Smooth pursuit eye movements in schizophrenics—What constitutes quantitative assessment?

J. psychiat. Rex, Vol. 22, No. 3, pp. 195-206, 1988. Printed in Great Britain.

0022-3956/88 $3.00+ .OO 0 1988 Pergamon Press plc

SMOOTH PURSUIT EYE MOVEMENTS IN SCHIZOPHRENICS:

QUANTITATIVE MEASUREMENTS WITH THE

SEARCH-COIL TECHNIQUE

S. LEVIN*, A. LUEBKE~, D. S. ZEE~, T. C. HAIN?, D. A. ROBINSON~ and P. S. HOLZMAN*

*School of Medicine, Harvard University, Cambridge, MA 02138, U.S.A. and tSchoo1 of Medicine, Johns Hopkins University, MD 21218, U.S.A.

(Received 17 August 1987; revised 26 January 1988; in final form 26 February 1988)

Summary-Eye movements of five schizophrenic and five normal subjects were measured with the magnetic-field search-coil technique. Subjects followed targets moving smoothly at various speeds, either unpredictably in a step-ramp fashion or predictably in a triangular wave. The tracking stimulus was either a small dot or a large, richly-textured image that occupied a large portion of the visual field. Tracking by schizophrenics was abnormal; it was punctuated by catch-up saccades that corrected for smooth following movements of inadequate velocity. We did not, however, find saccadic intrusions, such as square wave jerks. Under all tracking conditions steady- state gains (eye velocity/target velocity) and, in the case of step-ramps, average acceleration in the first 120 ms were lower in patients than in normal subjects. The differences were most pronounced for tracking of the small target, moving at the highest speed tested (30 degree/s), in the nonpredictable, step-ramp waveform. With this stimulus mean steady-state gain was 0.36 (SD +- 0.12) for the schizophrenic patients and 0.73 (SD f 0.11) for the normal subjects. When the target was changed to the large-field stimulus or moved in a predictable (triangular-wave) fashion, tracking improved in both patients and normal subjects, and even more so when these features were combined.

INTRODUCTION

EYE MOVEMENTS during tracking of a smoothly moving target are impaired in a large percentage of patients with schizophrenia (see reviews in LIPTON et al., 1983; HOLZMAN, 1985). Most early studies recorded eye movements using the electrooculogram, a method which, because it is inherently noisy, can make it difficult to measure the relative con- tributions of saccades and smooth pursuit in the overall pattern of tracking. Moreover, the early methods of grading tracking performance did not describe the exact nature of the smooth pursuit abnormality, either qualitatively or quantitatively. Accordingly we decided to use an accurate and sensitive method of recording eye movements-the magnetic-field search-coil technique-to study the ocular tracking disorders of schizophrenic patients.

A commonly used measure of pursuit performance is the closed-loop gain: the ratio of eye velocity to target velocity in the steady-state condition when eye and target velocity are momentarily constant. The pursuit gain of normal subjects, which has usually been investigated in individuals experienced in eye movement studies, is reported to be between 0.9 and 1.0 (ROBINSON et al., 1986). It is known, however, that pursuit capabilities are affected by many factors, including target velocity, target size and the predictability of target motion. In previous investigations of pursuit eye movements in schizophrenia the

195

196 S. LEVIN et al.

targets have usually been small, their trajectories predictable-sinusoids or triangular waves-and only a few velocities were tested. Consequently, we examined pursuit of schizophrenic and of normal subjects both with targets moving in a predictable, triangular wave form and with targets moving in a non-predictable, step-ramp wave form. We also varied both target speed and size to determine if a particular set of stimulus parameters might facilitate the identification of a pursuit abnormality in schizophrenia. For the step-ramp stimuli we also measured eye acceleration in the first 120 ms of tracking which is a measure of the open-loop performance of the pursuit system and is a more sensitive index of dysfunction than are closed-loop measurements (LISBERGER et al., 1981). For normal subjects we chose people inexperienced in eye movement studies with no prior knowledge of or expectations of ocular motor performance since it was clear from our experience that the performance of ourselves and of our colleagues was not representative of the normal population.

Subjects METHODS

We examined five patients with chronic schizophrenia and five normal subjects (see Table 1). All patients met criteria for a DSM-III diagnosis of chronic schizophrenia. Diagnoses were assigned by an experienced clinician (one of the authors, S.L.) who inter- viewed all the subjects and reviewed the patients’ hospital records. All patients were receiving neuroleptic medication (phenothiazines) at the time of testing.

All subjects were free of non-psychiatric medical illness and clinical neuro-ophthal- mological examination was normal. Normal subjects were free of psychiatric illness in themselves or their immediate families though one subject (No. 7) had a niece who had been hospitalized after a suicide attempt. No subjects had undergone previous testing of eye movements or had any special knowledge or interest in the study of eye movements.

TABLE 1. POPULATION DATA AND COMPARISON OF TRACKING PERFOR-

MANCEFORSMALLTARGETSMOVINGAT 30 DEGREE/S.FORQUALITATIVE RATINGS 1 IS WORST AND 5 IS BEST. FOR QUANTITATIVE RESULTS THE

MEANVALLJESOFGAINARESHOwN

Qualitative

rating Steady-state gains

Age Sex triangular Triangular Step-ramp

22 M 4.3 0.89 0.57 25 M 3.6 0.81 0.29 52 F 2.3 0.73 0.32 30 M 2.9 1.08 0.31 26 M 3.9 0.87 0.33

6

8 9

10

40 F 4.5 0.91 53 F 2.4 0.69 49 M 3.3 0.85 27 M 4.2 0.93 47 F 3.8 0.94

0.90 0.68 0.60 0.75 0.73

EYE MOVEMENTS IN SCHIZOPHRENICS 197

Eye movement recordings Eye movements were recorded with the magnetic-field search-coil technique. In this

method the subject wears a scleral contact lens in the form of an annulus that adheres to the eye just outside the junction of the cornea and the sclera. The annulus contains an embedded coil of wire that picks up a sinusoidal voltage generated by high-frequency (30 kHz), horizontal and vertical magnetic fields. Phase detection of this signal creates voltage outputs with a sensitivity of 9.25 degree (at the full-scale settings used here) and a bandwidth of O-l kHz. The outputs were recorded on a pen recorder that limited eye position bandwidth to O-70 Hz.

Target stimuli The tracking stimulus was either a small spot of light, subtending about 1 degree of arc,

or a large rectangular field subtending about 30 x 20 degrees of arc horizontally and verti- cally. The latter was a slide of a Jackson Pollock painting primarily consisting of randomly distributed dots on a white background. Both stimuli were rear-projected onto a large (1 x 1.75 m) translucent, featureless screen located at 1.0 m in front of the subject. No special fixation dot was present on the large-field stimulus though subjects were asked to focus their attention on a cluster of dots in the center of the slide. A mirror galvanometer allowed the visual stimuli to be moved horizontally. Target motion was under computer control. The unpredictable stimulus was a step-ramp designed to eliminate the need for a catch-up saccade in the initiation phase of pursuit tracking. If a target steps to one side and immediately begins moving at a constant velocity in the other direction, so that it crosses its initial position in about 200 ms, subjects often can initiate pursuit without a catch-up saccade. Target velocities were 5, 10, 20 and 30 degree/s presented in a random order and direction, with random timing. There were five trials at each target velocity. The predictable stimulus was a target moving in a triangular-wave fashion with a peak-to- peak amplitude of 40 degrees and velocities of 10, 20, 30, 40, 50, and 60 degree/s. Six cycles at each target velocity were elicited.

Data analysis The eye-position signal was differentiated with an analog filter of 40 Hz bandwidth. Eye

position, eye velocity and target position were displayed on a chart recorder. Data were analyzed by hand, directly from the on-line chart records run at high speed.

A steady-state gain (eye velocity/target velocity) was calculated during the triangular- wave tracking by measuring eye velocity as the eye passed through the primary position and during step-ramp tracking when eye velocity became nearly constant, which was usually about 750 ms after the onset of tracking. An average eye acceleration was calculated by approximating a straight line to the eye velocity trace for the first 120 ms of tracking and measuring its slope. Pursuit latencies were measured in the step-ramp paradigm using the time of onset of the step-ramp and the time of onset of smooth tracking as determined from the eye-velocity trace. Prediction was measured in the triangular-wave paradigm using the difference between the time when the target and the eye velocity changed sign.

S. LEVIN et al.

SCHIZOPHRENIC PATIENT NORMAL CONTROL

20*[[

/ 20. --

[ +

198

FIG. 1. Tracking of targets moving in a triangular-wave (upper panels) and step-ramp (lower panels) fashion, for a schizophrenic patient and for a normal subject. Target velocity was 30 degree/s. Note the excess of catch-

up saccades in the tracking record of the schizophrenic patient, especially for step-ramp tracking.

RESULTS

Figure 1 shows examples of step-ramp and triangular-wave tracking of a small target for a schizophrenic and for a normal subject. When the target moved as an unpredictable step-ramp, low-gain pursuit was clearly revealed in the schizophrenic patient by the numerous catch-up saccades. On the other hand, when the target moved predictably the difference in performance between the patient and the normal subject was much less.

Response to step-ramps Steady-state gain. Steady-state gains (called gains for short) for the responses to small

and large stimuli moving at different velocities with a step-ramp waveform are shown in Fig. 2 and also tabulated in Table 2. The values of gain were lower for patients than for control subjects. Mean gain (for all velocities) for tracking of the small target was 0.55 for patients and 0.87 for normal controls, values which were significantly different (Student’s t-test for independent means (P < 0.0001)). It is clear, however, from Fig. 2 that the defect was most pronounced at the higher target speeds and at 20 and 30 degree/s every patient but one had a gain considerably below that of the normal controls with the poorest performance. With the large-field stimulus there was an improvement in both groups but

Ens MOVEMENTS IN SCHIZOPHRENICS 199

STEP- RAMP SF TARGET

1.2-

1.0 -

b 3 0.6 -

k il! cr, 0.6 -

L

9 k 0.4 -

w

0.2-

0’ ’ ’ I 1 5 10 20 30

STEP-RAMP LF TARGET

-NC

“SZ

t

tt I I I 5 10 20 30

TARGET VELOCf TY fdeg/sec/

FIG. 2. Mean values of steady-state gain (eye velkarget vel) for the response to the small (SF) and the large-field (LF) target moving in a step-ramp fashion. Values for individual subjects are plotted as a function of target velocity. Schizophrenic patients (SZ) are represented by open triangles; normal subjects (NC) by filled circles.

Shaded area represents the range of values for normal subjects.

TABLE 2. MEAN VALUES (AND STANDARD DEVIATION) FOR STEADY-STATE GAIN, ACCELER- ATION AND LATENCY FOR STEP-RAMP TARGETS. EACH VALUE IS BASED UPON FIVE TRIALS

5”/s Stimulus speed

lO”/s 2O”/s 3O”/s

Steady-state gain

Small-field Schizophrenics Normals

Large-field Schizophrenics Normals

Acceleration (degree/s/s)



Latencies (ms)



0.66 (0.19) 0.96 (0.03)

0.82 (0.10) 1 .Ol (0.06)

28 ( 3) 34 (13)

32 (11) 36 ( 8)

239 (13) 210 (45)

221 (18) 188 (43)

0.70 (0.10) 0.95 (0.09)

0.90 (0.06) 1.02 (0.05)

43 ( 5) 48 (16)

50 (15) 60 (10)

203 (28) 163 (16)

148 (18) 145 (31)

0.48 (0.11) 0.83 (1.0)

0.78 (0.02) 0.94 (0.06)

48 ( 9) 62 (16)

70 (21) 85 (11)

188 (20) 172 (23)

119 (10) 130 (33)

0.36 (0.12) 0.73 (0.11)

0.67 (0.06) 0.85 (0.05)

61 ( 9) 79 (18)

88 (21) 112 (12)

188 (31) 190 (48)

121 ( 7) 147 (33)

200 S. LEVIN et al.

more so in the schizophrenic subjects so that the difference between patients and normals was less obvious than with the small target.

Analysis of variance with groups as the between subject factor, and the size of stimulus and the velocity as repeated measure, showed a significant main effect of group on gain (F(1)=63.61, P=0.0001), and there was a significant interaction of group by size of target (F(1) = 13.29, P= 0.0006). An inspection of the means with respect to this interaction indicated that the large-field condition had a greater effect on improving the gains of patients than it did on improving the gains of normal subjects. This finding may be due to a ceiling effect for the normal subjects.

Accelerufion. We also measured average acceleration in the first 120 ms of step-ramp tracking both with the small dot and with the large-field target (Fig. 3). Schizophrenic patients had significantly lower accelerations than normal subjects. Mean accelerations for the two subject groups are presented in Table 2. Analysis of variance for acceleration, with groups as a between subject factor, and size of target and velocity as repeated measure, revealed a significant main effect of group (P(l) = 5.8, P < 0.0426). There were also significant main effects for size of target and velocity (F(1) = 40.68 for field, F(1) = 78.09 for velocity, P < 0.0001 for both). As Fig. 3 shows, these significant main effects reflect the findings that for both groups acceleration increased as a function of both target velocity and size of stimulus. Furthermore, for both groups, acceleration increased significantly under large-field conditions when target velocity was 20-30 degree/s, as indicated by a significant size of target by velocity interaction (F(3) = 5.78, P < 0.0016).

Predictable tracking Gain for tracking a target moving in a predictable triangular waveform was also signifi-

cantly lower for patients than for normal subjects, although the effect was smaller than

STEP-RAMP SF TARGET

120 -

l Norm01 Control

A Schizophrenic

60-

--/

I I 2o 5

I I 10 20 30

STEP-RAMP LF TARGET

TARGET VELOCITY f&g/sec)

FIG. 3. Group mean values of average acceleration in the first 120 ms of step-ramp tracking with either a small field (SF) or large field (LF) stimulus, plotted as a function of target velocity. Schizophrenic patients are

represented by open triangles; normal subjects by filled circles.

EYE MOVEMENTS IN SCHIZOPHRENICS 201

STEP-RAMP TRIANGULAR 1.2 -

1.0 - .--d.,.

1 o.61 w : *

0.6

B

i: &

0.4 1”\’

NC D--d LF NC O---O SF

SZ-LF

0.2 - SZ-SF

o- ’ ’ I 1 I I I I I I

5 10 20 30 10 20 30 40 50 60

TARGET VELOCITY fdeghecl

FIG. 4. Group mean values of steady-state gain for targets moving in a step-ramp or a triangular-wave fashion, plotted as a function of target velocity and field size. Schizophrenic patients (SZ) are represented by triangles; normal subjects (NC) by circles. Open points are the small-field (SF) target, closed points the large-field (LF)

target. Further data are in Table 2.

for step-ramps (Fig. 4 and Table 3). An analysis of variance for gain for triangular-wave tracking with groups as a between subject factor, and size of stimulus and velocity as repeated measures, revaled a significant main effect of group (F(1) = 5.52, P < 0.05), of size of stimulus (F(1) = 39.02, PC 0.0001) and of velocity (F(1) = 9.26, P< 0.0001). There was also a significant interaction of group by size of field (F(1) = 8.89, P < 0.0038). An inspection of the relevant means indicates that the increase in gain of the patients for large-field tracking is greater than the improvement of normal subjects but as with the nonpredictable target this may be due to a ceiling effect.

TABLE 3. MEAN GAIN (AND STANDARD DEVIATION) OR FAN LATENCY (AND STANDARD DEVIATION) FOR TRIANGULAR-WAVE TARGETS. LATENCIES REFER TO THE TIME BETWEEN THE REVERSAL OF TARGET AND EYE VELOCITY

AND REFLECT PREDICTION. EACH VALUE IS BASED ON SIX CYCLES OF TRACKING

lO”/s 2O”/s Stimulus speed

3O”/s 4O”/S 5O”/s 6O”/s

Steady-state gain



Latency



0.87 (0.08) 0.99 (0.06)

0.94 (0.04) 1.09 (0.02)

133 (15) 72 (20)

117 (14) 75 (18)

0.86 (0.02) 0.90 (0.17) 0.85 (0.08) 0.95 (0.14) 0.86 (0.10) 0.87 (0.15)

0.88 (0.10) 0.91 (0.03) 0.89 (0.08) 1.06 (0.08) 1.03 (0.07) 1.00 (0.02)

123 (25) 94 (21) 88 (29) 83 (31) 81 (18) 74 (28)

114 (12) 101 (13) 102 (13) 58 (29) 60 (27) 70 (16)

0.76 (0.10) 0.84 (0.19)

0.85 (0.08) 0.97 (0.04)

91 (23) 79 (28)

94 (4) 76 (25)

0.71 (0.11) 0.81 (0.16)

0.90 (0.04) 0.95 (0.04)

90 (23) 73 (34)

93 (17) 64 (18)

202 S. LEVIN et al.

The gain of schizophrenics was significantly higher for targets moving in a triangular waveform than in a step-ramp fashion (Fig. 4). An analysis of variance for gain, with groups as a between subject factor and type of target motion (step-ramp or triangular) and size of stimulus as repeated measures, revealed a significant main effect for group (F(1) =45.22, P < 0.0001). For all subjects gain during triangular-wave tracking was higher than gain during step-ramp tracking as indicated by a significant main effect for type of target motion (F(1) = 34.81, P < 0.0001). The main effect of size of stimulus was also significant (F(1) = 46.10, P < O.OOOl), indicating that, overall, subjects had higher gains under the large-field condition. There was, in addition, a significant three-way interaction of group by size of target by type of target motion (F(1) = 12.12, P < 0.0003). This interaction reflects the fact that the increase in gain with a large-field stimulus was particularly pronounced for patients tracking targets that moved in a step-ramp fashion.

Latency and anticipation For step-ramp tracking we found no significant difference in latency between patients

and normal subjects (Fig. 5). Mean latencies for the two subject groups with different target size and velocity conditions are presented in Table 2. An analysis of variance of latency with groups as a between subject factor, and size of target and velocity as repeated measures, revealed a significant main effect for size of target (P’(l) = 52.71, P < 0.0001). A significant main effect for velocity (F(3) = 24.27, P < 0.0001) reflects the finding that for all subjects, latencies decreased with higher target velocities (20-30 degree/s).

In order to assess the extent to which subjects could benefit from prediction during triangular-wave pursuit, we measure the mean interval between the time the target velocity and the time the pursuit eye velocity changed direction. The smaller this interval, the

TRIANGULAR

NC*--. LF NCO---O SF

SZLILF SZA--ASF

401 ’ ’ I 5 10 20 ‘V 30 20 30 40 50 60

TARGET VELOCf TY (deg/secl

FIG. 5. Group mean values of latency in response to step-ramp targets and the interval between the time that target and eye velocity changed direction during triangular-wave tracking. Small dot (SF, open data points) or large-field (LF, closed data points) stimuli were used. Schizophrenic patients (SZ) are represented by triangles;

normal control subjects (NC) by circles.

EYE MOWMENTS IN SCHIZOPHRENICS 203

greater the effect of prediction. In tracking small targets, across all velocities, the mean interval for the schizophrenic patients was 104 (* 13) ms, compared to a mean of 72 (k 6) ms for the normal subjects. An analysis of variance with groups as a between subject factor, and size of target and velocity as repeated measures, revealed only a significant group main effect (F(1) = 24.2, P < 0.0012). Neither the main effects for size of stimulus or velocity, nor the interaction effects were significant, suggesting that the effect of predictability was independent of size of stimulus and velocity.

Saccade disturbances The main abnormality in our patient’s tracking records was low-gain pursuit. There

were catch-up saccades (Fig. l), to compensate for the inadequate pursuit movements, but no square-wave jerks or other types of saccades that took the eye away from the target.

Comparison of qualitative ratings and quantitative testing In order to compare our results to previous studies, records of triangular-wave tracking

of a small target moving at 30 degree/s were qualitatively rated by two of the authors (S.L. and P.H.) on a 5-point scale, with 1 representing severely impaired tracking and 5 perfect tracking (LEVIN et al., 1983). These ratings are shown in Table 1. Previous studies have established a score of 3.5, below which tracking is considered to be impaired (LEVIN et al., 1981). By this criterion 40% of our patients (Nos 3 and 4) had impaired tracking. This prevalence is close to the values that have been reported in other studies of patients with schizophrenia (50-60%). For comparison, Table 1 also shows gains for each subject tracking a small target moving in a triangular wave and a step-ramp at 30 degree/s.

DISCUSSION

The main finding in this study was clear cut even though the sample size was small. The small size of the sample, however, does call for replication. Smooth pursuit eye movements of our schizophrenic patients were not as accurate as those of normal subjects. The main abnormality was in the generation of the smooth component; disruption of tracking by inappropriate or inaccurate saccades was not a factor in our patients. These findings require replication in a larger sample.

Differences between schizophrenic and normal subjects were greatest for tracking of a small target, moving at the higher velocities, in a nonpredictable fashion. Under these conditions the performance of all of our patients was worse than that of any of the normal subjects and the performance of all patients except one could be clearly delineated from that of normals. Consequently, these testing conditions should be included when assessing tracking capabilities in schizophrenic patients. The present findings also suggest that prior studies, which primarily used targets moving in a predictable pattern, may have under- estimated the frequency and severity of tracking deficits in schizophrenic patients. While the gains of schizophrenic patients were significantly different from those of normal controls when targets moved in a predictable fashion, the size of the difference between the two groups was less than with the non-predictable stimulus. There was also a significant difference between the two groups in predictive capacity. The degree to which schizophrenic patients anticipated the motion of a predictable target was less than that of normal subjects.

It may be argued that the task of following an unpredictable step-ramp target impli-

204 s. LEVIN et al.

cates different cognitive functions than does a predictable oscillatory target. In particular, vigilance and focussed attention may be triggered to a greater extent by the requirement to track an unpredictable step-ramp. Since vigilance and simple reaction time are known to be impaired in most schizophrenic patients, the differences seen between the normals and the schizophrenics in the unpredictable step-ramp condition may simply reflect those cognitive impairments, and a dysfunction of the smooth pursuit system may be an incor- rect inference. If this argument is valid, one should perhaps expect to see differences in latencies between normal and schizophrenics in the step-ramp procedure. But, as Table 3 and Fig. 5 show, all latencies are within normal limits, and differences between groups are statistically non-significant. In fact, the more difficult the target is to follow (i.e. the faster its velocity and the smaller it is), the better the reaction time of all subjects, including the schizophrenic patients. Therefore, it is probably reasonable to attribute the finding of diminished gain in the schizophrenic group to the smooth pursuit system rather than to artifactual variables sometimes referred to as generalized deficit (CHAPMAN and CHAPMAN, 1973).

A number of other features of pursuit were similar in normal and in schizophrenic subjects. For step-ramp tracking both steady-state gain and average initial acceleration were higher for the large-field target. Steady-state gain was also higher for triangular-wave than for step-ramp tracking. Latencies for smooth pursuit during step-ramp tracking were less at higher target speeds.

Findings of low-gain pursuit in schizophrenics have been previously reported by YEE

et al. (1987) and ABEL et al. (1986). The degree of deficit, however, may have been under- estimated in those studies since only predictable targets were used. The finding of SCHMD BURGK et al. (1982) also imply that gain is lower in schizophrenic patients though the measure of gain used by these investigators was the amplitude of the smooth component divided by the amplitude of the entire eye track including saccades.

The finding that the pursuit gain of our patients (and normal subjects) improved with large-field targets is difficult to interpret without additional experimental manipulations of target characteristics. Nevertheless, there are some indications that the effect may be related to attention. (VAN DEN BERG and COLLEWIJN (1986) found that pursuit gain for normal subjects was highest when they pursued a large multicontoured pattern without a fixation spot compared to either a fixation spot alone or a spot superimposed on a large stationary background pattern. Other reports have shown that any manipulation of the target that increases its “attentional load”-such as numbers superimposed on a pen- dulum, targets with randomly changing colors, and targets with changing letters-results in some, though not complete, improvement of smooth tracking in schizophrenic patients (SHAGASS et al., 1979; HOLZMAN et al., 1978; LIPTON et al., 1983). Thus, our results are compatible with the notion that impairment of smooth pursuit in patients with schizophrenia is related to a disturbance of involuntary attention that is partially corrected by an external mobilization of attention (LEVIN et al., 1981).

For stimuli moving in a predictable fashion, both schizophrenic and normal subjects showed an ability to anticipate the motion of the target and their steady-state gain was higher than in response to an identical stimulus moving at the same speed in a nonpredictable fashion. This finding is consistent with the report of BUZZA and SCHM~D (1986) that target predictability is associated with improved pursuit in normal subjects. Yet, in the

Em MOVEI&ENTS IN SCHIZOPHRENICS 205

present study, schizophrenic patients showed less anticipation than did normal subjects. Recent experimental evidence in monkeys suggests that the frontal lobes may be import- ant for generating eye movements that anticipate the motion of the target (LYNCH and ALLISON, 1985). The abnormality of prediction that we found in our patients is another piece of evidence in favor of frontal lobe dysfunction in schizophrenia (LENIN, 1984).

Our finding that the steady-state gain in the normal control group was significantly lower than 1 .O for many of the stimulus conditions may raise doubts about how representative our control sample was. In most previous investigations of pursuit tracking, however, subjects were highly practiced and perhaps more motivated to track well. In the present study all subjects were unpracticed and unaware of expected eye movement performance. The true range of normal pursuit gains must be determined in a larger sample of unpracticed and naive-from an ocular motor point of view-normal subjects. Furthermore, in all normal subjects the family history must be scrutinized for psychiatric disease since there is evidence that pursuit tracking ability and a propensity for psychiatric illness are genetically-linked traits (MATHER, 1985; MATTHYSSE et al., 1986). These issues must be addressed in a separate study.

Table 1 shows the relationship between steady-state gain and the quantitative ratings of the records. With the exception of one patient (No. 4) there was good agreement for triangular-wave tracking at 30 degree/s. A linear regression for the other nine data points of gain versus rating had a correlation coefficient of 0.90. This suggests that a low gain requires frequent catch-up saccades leading to a low qualitative score. Patient No. 4 tended to lead the target (his gain was slightly greater than 1.0) and accordingly made several back-up saccades which contributed to his low qualitative score. For step-ramps the correlation was poor since all gains for the patients dropped large amounts according to each patient’s ability to use prediction to improve tracking.

A final point must be made regarding the possibility that medications, which all of our patients were taking, were responsible for the pursuit anomaly. Previous studies (LEW

et al., 1983; LEVY et al., 1984) have concluded that the integrity of pursuit eye movements is not affected by neuroleptic treatment. The effects of drugs upon pursuit gain or pursuit acceleration, however, have not been carefully investigated, especially with nonpredictable targets over a range of target velocities, though a preliminary report by ABEL et al. (1986) indicates that unmedicated schizophrenic patients had low pursuit gain and that neuroleptic or antidepressant medication, other than lithium, did not affect pursuit gain. Our results also suggest that a generalized neuroleptic effect on pursuit gain is unlikely since, under some target conditions, the performance of the patients and the normal control subjects was not significantly different. Nevertheless, we cannot conclude that the deficits shown by our patients did not reflect an effect of medications that became particularly apparent under conditions where smooth tracking becomes the most difficult even in normal individuals, i.e. small targets, moving at high speeds, in a nonpredictable waveform.

Acknowledgements-This research was supported by NIMH Research Scientist Development Award (MHO0460) to S.L., a General Medical Scientist training grant (GM 07057) to A.L., NIH research grants EY01849 (DSZ), EYO0598 (DAR), EY05505 (TCH), a core grant (BYO1765), and NIMH research grants MH31340 and MH31154. We thank Larry Tune, M.D., for referring patients, Donna Jenkins and Elaine Ken- nedy for assisting in data scoring and statistical analyses and Corena Bridges for assisting in data analysis and preparation of illustrations.

206 S. LEVIN et al.

REFERENCES

ABEL, L. A., HERTLE, R. W. and MELTZER, H. Y. (1986) Smooth pursuit in psychiatric disease: are there specific deficits in patients and their relatives? Sot. Neurosci. Absfr. 12, 1091.

BUZZA, A. and SCHMID, R. (1986) Velocity characteristics of smooth pursuit eye movements to different patterns of target motion. Exp. Brain Res. 63, 395401.

CHAPMAN, L. J. and CHAPMAN, J. P. (1973) Disordered Thought in Schizophrenia. Appleton-Century-Crofts, New York.

HOLZMAN, P. S. (1985) Eye movement dysfunction and psychosis. Int. Rev. Neurobiol. 27, 127-205. HOLZMAN, P. S., LEVY, D. L. and PROCTOR, L. R. (1978) The several qualities of attention in schizophrenia.

J. psychiat. Res. 14, 99-110. HOLZMAN, P. S., SOLOMON, C. M., LEVIN, S. and WATERNAUX, C. S. (1984) Pursuit eye movement dysfunctions

in schizophrenia: family evidence for specificity. Archs gen. Psychiat. 41, 136-139. LEVIN, S. (1984) Frontal lobe dysfunctions in schizophrenia--I. Eye movement impairments. J. psychiat. Res.

18, 27-55. LEVIN, S., LIPTON, R. and HOLZMAN, P. S. (1981) Pursuit eye movements in psychopathology: effects of target

characteristics. Biol. Psychiat. 16, 255-267. LEVIN, S., JONES, A., STARK, L., MERRIN, E. and HOLZMAN, P. S. (1983) Identification of abnormal patterns in

eye movements of schizophrenic patients. Archs gen. Psychiat. 39, 1125-I 130. LEVY, D. L., LIPTON, R. B., HOLZMAN, P. S. and DAVIS, J. M. (1983) Eye tracking dysfunction unrelated to

clinical state and treatment with haloperidol. Biol. Psychiat. 18, 813-819. LEW, D. L., LIPTON, R. B., YASILLO, N. J., PETERSON, J., PANDEY, G. and DAVIS, J. M. (1984) Psychotropic

drug effects and smooth pursuit: a summary of recent findings. In Theoretical and Applied Aspects of Eye Movement Research (Edited by GALE, A. G. and JOHNSON, F.), pp. 497-505. Elsevier-North-Holland.

LIPTON, R. B., LEW, D. L., HOLZMAN, P. S. and LEVIN, S. (1983) Oculomotor function in psychiatric patients: a review. Schizophr. Bull. 9, 13-32.

LISBERGER, S. G., EVINGER, C., JOHANSON, G. W. and FUCHS, A. (1981) Relationship between eye acceleration and retinal image velocity during foveal smooth pursuit in man and monkey. J. ieurophysiot. 46,229-249.

LYNCH, J. C. and ALLISON, J. G. (1985) A quantitative study of visual uursuit deficits followina lesions of the frontal eye fields in rhesus monkeys..Soc_ Neurosci. Absh. 11, 473. -

MATHER, J. A. (1985) Eye movements of teenage children of schizophrenics: a possible inherited marker of susceptibility to the disease. J. psychiat. Res. 19, 523-532.

MATTHYSSE, S., HOLZMAN, P. S. and LANGE, K. (1986) The genetic transmission of schizophrenia: application of Mendelian latent structure analysis to eye tracking dysfunctions in schizophrenia and affective disorder. J. psychiat. Res. 20, 57-65.

ROBINSON, D. A., GORDON, J. L. and GORDON, S. E. (1986) A model of the smooth pursuit eye movement system. Biol. Cybern. 55, 43-57.

SCW~ID-BURGK, W., BECKER, W., DIEKMANN, V., JERGENS, R. and KORNHUBER, H. H. (1982) Disturbed smooth pursuit and saccadic eye movements in schizophrenia. Archs psychiat. Nervenkr. 232, 381-389.

SHAGASS, C., ROEMER, R. A. and AMADEO, M. (1976) Eye tracking performance and engagement of attention. Archs gen. Psychiat. 33, 121-125.

VAN DEN BERG, A. V. and COLLE~WN, H. (1986) Human smooth pursuit: effects of stimulus extent and of spatial and temporal constraints of the pursuit trajectory. Vision Res. 26, 1209-1222.

YEE, R. D., BALOH, R. W., MADDER, S. R., LEVY, D. L., SAKALA, S. M. and HONRUBIA, V. (1987) Eye movements in schizophrenia. Invest. Ophthalmol. Vis. Sci. 28, 366-374.

Documents

Smooth pursuit eye movements in schizophrenics—What constitutes quantitative assessment?