabolic manoeuvres by the flight engineer. The Gx vector had
a maximum value of + 0.1 g, and this value was reached only
during the 1.8 Gz pull-up phase. The Gy vector was always less than
0.02 g. Thus, the resultant gravito-inertial vector is aligned with the
longitudinal axis
Fig. 2. The subject is sitting in front of a half-cylindrical projection
screen, with 0.6 m radius of curvature. The number 100 refers to a
length of 100 cm. Viewing distance is 0.6 m. The combination of
chair, screen and projection cylinder has been fixed to the bottom of
the aircraft. The subject is facing perpendicular to the aircraft's
longitudinal axis: the angular velocity of the aircraft during parabo-
las is effectively in the subject's frontal plane, and is not expected to
influence the nystagmus in the subject's sagittal plane
330
z-axis of the aircraft during the entire parabolic manoeuvre, within
an error range of about 3~. It follows that the subject, strapped in
the chair and with his head supported, did not experience any swing
of the resultant vector direction. Horizontal eye movements were
continuously measured during all parabolas, in order to check for
any horizontal gaze shift or nystagmus.
Method of optokinetic stimulation
The projection screen was homogeneous, white, and cylindrically
curved subtending a visual angle of 80~ horizontally and 95 ~ verti-
cally (see Fig. 2). The bar pattern was generated by means of an
Osram 8.5 V DC, 6A tungsten bulb (type Wi 9) with a 5 cm linear
filament, inside a rotating cylinder. The wall of the cylinder con-
tained 15 slits parallel to its main axis, enabling projection of dark
and light bars (12 ~ each) by shadow casting. The luminances of the
light and dark bars were 10 and 0.5 cd/m 2, respectively. The applied
angular velocity of the cylinder was 50 deg/s upward or downward.
The stimulus projector above the head was off axis with respect to
the axis of the cylindrical screen, implying a slight non-uniformity of
pattern velocity vertically across the screen. It is estimated that the
pattern velocity increased with about 1 deg/s for an upward gaze
shift of 4 ~ across the screen. The chair-screen combination was
surrounded by non-reflecting black curtains.
Protocol
In one flight, two sets of 15 parabolas each were flown. On board
were a subject and an operator, who exchanged roles after the first
set. Stimulus conditions besides the O K N stimulation were: visual
fixation of the stationary bar pattern and darkness.
The fixation condition was applied in parabolas 1 and 10 in
order to measure a possible drift in the E N G eye position signal, for
example due to electrode offset shift or vasovegetative reactions of
the subject. The projection lamp was switched on, the rotor
switched off and the subject fixated at the edge in front of him
during the complete parabola. In the darkness condition the projec-
tion lamp was switched off and an additional black curtain was
placed in front of the subject's eyes, which were kept open. This
condition was applied to measure the gaze shift and a potential
vertical vestibular nystagmus, which could both influence the
OKN-SPV. The planned sequence for each subject was as follows
(ok = optokinetic stimulus):
Parabola number Condition
1 fixation
2 dark
3 dark
4 ok up
5 ok down
6 ok up
7 ok down
8 ok up
9 ok down
10 fixation
11 dark
12 dark
13 ok up
14 ok down
15 ok up
The optokinetic stimulus started 10 s before pull-up and was
kept constant during the complete G~ profile. Between two subse-
quent parabolas there was an interval of straight and level flight
lasting for 2 min.
Subjects
Six male subjects (aged 2 6 4 7 years) underwent medical tests and
hypobaric chamber training according to ESA requirements. They
signed informed-consent forms of the experiment protocol and did
not use any medication. They were trained to perform a steady state
optokinetic nystagmus lasting 60 s in the stimulus setting that was
to be applied in flight. Their instructions were to look at as many
black bars as possible, not to pursue any missed bar, and to keep
their gaze in the middle of the screen within deviations of 15~
Eye movement analysis
Vertical and horizontal eye movements were measured with electro-
nystagmography (frequency range of amplifiers 0.04 Hz-100 Hz),
and recorded on an F M data-recorder. Calibration of the eye-
movement signals was performed before take-off. The FM recorded
(0-1 kHz) signals were low-pass filtered with a cutoff at 50 Hz, and
subsequently sampled at 100 Hz. The optokinetic nystagmus slow
phase velocity (OKN-SPV) was analysed off line in the following
steps:
1. Differentiation of eye position.
2. Removal of fast phases in eye velocity and interpolation of SPV
during fast phases. This was performed by an interactive computer
program that recognises fast phases whenever threshold limits in
eye velocity are exceeded (Barnes 1982; Bos and Kistemaker, per-
sonal communication). In this way, a continuous SPV curve over
time was obtained throughout each parabola, for each subject. Fig-
ure 3 shows an example of eye position and velocity during the
transition to microgravity.
3. The SPV curves from different parabolas were synchronised at
the time of "injection". Subsequently these curves were averaged
across parabolas, by subject and O K N direction. The number of
separate measures was thus equal to the number of these parabolas.
Figure 4 illustrates the effect of averaging for the upward OKN-SPV
traces: the signal-to-noise ratio is enhanced.
4. Calculation of mean SPV value ( • SD) over time during each Gz
condition: 1 Gz, 1.8 Gz pull-up, 0 Gz, 1.8 Gz recovery. The averaged
SPV response within each G condition was calculated over the time
interval starting at least 5 s after the G transition, and for as long as
the G value was in a steady state (_+0.1 g).
l i I I I I I
(g) 2 G,z
(dUpg)'O[~aeye posjtionA ^ A . ~
~ ............ ; .............. i. . . . . . . .
0 I 2 3 4 5 6
time (s)
Fig. 3. Example of eye movement analysis during the transition
from 1.8 Gz to 0 G. The eye position signal is differentiated in order
to obtain the velocity signal. The outliers of fast phase velocity are
then removed, and the slow phase velocity (SPV) is interpolated
over the time intervals of fast phases. In microgravity the SPV is
enhanced in the upward direction, when compared with the SPV
during 1.8 G z hypergravity
 331

PARABOLAS
1 I I I I I
P4 Oz )5,. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
I
: I I "~" I~ I I I

OKNSPV - .'~ '~-." .". ~~'-"' " "~ " . . . . I, ' ~

(=/s} 0 ......................................T .................................. i............... L..' ............................ "~< ......
I
9f I i I I I
2 1 ~ I I ~ ----'.1
P6
01 ........ i~,-J . ~. I
i f 1...................... -tl. . . . . . . . . . . .I. . . . . . . . . . . I. . . . . . . . . . . . . . .
I [ ^ I I
OKNSPV 40 ~.~.. . . . . . . " "i "~-', "-----~--J"~ ~-- .4", I "

(*/s} 0 ......... I ....................... ~ .................................. l ................ i.... : ....... i .................... :':':".....

I
2[ ~ i I i I 1
P8
0~; ........ t .............................?~--| ....................... 4 .4 ............l............... 1...........................
Fig. 4. Averaging of upward OKN-SPV traces
over parabolas 4, 6 and 8 for subject 2. For
OKNSPV [" '"," - "v'--J~ .~., I- " '~I ; ' ". . . . '~ ~'" " ~';'~"-I'-' ;'~" i~ ~ --
each separate parabola the G z profile and the
(~ O~ ..........i.................................... i .........................
[ ........................ i................ ................I"
SPV response are shown. The traces were syn-
89 [ ' , I, I I I , 10S , chronised at the time of injection. The vertical
4hi --- I
OKNSPV
[~
u ~'~..,,..~..,,/,..~,=..
/ I
i /
"t--
x,i, "~'-"'~'--'-~'~]'-"
I . I
~-~"~'J~
-I "
---~'-'~'x
..~
dashed lines delineate the time intervals in which
O~ ........... I .................................... t ........................ ~ ............................... I ..............................
the response was calculated for each G condi-
tion on the averaged trace 1/3(P4 + P6 + P8)
" u' ' 'L ' 3 'L 'J ' '
time P

Table 1. OKN-SPV and standard deviation in degrees per second, Table 2. OKN-SPV (~ values derived from Table 1, averaged
by subject, averaged over n parabolas across subjects

Subject Upward OKN 1 Gz 1.8 G z , pull_up 0 Gz 1.8 G . . . . . . . ,-y

1 Gz 1.8 G z , pull.u p 0 Gz 1.8 G z , recovery gt Upward +33.8 +24.8 +27.7 +21.7

Downward - 38.8 - 44.0 - 37.7 - 41.5
1 35-+3 30___5 36__.6 21-+5 3 i/2 (U+D) -2.5 -9.6 -5.0 -9.9
2 36+1 24_+3 28_+3 21-t-_3 3
3 42_+4 45_+3 40+4 46_+3 4
4 30_+3 18_-t-5 19_+5 10-+5 1
5 31-+5 16-+5 23_+4 18_+4 4 e q u i v a l e n t to 20 ~ eye position. H o w e v e r , these changes
6 29-+6 16_+4 20_+4 14_+5 3 were n o t s y s t e m a t i c in a n y one direction.
Subject Downward OKN

1 Gz 1.8 G z , pull_u p 0 Gz 1.8 G ......... y n The darkness condition

1 -45_+8 -55+2 -51• -52+4 2 A g a i n , offset shifts e q u i v a l e n t to 20 ~ were observed, m o s t

2 -44-+4 -49_+3 -44__3 -50_+2 2 m a r k e d l y d u r i n g G~ transitions. These shifts, however,
3 -46-+3 -45-+3 -52-+3 -53_+2 4 were n o t s y s t e m a t i c in the u p o r d o w n direction, except
4 -19_+4 -22_4 -15_+2 -15_+5 3
for subject 4 w h o s h o w e d a s y s t e m a t i c 20 ~ d o w n w a r d
5 -43+_1 -52_+5 -34_+5 -46-+5 4
6 -36_+2 -41_+4 -30-+3 -33_+3 3 gaze shift after injection to 0 G~. V e s t i b u l a r n y s t a g m u s
was o b s e r v e d o n l y in subject 2 d u r i n g p u l l - u p a n d recov-
ery, with a SPV of 4 deg/s d o w n w a r d .

5. Performance of an analysis of variance on these mean SPV values,

obtained from six subjects, two OKN conditions and four different The optokinetic condition
G conditions. If there was any significant effect of G level, the
post-hoc Newman-Keuls test (Wirier 1962) was applied to examine
which two G levels produced a significant difference between the The h o r i z o n t a l eye m o v e m e n t r e c o r d i n g s d i d n o t reveal
SPV values. a n y c o n s i s t e n t gaze shift o r n y s t a g m u s . This finding is in
6. From the SPV curves obtained in step 2, the modulation index a g r e e m e n t with the a b s e n c e of a n y v e c t o r tilt. T h e verti-
was calculated for each subject in each separate parabola. This was cal O K N - S P V of each subject was a v e r a g e d o v e r p a r a b o -
done in order to investigate the consistency of modulation within las, a n d s u b s e q u e n t l y time a v e r a g e d within each G c o n d i -
subjects.
tion. T h e resulting SPV values are given in Table 1 for the
G c o n d i t i o n s : 1 Gz j u s t before e n t e r i n g the p a r a b o l a ,
Results 1.8 G z pull-up, 0 Gz, a n d 1.8 Gz recovery. T h e m e a n value
a n d s t a n d a r d d e v i a t i o n were c a l c u l a t e d over the time in-
The fixation condition terval of each G level.
So far, responses to b o t h of the o p t o k i n e t i c stimulus
C h a n g e s in E N G offset o c c a s i o n a l l y o c c u r r e d after the d i r e c t i o n s have been a n a l y s e d separately. T h e h y p o t h e -
m o m e n t s of i n j e c t i o n a n d recovery, with a m p l i t u d e s sised SPV m o d u l a t i o n , however, s h o u l d manifest itself
332
Table 3. SPV modulation indices for sepa-
rate parabolas Subject Upward OKN stimulation
1 0.16 0.16 0.21
2 0.16 0.09 0.07
3 -0.03 -0.04 -0.11 -0.13
4 0.19
5 0.03 0.08 0.18 0.19
6 0.16 0.19 0.14
equally for both optokinetic directions; downward dur-
ing 1.8 G~, and upward during 0 G z. Linear addition of
vestibular and optokinetic signals is assumed here, as was
proposed by Robinson (1977) for angular acceleration
input. The modulation effect should therefore be inde-
pendent of the direction of the optokinetic stimulus.
Following this reasoning, an analysis of variance was
performed on the combined set of up and down respons-
es, resulting in a main effect due to G level (F3.15 = 9.4,
P < 0.01, 0.7% variance explained). 1 2 Post-hoc Newman-
Keuls analysis (Winer 1962) revealed significant differ-
ences (P<0.05): the SPV values at 1.8 G z (both during
pull-up and recovery) are downward enhanced when
compared with those at 1 G z and 0 Gz. However, no sig-
nificant difference was observed between 1 Gz and 0 Gz.
Table 2 presents the mean SPV values.
G modulation indices from separate parabola responses
In the analysis so far, the responses have been averaged
across parabolas. The question remains of whether, for
each subject, the modulation in SPV is constant over
separate parabolas. We did not observe a systematic
change of OKN-SPV over the course of each individual
parabolic manoeuvre, and thus we will focus on the aver-
age SPV response at each G level of an individual ma-
noeuvre. In order to quantify the modulation in SPV, an
index is defined as follows. The mean SPV response am-
plitudes at 0 Gz and 1.8 Gz are denoted A0 and A1.8, with
A1.8 determined by the mean of the 1.8 Gz,pull_up and
1.8 G z , r e c . . . . y responses. The index of SPV modulation is
defined as (A0-A1.8)/(A0 +A1.8) The values for each ma-
noeuvre are given in Table 3.
The index values in Table 3 show that the sign of the
modulation is highly consistent for each subject. As was
hypothesised in Fig. 1, the value of A 0 is larger than A1.8
during upward OKN, resulting in a positive index; the
reverse holds true during downward OKN. Only for sub-
ject 3 is the sign of modulation contrary to the prediction,
as already may have been apparent from Table 1. Be-
cause of the consistency of the effect in the other five
subjects, the overall effect remains significant.
The salience of this G effect is not apparent from the percentage
explained variance, because the difference between upward and
downward OKN already accounts for 89% of variance
2 In the combined up/down dataset we did not find a significant
interaction between the effect of Gz and upward or downward OK
direction. This finding supports the assumption of linear addition of
the vestibular and optokinetic signals, and justifies the approach of
combining the up and down responses
Downward OKN stimulation
-0.01 -0.07
-0.07 -0.04
0 0.04 0.04 0.06
-0.10 -0.08 -0.10
-0.06 -0.12 -0.40 -0.11
-0.19 -0.10 -0.16
O K N responses were not obtained from all scheduled
parabolas with O K N stimulation, the main reason being
motion sickness (subjects 2, 4, 6). No difference in re-
sponse was found between these subjects and the three
other subjects resistant to motion sickness. For subject 1,
the sequence of parabolas was interrupted and discontin-
ued because the aircraft had insufficient fuel. For the sub-
jects three and five, the whole series of parabolas was not
completed, because at the end of the sequence they were
fatigued which caused interrupted O K N responses dur-
ing the final parabolas.
Discussion
Potential artefacts of arousal, fatigue and gaze shifts
Arousal and fatigue are known to influence the nystag-
mus response. Therefore, these factors should be separat-
ed from the presumed Gz effect. Since five of the six sub-
jects were undergoing their first parabolic flight, arousal
might have occurred when entering the 0 G phase, result-
ing in the observed increase in amplitude of upward SPV.
However, the same arousal effect should cause an in-
crease in the downward SPV when entering 0 Gz, which
was not the case. A subject's oculomotor system might
become fatigued by the long duration of sustained op-
tokinetic eye movements, extending over more than 60 s.
Such fatigue may cause a continuous decline in SPV am-
plitude during a parabolic manoeuvre; it cannot, howev-
er, explain the increased upward SPV during 0 G z (com-
pared with 1.8 Gz pull-up), nor the increased downward
SPV during 1.8 Gz recovery (compared with 0 Gz).
A systematic gaze shift during the G z profile might
influence the results, because of the deviation in pattern
velocity across the screen (it was estimated that pattern
velocity increases with l~ for an upward gaze shift of
4~ This potential artefact, however, should cause the
same amplitude effect on up (positive) and down (nega-
tive) SPV, which would thus cancel itself out in the com-
bined up/down dataset.
Potential artefacts due to the ENG technique
E N G electrode artefacts might have influenced the offset
of the E N G signal during a parabola. The "fixation" con-
dition was included to check for any E N G offset arte-
facts. The eye position data showed some shifts, but these
were unsystematic in terms of their direction as com-
pared to the direction of the change in G. Admittedly, the

measurements of ocular position drift were limited to the
frequency-bandwidth of the ENG amplifiers (0.04-
100 Hz). The time constant corresponding with the lower
corner frequency is 4 s, and thus long-term offset shifts
may have been underestimated. Also, the ENG calibra-
tion factor of eye position might have changed during
parabolas. However, any calibration shifts would have
caused the same change in absolute value of eye velocity
in both the upward and downward OKN direction. The
'modulation', induced by this calibration shift, would be
opposed in amplitude, and be cancelled out when taking
up and down responses together.
In summary, it is unlikely that any artefact can ac-
count for the observed effect. Artefacts would potentially
induce an amplitude-effect on OKN-SPV, while the com-
bined set of up and down OKN responses, as shown in
Table 2, exhibits a directional effect on OKN-SPV, in-
duced by the Gz vector.
Role of semicircular canal stimulation
It is unlikely that any semicircular canal response could
have influenced our results. We already mentioned the
positioning of the subject in the aircraft (shown in Fig. 2):
during parabolic manoeuvres the axis of aircraft rotation
is parallel to the roll axis of the head, such that the pitch
movement of the aircraft would only cause torsional
rather than vertical ocular responses of the subject. 3 Also,
the angular acceleration of the aircraft's pitch manoeuvre
was restricted to a few degrees per second per second,
which would only cause a low amplitude of canal re-
sponses. However, only rigorous three-dimensional anal-
ysis of eye movements can give complete experimental
evidence of absence of any canal responses.
Reference to related observations
The Gz effect in the centrifuge was revealed in a vestibular
downwards SPV at a mean amplitude of 5 deg/s per g
(Marcus et al. 1989; Marcus and Van Holten 1990). The
vestibular nystagmus of 4 ~ observed in "darkness"
parabolas for one subject (number 2), is quantitatively
quite consistent with the centrifuge SPV. The mean G
modulation of OKN in parabolas amounts to 2.8 deg/s
per g, indicating that the OKN modulation per g unit in
flight is lower than the vestibular nystagmus amplitude in
the centrifuge. The direction of modulation however is
consistent with the centrifuge data.
The directional effect is also consistent with data of
Clement et al. (1989) in a study on two subjects in
parabolic flight. He observed a decrease of SPV down
and an increase of SPV up during the transition to
freefall. Clement and Lathan (1991) applied the method
of static tilt about the roll axis. They observed in man
that the gain of upward vertical OKN was increased in
3 The fact that the axis of aircraft rotation was not exactlythrough
the roll axis of the head does not mean that the semicircularcanals
in the sagittal plane of the subject were simulated
333
90~ roll position. In terms of vectors, the otolith-induced
SPV is opposed to OKN-SPV in the upright subject, and
orthogonal to OKN-SPV in the 90~ roll position. The
increased upward gain in roll can be explained by the
release of the downward otolith-induced SPV component
with respect to the subject's head.
DiZio and Lackner (1992) found that, in parabolic
flight, vertical vestibular nystagmus (induced by angular
acceleration on a rotating platform) showed no effect of
G level on peak SPV amplitude. In their experiment,
however, the subject was lying on his side, with the right
ear down to the floor of the aircraft. The subject therefore
must have experienced the G force along his lateral axis,
orthogonal to the vertical SPV direction, which probably
explains the lack of a significant G effect on vertical SPV
amplitude.
It has been shown in electrophysiological studies in
cats that responses to vertical linear acceleration are ini-
tiated by the otoliths (Xerri et al. 1987). Fukushima and
Fukushima (1991) reported on cats that otolith-induced
SPV responses are in phase with specific force G.
Raphan and Cohen (1988) have described another
mechanism in which velocity storage plays a role in the
orientation of optokinetic following with respect to the
earth's 1 g gravity. This mechanism, however, cannot ac-
count for the hypergravity effect in our conditions. Ex-
tended studies on the three-dimensional nature of the
otolith-ocular reflex and its influence on velocity storage
are continuing (De Jong and Oosterveld 1987; Angelaki
et al. 1991; Gizzi et al. 1991; DiZio and Lackner 1992).
Comparison between 1 Gz and 0 Gz responses
At this point in time, an explanation for the apparent
lack of a difference in SPV between the 0 Gz and the 1 Gz
condition is still missing. It may be due to the absence of
an immediate transition between 1 G z and 0 G z. The
"high-G barrier" between the two conditions may offset
the system. Perhaps a difference between 1 Gz and 0 Gz
might indeed be found if the transitions between 1 G z and
0 Gz were made more direct. This approach would, how-
ever, require another flight manoeuvre.
ConcLusion
During parabolic flight, the otolith responses to the Gz
profile were measured by analysing the vertical slow-
phase eye velocity. In darkness, vertical vestibular nys-
tagmus was observed in only one of six subjects. With
vertical optokinetic stimulation, however, the otolith re-
sponses were manifest in modulation of the vertical op-
tokinetic slow-phase velocity.
Statistical analysis revealed (1) In 1.8 G z hypergravity, the
amplitude of upward OKN-SPV is decreased, and the
amplitude of downward OKN-SPV is increased. This
holds true for the comparison between 1.8 Gz and 1 Gz,
and between 1.8 Gz and 0 Gz. (2) No significant differ-
ence, however, was found between the 1 Gz and 0 G z con-
334
dition. (3) N o significant i n t e r a c t i o n was f o u n d b e t w e e n
the O K s t i m u l u s d i r e c t i o n (up o r down), a n d the G level.
This s u p p o r t s the a s s u m p t i o n of linear a d d i t i o n of the
o t o l i t h i c a n d o p t o k i n e t i c eye velocity signals.
Acknowledgements. The parabolic flight facility was kindly provided
to us by the European Space Agency (ESA). Additional financial
support was given by the Space Research Organization Netherlands
(SRON). The technical support of the staff of the Department of
Medical Physics, Free University Amsterdam, is highly appreciated.
Dr. Mathew Sandor (ESTEC, The Netherlands) made helpful sug-
gestions for improving the text.
References
Angelaki DE, Anderson JH, Blakley BW (1991) Changes in the
dynamics of the vestibulo-ocular reflex due to linear accelera-
tion in the frontal plane of the cat. Exp Brain Res 86:27-39
Barnes GR (1982) A procedure for the analysis of nystagmus and
other eye movements. Aviat Space Environ Med 53(7): 676-682
Buizza A, Leger A, Droulez J, Berthoz A, Schmid R (1980) Influ-
ence of otolithie stimulation by horizontal linear acceleration on
optokinetic nystagmus and visual motion perception. Exp Brain
Res 39:165-176
Clement G, Andre-Deshays C, Lathan CE (1989) Effects of gravito-
inertial force variations on vertical gaze direction during oeulo-
motor reflexes and visual fixation. Aviat Space Environ Med
60:1194-1198
Clement G, Lathan CE (1991) Effects of static tilt abdut the roll axis
on horizontal and vertical optokinetic nystagmus and optoki-
netic after-nystagmus in humans. Exp Brain Res 84:335 341
De Jong HAA, Oosterveld WJ (1987) Rotation test in the weightless
phase of parabolic flight. Aviat Space Environ Med 58 [Suppl 9]:
A253-6
DiZio P, Lackner JR (1992) Influence of gravitoinertial force level
on vestibular and visual velocity storage in yaw and pitch. Vi-
sion Res 32:111-120
Fukushima K, Fukushima J (1991) Otolith-visual interaction in the
control of eye movement produced by sinusoidal vertical linear
acceleration in alert cats. Exp Brain Res 85:36-44
Gizzi MS, Rudolph SH, Cohen B, Raphan Th (1991) Gravitational
influences on human optokinetic nystagmus. Investig Ophthal-
tool Vis Sci 32(4): 895
Marcus JT, Bles W, Van Holten CR (1989) Influence of gravitoiner-
tial force on vestibular nystagmus in man, observed in a cen-
trifuge. Adv Space Res 9(11): 213-222
Marcus JT, Van Holten CR (1990) Vestibulo-ocular responses in
man to + Gz hypergravity. Aviat Space Environm Med 61:631-
635
Murasugi CM, Howard IP (1989) Up-down asymmetry in human
vertical optokinetic nystagmus and afternystagmus: contribu-
tions of the central and peripheral retinae. Exp Brain Res
77:183-192
Raphan T, Cohen B (1988) Organizational principles of velocity
storage in three dimensions: the effect of gravity on cross-cou-
pling of optokinetic after-nystagmus. Ann NY Acad Sci 545: 74-
92
Robinson DA (1977) Linear addition of optokinetic and vestibular
signals in the vestibular nucleus. Exp Brain Res 30:447-450
Van den Berg AV, Collewijn H (1988) Directional asymmetries of
human optokinetic nystagmus. Exp Brain Res 70:597-604
Winer BJ (1962) Statistical principles in experimental design. Mc-
Graw-Hill, New York, pp 80-85
Xerri C, Barthelemy J, Harlay F, Borel L, Lacour M. (1987) Neu-
ronal coding of linear motion in the vestibular nuclei of the alert
cat. Exp Brain Res 65:569-581