Brain (1994), 117, 1423-1432
Vestibular, cervical and visual remembered saccades in Parkinson's disease T. Nakamura,* A. M. Bronstein, C. Lueck, C. D. Marsden and P. Rudge MRC Human Movement and Balance Unit, Institute of Neurology, National Hospital for Neurology and Neurosurgery, London, UK
Correspondence to: Dr Adolfo Bronstein, MRC HMBU Section of Neuro-Otology, National Hospital for Neurology and Neurosurgery, Queen Square, London WCIN 3BG, UK *Present address: ENT Department, Yamagata University Medical School, Yamagata, Japan
Summary In order to assess both vestibulo-cervical perception of head rotation and saccadic function in Parkinson's disease, 14 patients with idiopathic Parkinson's disease were subjected to discrete sigmoid-shaped rotational displacements whilst fixating a target aligned with primary gaze in an otherwise dark room. The rotational stimuli were applied to (i) the whole body (vestibular stimulus); (ii) the trunk whilst the head remained stationary in space (cervical stimulus); (Hi) to the head alone whilst the trunk remained stationary (combined vestibular and cervical stimulus). The fixation target was then extinguished and the subjects had to estimate the angle travelled by the head or trunk with an ocularpointing task, using information from the preceding rotational stimulus (vestibular and cervical 'remembered' saccades). It was found that, although these saccades in Parkinson's
disease patients were multiple-step and hypometric, the final position of the eyes matched the rotational stimulus as accurately as in normal subjects. A complementary experiment in six patients showed that visual remembered saccades were hypometric, but significantly less so than vestibular remembered saccades. It is concluded that (i) vestibular and cervical perception of head/neck rotation is normal in Parkinson's disease; (ii) abnormalities of 'remembered' saccades, previously reported in Parkinson's disease, are not confined to the visual modality but involve other sensory modalities as well; (Hi) across different modalities of memory-guided saccades, visual input improves saccadic performance. This result demonstrates that the known increased visual dependence found in Parkinson's disease extends to memory-driven tasks.
Key words: Parkinson's disease; saccades; vestibular; visual; memory
Introduction The interest in ocular motor function in Parkinson's disease is two-fold. Some studies have attempted to provide additional clinical information to help in the differential diagnosis from other akinetic-rigid syndromes. In this regard, there now seems to be a consensus that visually guided eye movements (smooth pursuit, saccades) are either normal or only mildly affected in idiopathic Parkinson's disease (Kennard and Lueck, 1989; Rascol et al., 1989; Tanyeri et al., 1989; Stell and Bronstein, 1994). Reports of more severe abnormalities, mainly described in the earlier literature, are likely to have been due to inclusion of cases with additional vascular, anoxic or surgical central nervous system (CNS) damage (White et al., \9S3a,b) as well as probable cases of SteeleRichardson-Olszewski syndrome and multiple system atrophy (Corin etal., 1972), or to lack of age-matched normal controls (Shibasaki etal., 1979). Other studies have addressed specific questions as to the nature of the disorder of movement in Parkinson's disease. In these studies, the significant abnormalities were seen in © Oxford University Press 1994
special saccadic paradigms known as 'self-paced' (DeJong and Jones, 1971), 'predictive-anticipatory' (Bronstein and Kennard, 1984, 1985; Ventre et al., 1992), 'visually remembered' (Crawford et al., 1989; Lueck et al., 1990, \992b) and 'volitional' saccades (Lueck et al., 1990, 1992a) in which delayed or hypometric saccades occurred. Of note, the programming of saccades in such paradigms is largely based on memorized visuo-spatial information. The fact that they are all affected in Parkinson's disease is consistent with experimental work in monkeys in which basal ganglia neurons have been identified that become selectively activated prior to saccades made to the remembered location of visual targets (Wurtz and Hikosaka, 1986; Hikosaka, 1991). Spatial information for the generation of saccades towards the remembered position of a target, however, can arise from modalities other than vision. For instance, information of head/neck motion can be stored and used to generate saccades towards a remembered position in space, experimental paradigms known as vestibular (Bloomberg etal., 1988,1991)
1424
T. Nakamura et al.
and cervical (Nakamura and Bronstein, 1993) 'remembered' saccades. Figure 1 can be examined for details of these paradigms but, in brief, the function explored with such experiments underlies the ability that man has to orientate and operate in space—even if in the dark and during motion. In this paper we investigate the accuracy of saccades in response to preceding angular motion of the head or trunk in patients with Parkinson's disease in order to define whether the deficit reported in visual remembered saccades is specifically related to visual memory or if, on the contrary, it involves other sensory modalities. The study of vestibular and cervical afferent function is particularly relevant to Parkinson's disease because it has often been suggested that impairment in vestibular and proprioceptive function may partly contribute to the postural disorder in this condition (Martin, 1965, 1967; Reichert et al., 1982; White et al., 1983a). This study therefore has two interrelated aims, one to examine the perception of head/neck angular displacement in Parkinson's disease and the other to explore further the nature of the observed saccadic abnormalities.
Material and methods Fourteen patients with idiopathic Parkinson's disease were tested. There were 11 males and three females with a mean age of 51.8 years (range 40-64 years). The clinical severity of akinesia, rigidity, tremor and postural disorder, and HoehnYahr staging were assessed immediately before or after the experiments. Ten patients were in stage 1 or 2, three patients in stage 3 and one patient in stage 4; there was good agreement between clinical neurological and Hoehn-Yahr severity. In all the patients, clinical examination of eye movements was normal. There was no patient with clinical evidence of additional CNS disorder or dementia, or previous history of vestibular disease. Six of the 10 patients with mild to moderate disease (Hoehn-Yahr, 1-2) and three out of four with more severe Parkinson's disease (Hoehn-Yahr, 2>-A) were on no medication at the time of testing. The control group consisted of 16 age-matched volunteers; 12 males and four females with a mean age of 53.2 years (range 40-68 years). All subjects gave informed consent to the experiments. Horizontal eye position was recorded by DC electrooculography using bitemporal electrodes with a bandwidth of 0-70 Hz. Eye position was calibrated with fixation points ±8°, 16°, 24°, 32° and 40° before each test condition. The experiments will be described in two groups, those concerned with vestibular and cervical stimuli, and those concerned with visual stimuli.
Vestibular and cervical stimuli Body rotation was delivered by an electric servomotor turntable system (Contraves-Goerz Inc.; 120 Nm torque) controlled by computer during whole body rotation (vestibular stimulation) or turned by hand, for safety reasons, when the trunk was rotated with the head stationary in space (cervical
stimulation). The chair was fitted with a rigid rod carrying a small light-emitting diode (LED) on its end, positioned at a distance of 70 cm in front of the chair rotational axis (chairfixed target; Fig. 1). Another LED was mounted on the floor (earth-fixed target) in line with primary gaze, some 2-3 cm further than the chair-fixed target. For the experiments requiring head rotation on the trunk (combined vestibular and cervical stimulation), a helmet connected to a low torque potentiometer was used to monitor horizontal head position. A 50 cm long rigid rod was attached to the front of the helmet, carrying a LED on its end (head-fixed target). Chair velocity was transduced by a tachometer and chair angular displacement was obtained by digital integration of the tachometer signal. Chair position was frequently monitored visually by the experimenter but there was never disagreement between the direct readings and those obtained by digital integration. Head velocity was derived from digital differentiation of the helmet potentiometer signal. Eye, chair, head and target signals were digitized with a sampling rate of 250 Hz through a 12-bit analogue/digital converter and stored on microcomputer for off-line analysis. Experiments were performed in complete darkness, except for the LEDs switched 'on' and 'off as described below. In all four experimental conditions, motion stimuli were discrete sigmoid-shaped displacement; during motorized rotation this was obtained by a raised cosine angular velocity stimulus generated by the computer. Each experimental paradigm is described below and illustrated in Fig. 1.
Vestibular stimulus (VS): whole body (head and trunk) rotation Subjects were seated on the rotating chair with the head firmly supported by a head rest and a binaural head clamp. Before each stimulus the chair was positioned so that both the chair-fixed target and the earth-fixed target were in line one behind the other. The subjects fixated the chair-fixed target for 3-8 s and then, with the earth-fixed target extinguished, the chair was smoothly rotated towards the right or left at predetermined amplitudes of 10°, 20°, 30° and 40° with a peak velocity of 30° or 407s which were randomly selected (Fig. 1, top). Approximately 1 s after stopping the chair, the chair-fixed target was turned 'off and the subjects were instructed to look, in total darkness, to the point from which rotation had started, i.e. to locate the imaginary earth-fixed target. Finally, the earth-fixed target was illuminated and fixated by the subjects so that any residual error could be measured (Fig. 1, top right).
Cervical stimulation (CS): trunk rotation with earth-stationary head In this experiment, the subjects sat on the rotating chair with the head rigidly fixed by a chin rest and a binaural head clamp attached to a rigid wall-mounted frame. Before starting
Vestibular saccades in Parkinson's disease
1425
VS
CS
VS + CS
Fig. 1 Schematic drawing of the experimental paradigms. The open circles represent 'target on' and the solid circle, 'target off. The dotted lines indicate the direction of gaze and the arrows, the saccadic eye movements. Top: vestibular stimulation (VS) paradigm to assess 'vestibular remembered saccades'. The subject's whole body is rotated whilst fixating on the chair-fixed target. When the chair-fixed target is switched off, the subject has to locate the starting position with his/her eyes, using stored information from the preceding rotation. Error in the task can be detected by the presence of a saccade required to refixate the earth-fixed target. Bottom left: cervical stimulation (CS). The subject's head is fixed to the wall so that it remains stationary in space during trunk rotation. During motion, optic fixation is maintained on the earth-fixed target. After stopping the chair, the subject is required to look in the direction of his/her trunk or legs when the earth-fixed target is switched off. Note the different orientation of the trunk and head which is, as sensed by neck receptors, the source of information for this task. Horizontal trunk position in space is signalled by the chair-fixed target. Bottom right: passive and active combined vestibular/cervical stimulation (VS+CS). The subject's head is turned upon the shoulders whilst fixating the head-fixed target. On completion of the movement, the subject's task is to redirect the eyes to the starting position.
rotation, both chair-fixed target and earth-fixed target were lit and in line. The subjects fixated the earth-fixed target for 3-8 s and then, with the chair-fixed target extinguished, the chair (trunk) was smoothly displaced by hand varying direction, amplitude and velocity. About 1 s after stopping the chair, the earth-fixed target was turned off and the subjects were asked to look in the direction in which their trunk, chest or legs were pointing (all these words were used to facilitate the task), i.e. locate the imaginary chair-fixed target. Finally, the chair-fixed target was presented for fixation.
Passive and active combined vestibular and cervical stimulation (passive and active VS+CS): head rotation on trunk In these experiments, the subject's head (fitted with the helmet) was free to move upon the shoulders. For passive movements, the experimenter moved the subject's head by holding the head and helmet whilst standing behind the seated subject. The subjects were instructed to relax the neck.
After all the passive head turns were completed, the active experiments were conducted by instructing the subjects to move the head in a similar way to the preceding passive movements, i.e. single displacements varying direction, amplitude and velocity. In these experiments, both the headfixed target and earth-fixed target lights were initially illuminated and in line. The earth-fixed target was then switched 'off and the head was turned whilst maintaining optic fixation on the head-fixed target. About 1 s after completion of the head movement, the head-fixed target was extinguished and the subjects had to look back, with the eyes only, to the starting position, i.e. towards the imaginary earthfixed target. The earth-fixed target was finally lit for refixation. A total of 36-80 stimuli in each test were obtained on each individual subject. In four patients, only the whole body rotation paradigm was examined but the other cases had all paradigms. During hand-driven stimuli or active movement, a trial was excluded if stimulus amplitude was outside 5-45°, or stimulus velocity outside 25-457s. During the trials there was no visual feedback of head/body position
1426
T. Nakamura et al.
available to the subjects because they always maintained foveation of the target in front of the head during rotation and the amount of room illumination provided by the LED was negligible. All experiments were monitored through an infrared camera providing detailed view of the subject's face, eyes, LEDs and upper half of the rotating chair. The experimenter used this view to check subject's arousal and absence of blink/facial artefacts before switching off the fixation targets; in this way the time interval between motion stimuli and the instruction to make the saccade was variable (-1 s) and saccadic latency measurements were therefore discarded. Trial runs were given to familiarize the subjects with the tasks.
A
1 s
\ , Eye Position
Chair Position
\ 20
\
Chair Velocity
\
/ \
40 /s
/
| 'otf
Chair-fixed Target
'on
Visual and vestibular stimuli In six patients, visually guided and visually remembered saccades were studied for comparison with their vestibular remembered saccades. The procedures were similar to those described previously (Crawford et al., 1989; Lueck et al., 1990). Visual targets (squares sustaining 1.3° of visual angle) were generated by a computer controlled video-projection system onto a screen placed 130 cm in front of the subjects. Subjects sat on the chair with the head fixed, in the dimly illuminated room. For visually guided saccades, subjects maintained fixation on a central target for 800 ms; this was switched off at the same time that one of the lateral targets at ±7.5°, 15° or 25° was randomly illuminated. Subjects were instructed to look at these targets as accurately as possible. For visual remembered saccades, one of the lateral targets was randomly flashed for 200 ms whilst the subjects were requested to maintain fixation on the central target. The central target was extinguished 500 ms after the flash and at this point the subjects were asked to look, as accurately as possible, at the remembered position of the lateral flashed target. Two seconds after the extinction of the central light, the position of the previously flashed target was illuminated for 2 s so that, on re-fixation, a direct measure of the error could be obtained. As this study was concerned with motion perception and eye movement accuracy, the parameters considered were: (i) the amplitude of the initial saccade towards the remembered target; (ii) the final position of the eye, as measured immediately before illumination of the real target. These parameters were expressed as gains, i.e. as a ratio to the amplitude of the stimulus. Non-parametric statistical tests were used from a PC-based package (CSS Statistica): MannWhitney 'IT test, Wilcoxon matched pairs tests and the Friedmann ANOVA, with a significance level of P < 0.05.
Results Vestibular and cervical stimuli Figure 2A shows typical recordings during the vestibular paradigm VS in a control subject. In all illustrations, upwards deflections in eye, head, whole body and trunk movement
20°
F
Earth-fixed Target
J
B Eye Position
20°
20
Chair-fixed Target Earth-fixed Target
'off'
Fig. 2 (A) Raw recordings from a normal subject during the whole body rotation (vestibular) task. In this and all other illustrations, upwards deflections in eye, chair or head traces indicate rightwards motion and downwards deflections leftwards motion. In this case, when the chair-fixed target is switched 'off. a single saccade takes the eye to its final eye position (F). C = the corrective saccade to refixate the earth-fixed target, a measure of the inaccuracy in the task. (B) Raw recordings in a normal subject in the cervical paradigm. After trunk (chair) rotation to the right, the subject attempts to locate its position with a large primary saccade and a small (-2°) secondary saccade. When the chair-fixed target is illuminated a corrective saccade indicates the error in the task.
recordings indicate rightwards motion; downwards deflections indicate leftwards motion. During whole body rotation of 20° to the left no eye movements were seen as the subject appropriately fixates upon the chair-fixed target. When the chair-fixed LED was extinguished ('off'), -1 s after stopping the chair, the subject produced a single saccade of 18° to the right (F) attempting to locate the position from where rotation started. On illumination of the earth-fixed target, a small corrective saccade occurred (C); such small inaccuracies were common in the normal subjects (Table 1). Figure 2B shows an example from the cervical CS
Vestibular saccades in Parkinson's disease Table 1 Summary of findings in the Parkinson's disease patients and in the normal
1427
controls
Gain of final eye position
Gain of initial saccade Patient (n = 14)
Normal (n = 16)
Patient
Normal
Vestibular 10° 20° 30° 40°
0.749+0.271 0.666+0.182 0.704+0.185 0.673+0.217
1.072+0.253 (0.003) 0.999+0.130(0.000) 0.908+0.153 (0.003) 0.847+0.164 (0.036)
1.106+0.135 0.906+0.094 0.853+0.114 0.815+0.089
1.141+0.225 1.007+0.126 (0.016) 0.911+0.130 0.834+0.142
Cervical 10° 20° 30° 40°
0.694+0.230 0.722+0.260 0.753+0.205 0.894+0.124
1.041+0.230(0.003) 0.956+0.198(0.009) 0.930+0.112(0.016) 0.850+0.087
1.002+0.102 0.932+0.103 0.876+0.126 0.908+0.093
1.130+0.104 (0.012) 1.027+0.164 0.928+0.083 0.865+0.117
Passive 10° 20° 30° 40°
0.882+0.335 0.805+0.279 0.826+0.225 0.786+0.235
1.145+0.380 1.031+0.220(0.045) 1.010+0.127(0.025) 0.966+0.107
1.108+0.227 1.066+0.026 1.015+0.063 0.947+0.054
1.164+0.185 1.079+0.116 1.010+0.078 0.962+0.117
Active 10° 20° 30° 40°
0.819+0.547 0.798+0.253 0.778+0.239 0.789+0.314
1.113+0.375 1.129+0.203 (0.002) 0.987+0.224 (0.040) 0.951+0.067
1.171+0.131 1.100+0.093 1.011+0.090 0.974+0.031
1.220+0.207 1.131+0.139 1.062+0.078 0.977+0.052
Probability levels reaching statistical significance between patients and controls are shown in parentheses.
paradigm in a normal subject. The subject's trunk was rotated some 20° to the right whilst his head remained stationary in space. As the earth-fixed target was switched 'off, the subject made a saccade in the same direction of chair rotation attempting to match eye and trunk position. Again, a small error was detected as illustrated by the corrective saccade when the chair-fixed target was switched 'on'. Eye-movement recordings following combined 'active' and 'passive' head turns (combined vestibular/cervical stimulation) in the control group were similar to those illustrated in Figs 2 and 3 and are therefore not shown. Figure 3 displays raw recordings from all four paradigms in a Parkinson's disease patient. As the results in all testing conditions were very similar they will be described collectively. It can be seen that, following the motion stimuli, the saccades produced were multiple-step, i.e. consisted of three or more saccades to attain final eye position. In spite of this, the eye position finally achieved accurately matched the preceding rotational stimulus, as confirmed by the virtual absence of corrective saccades on illuminating ('on') the visual target. In the case shown, the gain of the initial saccade was -0.40, a value much lower than that seen in the normal subject illustrated in Fig. 2, whereas the gain of the final eye position approached unity. The number of multiple-step saccades present, defined as those consisting of three or more saccades to attain final eye position, was determined. In Parkinson's disease patients they occurred in 42.4% of all vestibulo-cervical trials, whereas they were infrequent in normal subjects (14.1%). In normal
controls, final eye position was usually achieved in a single saccade (55.5%), or with one corrective saccade (30.4%). All the results in patients and control subjects are shown in Table 1 and Fig. 4A-D. The mean gain of initial saccade amplitude and final eye position in control subjects decreased as stimulus displacement increased, although this tendency was more frequently seen in VS and CS paradigm rather than in VS+CS paradigms (aspects of the normal response can be found in Nakamura and Bronstein, 1993). The amplitude of the initial saccade [initial saccadic gain, Fig. 4) in both VS and CS paradigms was considerably lower in patients than in controls; this difference was statistically significant (except at 40° in CS paradigm). Initial saccadic gains were also lower in Parkinson's disease when compared with the control group in the combined VS+CS experiments; this difference was significant at 20° and 30° of stimulus amplitude. Probability (f) values can be found in Table 1. In the patients, the mean initial saccadic gain in the combined VS+CS paradigms was higher than that in VS or CS paradigms, but not reaching statistical significance. In contrast, the gain of final eye position (Fig. 4) in patients closely resembled that of controls in all paradigms and, apart from two exceptions (20° in VS and 10° in CS paradigms), there was no statistical difference between subject groups. In the normal control group, there was no statistical difference between initial and final eye position gain, whereas in the Parkinson's disease group there was a significant difference between initial and final eye position gain (P < 0.05). In Fig. 4 the performance of each individual patient,
1428
T. Nakamura et al.
Cervical
Vestibular 1s
Chair-tixed Target I ' o t f '
'on1
Chair-fixed Target
'on' Earth-fixed Target
Earth-fixed Target
I 'off
Active
Passive
20°
20°
Head-fixed target
'Off'
Earth-fixed Target
Head-fixed target
'Off'
Earth-fixed Target
Fig. 3 Examples of saccades following vestibular, cervical and combined vestibular/cervical (passive and active) stimuli in a patient with Parkinson's disease. Although saccades were 'multiple-step', final eye position matched the amplitude of the preceding rotational stimulus in all conditions (note the small, or absence of, corrective saccades when targets were switched 'on').
compared with the mean and standard deviation (SD) of the normal group, can be seen. The amplitude of the initial saccade in all paradigms was very variable amongst different patients, with the more severely affected patients (solid circles) usually showing lower gains. Since there were only four severely affected patients, statistical analysis within the Parkinson's disease group was not performed. In contrast, the gain of final eye position was usually within the normal SD, regardless of clinical severity; this trend was present in all paradigms. Four patients had asymmetrical clinical manifestations. In these patients the gains towards or away from the more symptomatic side were compared, but no consistent differences were found. Similarly, a comparison of saccadic gains between patients with or without treatment showed no significant differences.
Visual and vestibular stimuli Figure 5 shows the data for the six patients who underwent vestibular and visually remembered, as well as visually guided saccades. Initial and final saccadic gains are presented for remembered saccades but only initial gain can be computed for visually guided saccades because the target remains visible throughout. A non-parametric, Friedmann analysis of variance showed that there were significant differences within these data (P < 0.002), which were further investigated with the Wilcoxon matched pairs test. Of note, the gain of the initial saccade was significantly lower during the vestibular (0.60) than during the visual (0.81) remembered test (/> = 0.02) (Fig. 5). This finding could not be explained by differences in duration to hold in memory the location of the target (0.5 s in visually remembered saccades but variable in vestibularly remembered
Vestibular saccades in Parkinson's disease
Vestibular initial gain
final gain
1.4
1.4
°
1.2 1.0 •
o
]° *\
J J
0
_
Cervical initial gain
•
final gain
1.4 -
1.4 •
1.2 -
1.2
1.2 •
1.0 -
1.0 •
1.0 •
0
o
0
0.8 -
0.8 -
0.8 -
0.6 •
0.6 -
0.6 -
0.6 •
0.4 •
0.4 •
0.4 -
0.4
0.2 •
0.2
0.2 •
0.2 -
0.8 •
0
e
§
§•
8
8
0.0
0.0
0.0
10° 20° 30° 40° Chair Displacement
initial gain •
1.2
•
1.4
10° 20° 30° 40° Chair Displacement
8 1.0
>
•
o
0.8
-
I
"I
J
•
0.4 -
1.4 -
•
1.0
\o
1.2
1.4
-.
•
0.8
Hi
1.0
•
• •
1.2
0.6
0.6 -
0.4
0.4 •
0.2 •
10° 20° 30° 40° Head Displacement
°
~s O o
8
*\ • 0
1.0
^r
0.8
0 #
0.6 0.4 •
C
0.2 •
0.0
0.0
^
0
-
0.2 i r 10° 20° 30° 40° Head Displacement
o o
-
0.8 -
•
0.2 • nn
final gain o
03
°
0
10° 20° 30° 40° Chair Displacement
o
1 §
o
0.6
initial gain
1.2 -
CP
18
1 J°
Active
final gain o
I
0.0
10° 20° 30° 40° Chair Displacement
Passive 1.4
1429
T
10° 20° 30° 40° Head Displacement
0.0
10° 20° 30° 40° Head Displacement
Fig. 4 Saccades during the vestibular, cervical, and combined vestibular/cervical passive and active paradigms: summary of results in the control group (mean and SD are represented by a solid triangle and vertical bars, respectively) and in the parkinsonian patients. Open circles are patients mildly affected and solid circles patients more severely affected (Hoehn-Yahr III and IV). Results are expressed as gain, i.e. the ratio between eye displacement and chair or head displacement. The horizontal dotted line through unity indicates the 'ideal' error-free performance. On the left of each part is shown the amplitude of the initial (primary) saccade; on the right thefinaleye position.
saccades) because in Parkinson's disease patients the initial saccadic gain of vestibular remembered saccades occurring after 'storage' time shorter than 0.5 s was not statistically different to those following storage longer than 0.5 s (storage time was defined as the interval between end of chair motion and chair-fixed target extinction). Nor was there a significant correlation between storage time (range 0.2-3.4 s) and initial saccadic gain (r = 0.048, P = 0.35, n = 382). The mean gains of initial saccades during vestibular and visually remembered tasks were lower than their respective final saccadic gains [vestibular 0.60 and 0.87 (P = 0.02); visual 0.81 and 0.94 (P = 0.02)]. The initial gain of visually
remembered saccades (0.81) was lower than that of visually guided saccades (0.88) but this difference was just outside the significance level (P = 0.07). There were no statistically significant differences between the final gain of vestibular and visual remembered saccades and visually guided saccadic gain, all of these -0.90 (Fig. 5).
Discussion The experiments undertaken had two interrelated aims: to quantify vestibulo-cervical perception of head rotation in Parkinson's disease and to find out if the abnormality of
T. Nakamura et al.
1430
1.1 1.0 0.9 o CD o
0.8
8 0.7 u o
*" 0.6 0.5 0.4 Vestl
VestF
Visl
VisF
VisG
Fig. 5 Saccadic gains in six patients with Parkinson's disease. Means and SD are presented, from left to right of thefigure,for initial vestibular remembered saccadic gain (Vestl), final vestibular remembered gain (VestF), initial visual remembered gain (Visl), final visual remembered gain (VisF) and visually guided gain (VisG).
'remembered' saccades in Parkinson's disease is selective for visual memory. Memory-contingent or 'remembered' saccades to imaginary positions in space have been described following visual (Wurtz and Hikosaka, 1986; Crawford et al., 1989) and vestibular stimulation (Bloomberg et al., 1988, 1991). Saccades in the cervical experiment are called 'remembered' because of the similarity with the vestibular remembered paradigm (Nakamura and Bronstein, 1993). This nomenclature will be retained for convenience but, because neck afferents carry tonic as well as dynamic signals (Richmond and Abrahams, 1979), it is plausible that the programming of such saccades is based not only on memorized information but also on direct detection of neck angle. The main finding in the vestibulo-cervical experiments is that, although patients with Parkinson's disease produced saccades of initial low amplitude, they nevertheless achieved a normal final eye position. The latter implies that the actual detection of head angular displacement (subserved by vestibulo-cervical receptors, and conceivably by an 'efference copy' signal during the active movement experiment) is normal. This is inconsistent with previous suggestions that vestibular function is abnormal in patients with Parkinson's disease (Reichert et al., 1982; White et al., 1983a). Whilst undoubtedly some reactions of vestibulo-proprioceptive origin are abnormal in Parkinson's disease, e.g. tilting reactions and postural responses (Martin, 1965, 1967), the results of our experiments are evidence against a primary sensory/perceptual contributing factor. These conclusions are also supported by recent experiments using galvanic, vestibularly induced body sway in which Parkinson's disease
patients showed normal, or even enhanced, sway reactions (Pastor et al., 1993). In spite of achieving a normal final eye position, the initial saccades following the rotational stimuli were abnormal in Parkinson's disease. The saccades had a low initial gain and were often multiple-step, irrespective of whether the stimulus was purely vestibular, cervical or combined both sensory modalities during passive and active head turns. Previous studies have reported that saccades to the memorized position of visual targets (visual remembered saccades) in Parkinson's disease show a similar abnormality, with low initial saccadic amplitude but normal final eye position (Crawford et al., 1989; Lueck etai, 1990, 1992a). Such findings were interpreted in the light of experiments in the monkey, demonstrating the existence of saccadic-related cells in the basal ganglia (Hikosaka and Wurtz, 1985; Joseph and Boussaoud, 1985; reviewed in Wurtz and Hikosaka, 1986; Hikosaka, 1991). Of significance, some of the cells in caudate nucleus and substantia nigra pars reticulata changed their firing rate prior to saccades made to the remembered spatial location of a briefly exposed target but not when the target remained visible. The interpretation was that the abnormality of visual remembered saccades in Parkinson's disease was due to involvement of such cells, again supported by experimental work showing that visually elicited saccades were relatively spared in comparison with visually remembered saccades after muscimol injections into the substantia nigra (Hikosaka and Wurtz, 1985). However, recent experiments in the monkey have shown that dopamine-antagonist agents locally applied to the dorsolateral prefrontal cortex interfere with visual memory-guided saccades but not visually guided saccades (Sawaguchi and Goldman-Rakic, 1991). It is therefore possible that frontal lobe involvement is primarily responsible for the abnormality of remembered saccades found in Parkinson's disease. In agreement with previous findings, our patients showed a lower gain of initial saccades to a visually remembered target than to a permanently visible target (Crawford et al., 1989; Lueck et al., 1990, 1992a). Of note, the experiments reported here demonstrated that in Parkinson's disease the initial gain of vestibular and cervical remembered saccades was also reduced. These findings therefore indicate that the 'remembered' saccade abnormality in Parkinson's disease is not specific to the visual modality. This, in turn, implies that the abnormality must be mediated by dysfunction in common neural systems. One possibility is that the activity of the memory-contingent cells in the basal ganglia or frontal cortex is not modality-specific but, to our knowledge, this has not yet been investigated in experimental animals. It has previously been suggested that Parkinson's disease patients are abnormally dependent on visual input for the generation of a normal saccade (reviewed in Kennard and Lueck, 1989; Lueck et al., 1992a; Stell and Bronstein, 1994). The experiments comparing visual and vestibular remembered saccades showed that the initial saccade was more inaccurate (hypometric) in the vestibular paradigm (though the final eye
Vestibular saccades in Parkinson's disease position was normal in both cases). It is noteworthy that visual remembered saccades do have a visual position cue, albeit well before the saccade is made. Vestibular remembered saccades have no such visuo-positional cue and it is therefore possible that the absence of this visual input is responsible for the added degradation found in the vestibular paradigm. The findings therefore extend the previous concept that Parkinson's disease patients are dependent on visual information for motor control (Flowers, 1976; Cooke et al., 1978; Bronstein et al., 1990) to the domain of memoryguided tasks involving different sensory modalities. In conclusion, the experiments showed that vestibulocervical perception of head and neck angular displacement is normal in Parkinson's disease. The finding that during the vestibulo-cervical experiments the remembered saccades were hypometric and multiple-step indicates that the abnormality of visual remembered saccades previously described in Parkinson's disease is not confined to the visual modality. In turn this implies that the 'remembered saccade' cells of the basal ganglia may have a wider role than previously assumed. The finding that vestibular remembered saccades were more hypometric than visual remembered saccades indicates that the increased visual dependence in Parkinson's disease patients includes the domain of memorydriven tasks.
1431
Flowers KA. Visual 'closed-loop' and 'open-loop' characteristics of voluntary movement in patients with Parkinsonism and intention tremor. Brain 1976; 99: 269-310. Hikosaka O. Role of the forebrain in oculomotor function. [Review]. Prog Brain Res 1991; 87: 101-7. Hikosaka O, Wurtz RH. Modification of saccadic eye movements by GABA-related substances. II: Effects of muscimol in monkey substantia nigra pars reticulata. J Neurophysiol 1985; 53: 292-308. Joseph JP, Boussaoud D. Role of the cat substantia nigra pars reticulata in eye and head movements. I: Neural activity. Exp Brain Res 1985; 57: 286-96. Kennard C, Lueck CJ. Oculomotor abnormalities in diseases of the basal ganglia. [Review]. Rev Neurol (Paris) 1989; 145: 587-95. Lueck CJ, Tanyeri S, Crawford TJ, Henderson L, Kennard C. Antisaccades and remembered saccades in Parkinson's disease. J Neurol Neurosurg Psychiatry 1990; 53: 284-8. Lueck CJ, Tanyeri S, Crawford TJ, Henderson L, Kennard C. Saccadic eye movements in Parkinson's disease: 1. Delayed saccades. Q J Exp Psychol [A] 1992a; 45: 193-210. Lueck CJ, Crawford TJ, Henderson L, Gisbergen JAM Van, Duysens J, Kennard C. Saccadic eye movements in Parkinson's disease: II. Remembered saccades—towards a unified hypothesis? Q J Exp Psychol [A] 1992b; 45: 211-33. Martin JP. Tilting reactions and disorders of the basal ganglia. Brain 1965; 88: 855-74.
References Bloomberg J, Jones GM, Segal B, McFarlane S, Soul J. Vestibularcontingent voluntary saccades based on cognitive estimates of remembered vestibular information. Adv Otorhinolaryngol 1988; 41: 71-5. Bloomberg J, Melvill Jones G, Segal B. Adaptive modification of vestibularly perceived rotation. Exp Brain Res 1991; 84: 47-56. Bronstein AM, Kennard C. Predictive eye movements in normal subjects and in Parkinson's disease. In: Gale AG, Johnson F, editors. Theoretical and applied aspects of eye movement research. Amsterdam: North-Holland, 1984; 463-72. Bronstein AM, Kennard C. Predictive ocular motor control in Parkinson's disease. Brain 1985; 108: 925-40. Bronstein AM, Hood JD, Gresty MA, Panagi C. Visual control of balance in cerebellar and parkinsonian syndromes. Brain 1990; 113: 767-79. Cooke JD, Brown JD, Brooks VB. Increased dependence on visual information for movement control in patients with Parkinson's disease. Can J Neurol Sci 1978; 5: 413-15. Corin MS, Elizan TS, Bender MB. Oculomotor function in patients with Parkinson's disease. J Neurol Sci 1972; 15: 251-65.
Martin JP. The basal ganglia and posture. London: Pitman Medical, 1967. Nakamura T, Bronstein AM. Perception of neck rotation assessed by 'remembered saccades'. Neuroreport 1993; 4: 237-9. Pastor MA, Day BL, Marsden CD. Vestibular induced postural responses in Parkinson's disease. Brain 1993; 116: 1177-90. Rascol O, Clanet M, Montastruc JL, Simonetla M, Soulier-Esteve MJ, Doyon B, et al. Abnormal ocular movements in Parkinson's disease. Evidence for involvement of dopaminergic systems. Brain 1989; 112: 1193-214. Reichert WH, Doolittle J, McDowell FH. Vestibular dysfunction in Parkinson's disease. Neurology 1982; 32: 1133-8. Richmond FJR, Abrahams VC. Physiological properties of muscle spindles in dorsal neck muscles of the cat. J Neurophysiol 1979; 42: 604-17. Sawaguchi T, Goldman-Rakic PS. Dl dopamine receptors in prefrontal cortex: involvement in working memory. Science 1991; 251: 947-50. Shibasaki H, Tsuji S, Kuroiwa Y. Oculomotor abnormalities in Parkinson's disease. Arch Neurol 1979; 36: 360-4.
Crawford TJ, Henderson L, Kennard C. Abnormalities of nonvisually guided eye movements in Parkinson's disease. Brain 1989; 112: 1573-86.
Stell R, Bronstein AM. Eye movement abnormalities in extrapyramidal diseases. In: Marsden CD, Fahn S, editors. Movement disorders 3. Oxford: Butterworth Heinemann, 1994; 88-113.
DeJong JD, Jones GM. Akinesia, hypokinesia, and bradykinesia in the oculomotor system of patients with Parkinson's disease. Exp Neurol 1971; 32: 58-68.
Tanyeri S, Lueck CJ, Crawford TJ, Kennard C. Vertical and horizontal saccadic eye movements in Parkinson's disease. Neuroophthalmol 1989; 9: 165-77.
1432
T. Nakamura et al.
Ventre J, Zee DS, Papageorgiou H, Reich S. Abnormalities of predictive saccades in hemi-Parkinson's disease. Brain 1992; 115: i 147_AS White OB, Saint-Cyr JA, Sharpe JA. Ocular motor deficits in Parkinson's disease. I: The horizontal vestibulo-ocular reflex and its regulation. Brain 1983a; 106: 555-70. White OB, Saint-Cyr JA, Tomlinson RD, Sharpe JA. Ocular motor deficits in Parkinson's disease. II: Control of the saccadic and
smooth pursuit systems. Brain 1983b; 106: 571-87. ... . „ „ u . , , „ n . c.. , . .. . ., . . . . Wurtz RH, Hikosaka O. Role of the basal ganglia in the initiation of saccadic eye movements. [Review]. Prog Brain Res 1986; 64: 175-90.
Received March 17, 1994. Revised May 17, 1994. Accepted July 19, 1994
