Brain (1992), 115, 1343-1358

A QUANTITATIVE STUDY OF EYE AND HEAD MOVEMENTS DURING SMOOTH PURSUIT IN PATIENTS WITH CEREBELLAR DISEASE by JOHN A. WATERSTON, GRAHAM R. BARNES and MADELEINE A. GREALY (From the MRC Human Movement and Balance Unit, National Hospital for Neurology and Neurosurgery, London, UK)

SUMMARY Eye and head movements were analysed during smooth pursuit in 16 patients with various forms of cerebellar disease. Smooth pursuit gain was reduced across all frequencies and velocities of target motion for the patient group as a whole, during both sinusoidal and pseudo-random target motion. The graded breakdown in the pursuit response, as pseudo-random target motion became less predictable, was of a similar magnitude in patients and controls, implying that the predictive pursuit mechanisms were intact in these patients. During head-free pursuit, when vestibulo-ocular reflex (VOR) suppression was necessary, performance was not significantly different from that observed during head-fixed pursuit in the patient group. This finding is similar to that noted in control subjects, and is consistent with the observation that the VOR gains associated with head movements in darkness were similar in the patient and control groups. The deficits in pursuit and VOR suppression in patients with cerebellar disease therefore represent a decrease in gain in the closed-loop visual feedback pathways with apparent sparing of the predictive pathways. INTRODUCTION

The critical role of the cerebellum in smooth eye movement control has been well established in numerous animal (Westheimer and Blair, 1974; Zee et al., 1981) and human studies (von Noorden and Preziosi, 1966; Baloh et al., 1975, 1986; Zee et al., 1976; Dichgans et al., 1978; Monday et al., 1978; Avanzini et al., 1979; Estanol et al., 1979; Wennmo et al., 1983; Ell et al., 1984; Furman et al., 1986; Yamamoto et al., 1988; Pierrot-Deseilligny et al., 1990). Cerebellar lesions and degenerations usually cause disruption of smooth tracking by reducing maximal smooth eye velocity, resulting in the characteristic broken eye displacement trace consisting of low velocity slow phases punctuated by catch-up saccades. These features have been a consistent finding in all of the above studies where pursuit performance has been assessed using predictable target motion. In constructing control systems models of the smooth pursuit system, it has generally been assumed that at least two feedback loops are responsible for optimal tracking performance (Dallos and Jones, 1963; Robinson, 1965; Young, 1971; Yasui and Young, 1984; Becker and Fuchs, 1985; Barnes and Asselman, 1991). It has been postulated that a closed loop, visual feedback pathway corrects eye velocity according to retinal velocity error, but such a simple system cannot explain how it is possible to maintain Correspondence to: Dr G. R. Barnes, MRC Human Movement and Balance Unit, National Hospital for Neurology and Neurosurgery, Queen Square, London WC1N 3BG, UK. © Oxford University Press 1992

1344

J. A. WATERSTON AND OTHERS

steady state pursuit of a moving target with the low levels of phase error observed experimentally in the face of the established delays in the visual pathways. An additional predictive pathway is needed to enable the oculomotor system to produce eye movements which attempt to match future estimates of target motion based on information derived from the previous target trajectory. These predictive strategies are able to eliminate the inherent phase lags resulting from time delays in the visual and oculomotor pathways, and would therefore be expected to make an important contribution to pursuit performance, particularly as target frequency increases. Cerebellar disease can result in the most severe deficits of smooth pursuit, but predictive behaviour in the pursuit system has not been specifically addressed in patients with cerebellar disease. Recordings from the primate cerebellar flocculus indicate that a predictive pattern of firing behaviour can be observed in neurons which encode eye velocity (Lisberger and Fuchs, 1978; Noda, 1986), but the exact source of this activity is unclear. It might be anticipated that, if the predictive mechanisms were disturbed by cerebellar lesions, performance would be relatively preserved at low target frequencies, as it is in some patients with mild cerebellar disease (Zee et al., 1976). However, because it is difficult to examine prediction in isolation, it has not been conclusively demonstrated that the predictive mechanisms are functioning normally in these patients. Phase differences between target and eye velocity give some indication of predictive behaviour, but a number of problems arise in the interpretation of phase changes in patients with low gain pursuit. A reduction in gain of itself will cause an increase in the degree of phase lag, thereby masking any predictive effects which might be occurring, and documentation of phase changes becomes less accurate when gain falls, particularly at higher target frequencies. Using an alternative approach to this problem, it has been demonstrated that pseudo-random stimuli can be used with great effect to study the breakdown in the predictive behaviour of the smooth pursuit response (Barnes et al., 1987; Barnes and Ruddock, 1989). The predictability of a pseudo-random stimulus composed of the sum of two or more sinusoids is determined not by the number of frequency components contained within the stimulus, but by the frequency of the highest frequency component. Low frequency pseudo-random stimuli are very predictable but, if the stimulus contains a frequency component which is above a critical level of 0.4 Hz, the target motion becomes less predictable and a breakdown in the gains of all the lower frequency components occurs (Barnes et al., 1987). When the velocity of the high frequency component is increased with respect to the velocity of the other components, further breakdown occurs. We have therefore performed a study of smooth pursuit in patients with various forms of cerebellar disease using pseudo-random and sinusoidal stimuli to investigate predictive smooth pursuit function and to quantify the frequency and velocity characteristics of the pursuit deficits in these patients. When both head and eye movements are used to track smooth target motion, accurate gaze control depends on adequate suppression of the vestibulo-ocular reflex (VOR) which, if unopposed, would produce eye movements of opposite polarity to head movement. The ability to suppress the VOR generated by either rotational or caloric stimuli is reduced in patients with cerebellar disease and is generally considered to correlate directly with the smooth pursuit deficit, implying a common pathway involving the cerebellar connections for both functions (Zee et al., 1976; Dichgans et al., 1978; Henriksson

SMOOTH PURSUIT IN CEREBELLAR DISEASE

1345

ex al., 1984; Baloh et al., 1986). Therefore, we have also investigated head-free pursuit and compared the measures of smooth combined eye and head, or gaze velocity, with those obtained under head-fixed conditions during pseudo-random target motion tracking. Because the underlying level of VOR gain will be expected to influence head-free pursuit performance, an estimate of the VOR gain at the frequencies used in the pseudorandom stimulus was also obtained in each of the subjects, using various real and imaginary target paradigms during active head movements and passive whole body rotation. METHODS Subject selection A total of 16 patients with cerebellar disease (mean age 44.8 yrs, range 23—66 yrs) resulting from a number of different pathological processes were studied (Table 1), though not all patients did every test. Patients had either pure cerebellar involvement in the form of hereditary or sporadic cerebellar degeneration with no clinical or investigative evidence of other neurological involvement, or cerebellar signs in combination with other central neurological features. Patients in the latter group, particularly those with multiple sclerosis, were chosen because the major clinical features were consistent with dysfunction in cerebellar pathways, and all had clinical evidence of bilateral involvement. The patients with multiple sclerosis were examined to ensure that there were no other signs of oculomotor involvement, such as ophthalmoplegia or saccadic slowing, which might have influenced the eye movement recordings. In the one patient with a mild internuclear ophthalmoplegia, recordings were taken from the unaffected eye. Patients with dementia or uncontrolled head tremor were excluded from the study. A similar number of naive, age-matched controls performed each experiment (mean age 45.8 yrs, range 23—69 yrs). All subjects participated with informed consent and the experiments were approved by the local ethics committee. Apparatus The subjects were seated in the centre of a darkened room in front of a semicircular screen of radius 1.5 m. Eye movements were recorded using an infrared limbus reflection technique (Ins 6500 system, Skalar Medical) with a resolution of 10 min of arc and a linear range of at least ±20 degrees. The eye

TABLE 1. SUMMARY OF PATIENTS Case w. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Age/sex 43/F 42/F 62/F 56/F 65/F 52/M 58/M 28/M 23/F 26/F 39/F 53/F 39/F 33/M 66/F 31/M

Diagnosis Cerebellar degeneration Cerebellar degeneration Cerebellar degeneration Cerebellar degeneration Cerebellar degeneration Cerebellar degeneration Spinocerebellar degeneration Spinocerebellar degeneration Spinocerebellar degeneration Multiple sclerosis Multiple sclerosis Multiple sclerosis Friedreich's ataxia Friedreich's ataxia Multiple system atrophy Hereditary spastic paraparesis with cerebellar features

Other oculomotor signs Gaze evoked nystagmus Gaze evoked nystagmus Gaze evoked nystagmus Gaze evoked nystagmus Gaze evoked nystagmus, square wave jerks Gaze evoked nystagmus, saccadic dysmetria Gaze evoked nystagmus Gaze evoked nystagmus, slow saccades Square wave jerks, saccadic dysmetria Gaze evoked nystagmus, R. internuclear ophthalmoplegia Gaze evoked nystagmus Gaze evoked nystagmus Square wave jerks, saccadic dysmetria Gaze evoked nystagmus, square wave jerks Gaze evoked nystagmus, hypometric saccades Gaze evoked nystagmus

1346

J. A. WATERSTON AND OTHERS

movement recorders were mounted on a helmet assembly which was attached firmly to the subject's head. A single turn potentiometer, attached to the top of the helmet via a flexible assembly, was used to record head displacement. During experiments performed under head-fixed conditions, a head clamp and chin rest were used to stabilize the head. The target consisted of a small white cross within a circle, the diameter of which subtended 70 min of arc at the eye, and its motion was controlled by a motor-driven mirror situated above the subject's head. A motorized turntable (Toennies, 200 Nm) was used to assess vestibular function during sinusoidal oscillation. Eye movement calibrations were performed prior to each subtest. Experimental design Experiment (i). Pursuit of predictable sinusoidal target motion was investigated under head-fixed conditions at frequencies of 0.20, 0.40, 0.82, 1.20 and 1.56 Hz. For each set of stimuli the peak amplitude was set at ± 5 , 10 or 15 degrees, thus producing peak stimulus velocities between 6 and 147 degrees/s. A total of 15 patients performed this test. Experiment (ii). Head-fixed performance was examined during pursuit of a pseudo-random stimulus composed of the sum of three or four sinusoids of frequency 0.11, 0.24, 0.37 and 1.56 Hz (Fig. 1). The velocity of the three lowest frequency components remained constant at 8 degrees/s, while the velocity of the 1.56 Hz component was varied as a ratio (velocity ratio, VR) of the lower frequency velocities between 0 and 2, the 1.56 Hz component therefore being absent when the velocity ratio was zero. A total of 14 patients performed this test. Experiment (Hi). A direct comparison of head-fixed and head-free pursuit was performed in 10 patients. The pursuit stimuli were the same pseudo-random waveforms used in experiment (ii). All conditions were presented in random order. For the head-free conditions the subject was asked to use natural, combined eye and head movements to track the target (Fig. 1). Experiment (iv). The vestibular response was recorded during active head movements and passive whole body turntable rotation. Each of the three lowest frequencies used in the pseudo-random stimulus (0.11, 0.24, 0.37 Hz) was presented as a sinusoidal waveform with a peak velocity of 16 degrees/s to approximate the root mean square velocity of the pseudo-random waveform. Using combined eye and head movements, subjects were initially instructed to follow the target in time with a stationary, sinusoidally modulated tone located above the subject's head. Subsequently, the target was extinguished and the subject was asked to follow an imaginary target in complete darkness, moving his head in time with the tone (active VOR). We have shown previously that this paradigm does not result in any significant degree of VOR suppression in normal subjects (Waterston and Barnes, 1991). Head movements were monitored on an oscilloscope by an observer and the subject was prompted if either the timing or the amplitude of head movement was inappropriate. The target frequencies were presented in randomized fashion, but the imaginary target conditions always followed the real target condition. Because some patients had difficulties performing the test, results from eight patients only were used for analysis. The sequence of real and imaginary target tracking was then repeated using a stationary fixation target (earth-fixed target, EFT and imaginary earth-fixed target, IEFT); the results were analysed from six patients. Vestibular responses during sinusoidal whole-body rotation (passive VOR) were recorded in 11 patients at the same frequencies and velocities, while subjects performed mental arithmetic to maintain alertness. On the basis of these results, subjects who performed the head-free task were divided arbitrarily into a low VOR gain (gain at two or more frequencies

A quantitative study of eye and head movements during smooth pursuit in patients with cerebellar disease.

Eye and head movements were analysed during smooth pursuit in 16 patients with various forms of cerebellar disease. Smooth pursuit gain was reduced ac...
940KB Sizes 0 Downloads 0 Views