You are on page 1of 8

Psychopharmacology (2001) 157:284–291

DOI 10.1007/s002130100803


S. B. Hutton · T. J. Crawford · H. Gibbins

I. Cuthbert · T. R. E. Barnes · C. Kennard
E. M. Joyce

Short and long term effects of antipsychotic medication

on smooth pursuit eye tracking in schizophrenia
Received: 27 November 2000 / Accepted: 9 April 2001 / Published online: 20 July 2001
© Springer-Verlag 2001

Abstract Rationale: Smooth pursuit abnormalities have catch-up saccades than both first-episode schizophrenic
been observed in antipsychotic naive first-episode pa- patients and controls. First-episode patients were im-
tients, suggesting that they are intrinsic to the illness. paired relative to controls only on the measure of veloci-
However, it is not clear whether these abnormalities are ty gain. There were no differences between antipsychotic-
as severe as those observed in more chronic patients. In naive and treated first-episode patients. Antipsychotic-
addition, although research suggests that there are no free chronic patients were significantly less impaired on
short-term effects of conventional antipsychotic medica- velocity gain than matched continuously treated chronic
tion, the effects of long-term antipsychotic medication patients. These results were not influenced by group dif-
on smooth pursuit eye movements are relatively un- ferences in age and symptom severity. Conclusions: These
known. Objectives: To determine the short and long term results show that: 1) the main index of smooth pursuit, ve-
effects of antipsychotic medication on the smooth pur- locity gain, is impaired early in the course of schizophre-
suit performance of first-episode and chronic patients nia; 2) whereas velocity gain is unaffected by short-term
with schizophrenia. Methods: We compared the smooth (weeks) medication, it is worsened by chronic (years)
pursuit performance of antipsychotic-treated and untreat- treatment; 3) other indices of smooth pursuit, catch-up
ed first-episode and chronic schizophrenic patients with saccades and ability to predict target movement, are ad-
healthy controls using a comprehensive range of perfor- versely influenced by illness chronicity rather than medi-
mance measures. This included velocity gain, the num- cation.
ber, type and size of intrusive and corrective saccades,
and the average time between the change in direction of Keywords Smooth pursuit · Eye tracking · Schizophrenia ·
the target and the change in direction of the eye move- Antipsychotic medication · Eye movement
ment, a measure of subjects’ ability to predict target
movement. Results: Chronic schizophrenic patients had
significantly reduced velocity gain, took longer to re- Introduction
spond to the change in target direction and made more
The study of oculomotor abnormalities in schizophrenic
patients provides a potentially powerful tool, both for
S.B. Hutton · H. Gibbins · I. Cuthbert · T.R.E. Barnes understanding the neuropathology underlying schizo-
E.M. Joyce (✉) phrenia, and for identifying genetically relevant pheno-
Department of Psychiatry, typic markers (Clementz 1998). Eye movements can be
Division of Neuroscience and Psychological Medicine,
Imperial College School of Medicine, St Dunstan’s Road, broadly categorised as either fast saccadic movements,
London W6 8RP, UK which serve to foveate an object of interest, or slower
e-mail: smooth pursuit movements, which allow a moving object
Tel.: +44-20-88467336, Fax: +44-20-88467372 to remain foveated. Abnormal smooth pursuit eye track-
C. Kennard ing has been extensively documented in chronic schizo-
Department of Sensorimotor Control, phrenic patients and their first-degree relatives, and is
Division of Neuroscience and Psychological Medicine, currently considered to be an important indicator of ge-
Imperial College School of Medicine, St Dunstan’s Road,
London W6 8RP, UK netic vulnerability for the illness (see Levy 1994; Hutton
and Kennard 1998 for reviews). The observation of ab-
T.J. Crawford normal smooth pursuit in antipsychotic naive first-epi-
Mental Health and Neural Systems Research Unit,
Department of Psychology, Lancaster University, sode patients has strengthened arguments that this abnor-
Lancaster LA1 4YF, UK mality represents a core and enduring trait deficit in

schizophrenia, which is present at an early stage in the treated, chronic patients than in a group of treated first-
illness (Campion et al. 1992; Sweeney et al. 1994; episode patients. None of these studies was able to dis-
Hutton et al. 1998a). However, it is not clear whether the entangle the effect of long-term medication from that of
nature and severity of the smooth pursuit dysfunction is illness duration.
the same in first-episode patients as it is in chronic pa- The aim of this study was to assess systematically the
tients. This is an important issue, as neurodevelopmental effects of illness chronicity and antipsychotic medication
models of schizophrenia would predict that deficits ob- on smooth pursuit function in schizophrenia. In order to
served at first-episode ought to be qualitatively and examine the effects of illness chronicity, we compared
quantitatively similar to those observed in chronic pa- the eye tracking performance of both first-episode and
tients. If, however, schizophrenia involves progressive chronic schizophrenic patients with a group of healthy
neurodegenerative pathology, then the severity and/or controls. We used a comprehensive range of perfor-
nature of smooth pursuit dysfunction would be expected mance measures, including smooth pursuit velocity gain
to change over the course of the disorder. (the ratio of eye velocity to target velocity) and a mea-
In order for the effects of illness chronicity on smooth sure of the predictive component of smooth pursuit. In
pursuit performance to be accurately assessed, the poten- addition, we identified and quantified all intrusive and
tially confounding effects of antipsychotic medication corrective saccades made during pursuit as these can
must also be addressed. There are several reasons why provide important information concerning the nature of
antipsychotic medication might be expected to influence the pursuit dysfunction (Friedman et al. 1992a). All pa-
smooth pursuit performance. Firstly, the common action tients were tested with the same stimulus conditions, and
of antipsychotic drugs is dopamine receptor blockade the measures were computed using identical procedures.
(Creese et al. 1976) and dopamine innervation of the In order to examine the contribution of short- and
striatum is though to regulate fronto-striatal output path- long-term antipsychotic treatment to smooth pursuit dys-
ways which include oculomotor circuitry (Alexander et function, and any possible interaction this may have with
al. 1986). Although this oculomotor circuitry is tradition- illness chronicity, the first-episode patients were divided
ally considered to be involved in saccadic control, neuro- into those who were antipsychotic-naive at the time of
imaging evidence suggests smooth pursuit and saccades testing, and those who had been receiving antipsychotic
are controlled by similar neural areas (Berman et al. medication. The chronic patients were divided into those
1999; O’Driscoll et al. 2000). Secondly, in patients with who had been antipsychotic free for at least 6 months,
schizophrenia, the dopamine receptor blockade caused and those who had been treated continuously. In order to
by antipsychotic medication often leads to parkinsonian establish any differences in the effects of treatment with
side effects, and smooth pursuit deficits have been ob- conventional and atypical antipsychotic drugs, we com-
served in patients with Parkinson’s disease (White et al. pared first episode patients receiving traditional antipsy-
1983; Waterston et al. 1996). These deficits can be re- chotics with those receiving atypical antipsychotics.
versed with L-dopa treatment (Gibson et al. 1987). Final-
ly, single doses of conventional antipsychotic medication
also lead to smooth pursuit dysfunction in healthy con- Materials and methods
trols (King 1994; Malaspina et al. 1994).
The majority of studies addressing medication effects Subjects
conclude that antipsychotic treatment does not have a The study participants were recruited from Tower Hamlets, River-
major impact on smooth pursuit performance in schizo- side, Kingston and Richmond, Merton Sutton and Wandsworth
phrenic patients (Holzman et al. 1974; Levy et al. 1983; Health Authorities, London, UK. The Local Research Ethics Com-
Siever et al. 1986; Spohn et al. 1988; Litman et al. 1989; mittees for these Health Authorities reviewed and approved the
Hutton et al. 1998a). However, these studies have only study. All patients provided written informed consent.
Patients with first-episode schizophrenia (n=67) were part of a
addressed the effects of relatively short-term treatment or larger ongoing prospective study of first-episode schizophrenia in
treatment withdrawal and few studies have addressed the West London where patients are recruited from a geographically
long-term effects of antipsychotic medication. Campion defined catchment area. Consenting patients were assessed by re-
and colleagues (1992) compared antipsychotic-naive pa- search clinicians using the Scale for the Assessment of Positive
Symptoms (SAPS; Andreasen 1984b), the Scale for the Assess-
tients with patients who had been treated for more than ment of Negative Symptoms (SANS; Andreasen 1984a) and the
5 years and found equivalent levels of smooth pursuit Comprehensive Psychopathological Rating Scale. All patients
dysfunction in both groups suggesting that long-term were re-assessed at 6 months and a DSM-IV diagnosis of schizo-
medication is not associated with smooth pursuit dys- phrenia was confirmed for each patient in a diagnostic review with
function. However, Kufferle and colleagues (1990) two psychiatrists, applying the SCID criteria. The velocity gains
of 31 of these patients have been presented previously (Hutton et
found that a group of previously treated patients who had al. 1998a). Patients with chronic schizophrenia (n=36) were iden-
not received medication in the 2 years prior to testing tified from outpatient case-notes by a research clinician. Consent-
were significantly less impaired than a group of patients ing patients were assessed with the SAPS and SANS and DSM-
who had been taking medication during that time. Fur- III-R criteria were applied. The velocity gains of 27 of these pa-
tients have been presented previously (Crawford et al. 1995). The
ther, Sweeney and colleagues (1992), using precise oc- data on the chronic group were reanalysed by SBH for the present
ulographic techniques, found that smooth pursuit perfor- study in order to derive velocity gain and additional performance
mance was significantly more impaired in a group of measures described below. Fifty-four healthy volunteers served as
Table 1 Group details: mean (SD). SAPS Scale for the Assessment of Positive Symptoms, SANS Scale for the Assessment of Negative
Symptoms, CPZ Chlorpromazine equivalents, mg/day

Group n Age SAPS SANS Duration of Duration of CPZ

(years) total total illness (weeks) treatment

Controls 54 26.9 (5.3) – – – – –

Anitpsychotic naive 20 27.6 (7.8) 36.3 (16.6) 28.2b (23.8) 82.2b (102.9) – –
first-episode patients
Antipsychotic-treated 47 24.3 (6.9) 37.5 (19.8) 23.5b (16.2) 61.5b (93.4) 7.32c (5.5) 422.4c (289.2)
first-episode patients
Antipsychotic-free 16 37.5a (11.7) 23.1 (20.8) 15.9 (16.0) 756.9 (726.2) – –
chronic patients
Antipsychotic-treated 20 37.8a (12.0) 28.3 (15.0) 12.7 (17.0) 621.1 (548.7) 656 (582.4) 1216.0 (229.9)
chronic patients
a Significantly different from controls, antipsychotic-treated and antipsychotic-naive first-episode patients
b Significantly different from antipsychotic-treated and antipsychotic-free chronic patients
c Significantly different from antipsychotic-treated chronic patients

controls with no history of psychiatric or neurological disorder in PDP 11/73 computer or an IBM compatible PC. The smooth pur-
themselves and in their first-degree relatives, and no history of al- suit paradigm was preceded by a calibration trial in which nine
cohol or drug abuse. equally spaced LED lights with known horizontal positions were
Patients were considered first-episode if this was their first illuminated sequentially, the subject being asked to fixate each tar-
presentation with a psychotic illness and if they had been receiv- get in turn.
ing antipsychotic medication for no longer than 12 weeks (Lieber-
man et al. 1992). As they were consecutively recruited they were
not selected for medication status; some patients had already be- Data analysis
gun treatment by the time they were identified by our team where-
as others were still antipsychotic-naive. Twenty first-episode pa- When tracking a smoothly moving object, the smooth pursuit and
tients were antipsychotic-naive at the time of testing. Within the saccadic systems work together interactively. A measure of the
47 treated first-episode patients, 19 were receiving the atypical an- smooth pursuit system’s ability to match eye velocity to target ve-
tipsychotic drugs olanzapine (n=12) or risperidone (n=7) at the locity is provided by a measure of velocity gain, the ratio of the
time of testing, and the remainder were receiving conventional an- eye velocity to target velocity. A velocity gain value of 1 indicates
tipsychotics. Of the treated first-episode patients, 36% were re- perfect smooth pursuit performance, values lower than 1 indicate
ceiving concurrent anticholinergic medication. Two patients were that the eye is moving more slowly than the target. The number
receiving benzodiazepines as hypnotics. and type of saccades made during smooth pursuit can also provide
Medication status was a key criterion in the recruitment strate- important information concerning the nature of the underlying eye
gy for the chronic patient group. In order to avoid selection bias, tracking deficit (Friedman et al. 1992a). The saccades made dur-
care was taken to identify treated and untreated patients from the ing smooth pursuits are typically either corrective [catch-up sac-
same population. Thus the majority of patients were attenders at cades (CUS) and back-up saccades (BUS)] or intrusive [anticipa-
an outpatient depot clinic. Antipsychotic free patients were de- tory saccades (AS) and square wave jerks (SWJ)]. Accurately
fined as those who had not attended at least for the past 6 months tracking a smoothly moving object involves predictive processes
but had attended previously. Sixteen patients had been free of anti- (van den Berg 1988). One measure of the predictive component of
psychotic medication for at least 6 months, and 20 patients had smooth pursuit is the time difference (lag) between the target and
been receiving conventional antipsychotics continuously. Of these, the eye changing direction at the end of each ramp. Large lag val-
approximately 46% were also receiving anticholinergic medica- ues indicate poor prediction of target movement.
tion and none were receiving benzodiazepines. We also assessed Smooth pursuit analysis was performed using the Eyemap
premorbid IQ (NART; Nelson 1982) current level of social func- analysis package (Amtech GmbH, Heidelberg, Germany). Saccad-
tioning and parental social class, in order to establish the extent to ic eye movements were identified by the software using accelera-
which the drug-treated and drug-free chronic patients were tion (an increase in eye velocity of 30 degrees per second) and
matched for education and socio-economic status. amplitude (>0.5 of a degree) criteria. Based mainly on the criteria
Duration of psychosis was calculated for each patient, and du- of Abel et al. (1991) and Friedman et al. (1992a), the saccades
ration of treatment was calculated for each treated patient (see Ta- were then classified by the operator (SBH) as CUS, BUS, AS or
ble 1). SWJ. Briefly, CUS were defined as saccades that occurred in the
direction of target motion during smooth pursuit, and which took
the eye from a position behind the target to a position on or near
Apparatus the target. BUS were defined as saccades which occurred in a di-
rection opposite to the target motion and serve to bring the eyes
The smooth pursuit stimulus was a bright red laser spot back pro- from a position ahead of the target to a position on or near the tar-
jected onto a semi-opaque screen. The target oscillated horizontal- get. SWJ were defined as intrusions into smooth pursuit consisting
ly with a triangular waveform of amplitude 22.5 degrees. Four ve- of an initial small saccade in either direction, followed by a short
locities, 10, 20, 30 and 36 deg/s were used, and six full cycles re- period of continued pursuit and terminated by another small sac-
corded at each velocity. Subjects were comfortably seated 1.5 m cade, similar in size to the first, in the opposite direction. Finally,
from the screen and head movements were restrained using an ad- following Ross et al. (1999b), who have argued that small sac-
justable headrest. All paradigms were conducted in the dark. Eye cades (<4 degrees) which move the eye ahead of the target should
movements were recorded using a Skalar IRIS infrared limbus re- be termed “leading saccades”, true AS were defined as large (>4
flection device. A hardware anti-aliasing filter (cut-off frequency degrees) intrusive saccades which took the eye to a position ahead
200 Hz) was used to filter eye position and the sampling rate was of the target, and were followed by a short period of low, or zero
500 Hz. Stimulus display and data sampling were controlled by a velocity gain, and/or a BUS.
Smooth pursuit velocity gain (GAIN) was measured by first (t=1.72, P=0.09) but there were no differences between
identifying and excluding saccadic movements and blinks from these groups for symptom severity (SAPS, t=–0.23,
the data. In each ramp a 50-ms portion of smooth pursuit eye
movement was selected and expressed as peak velocity gain (eye P=0.8; SANS, t=0.94, P=0.35) or duration of illness
velocity/target velocity). This portion was always collected from (t=0.79, P=0.43). Antipsychotic-free chronic patients did
the middle third of each half cycle, to avoid acceleration and de- not differ from their treated counterparts in age (t=0.07,
celeration transients at the beginning and end of each ramp. The P=0.94), symptom severity (SAPS, t=0.8, P=0.43;
ramp was selected by the operator determining the portion of the
ramp in which pursuit was optimal (i.e. the velocity gain was near-
SANS, t=–0.54, P=0.59) or duration of illness (t=–0.54,
est to 1). P=0.59). Nor did the antipsychotic-free and antipsychot-
A temporal measure of smooth pursuit lag (LAG) was calculat- ic-treated patients differ in terms of their premorbid IQ
ed by measuring the time at which the eye movement changed di- (t=0.62, P=0.54), current level of social functioning
rection at the end of each ramp, and subtracting from this the time (t=–0.44, P=0.66) or parental social class (t=–1.03,
at which the target changed direction.
The first-episode patients receiving atypical antipsy-
Statistical analyses chotics did not differ in age (t=0.38, P=0.71), symptom
severity (SAPS, t=–0.39, P=0.70; SANS, t=0.67,
Performance on the measures of GAIN and LAG was compared
between the two groups of schizophrenic patients and the controls P=0.50) or duration of treatment (t=0.7, P=0.95), from
using ANCOVA with target speed (Speed) as a within subjects- those receiving conventional antipsychotics. There was a
factor and Group (first-episode, chronic, control) as a between- trend for first-episode patients receiving atypical anti-
subjects factor. As the chronic patients were older than the first- psychotics to have been ill for longer than patients re-
episode patients and controls, and increased age has been shown
to be associated with decreased smooth pursuit function (Ross et
ceiving conventional antipsychotics (t=–1.89, P=0.07).
al. 1999a), age was entered as a covariate. The number of sac- The mean daily dose of antipsychotic medication was
cades were compared using ANCOVA, with group as the between converted to chlorpromazine equivalents, mg/day. Reli-
subjects factor and age as the covariate. Significant main effects able data were obtained for 46 out of the 47 treated first
and interactions were further explored using post-hoc t-tests. episode patients and 15 out of the 20 treated chronic pa-
The interactions between acute and chronic medication and
disease chronicity were explored by ANCOVA with Treatment tients. Mean daily dose was greater in the treated chronic
(antipsychotic-free/antipsychotic-naive versus antipsychotic-treat- schizophrenic patients than in the treated first-episode
ed) and Chronicity (first-episode versus chronic) as between sub- patients (t=2.32, P=0.02). The duration of treatment was
ject factors. The GAIN and LAG measures were collapsed across also significantly longer in the treated chronic patients
target speed. In order to control for group differences in age and
symptom severity age, SAPS and SANS scores were entered as compared to the treated first-episode patients (t=–7.46,
covariates. The SAPS and SANS scores for three antipsychotic- P<0.001), as was duration of illness (t=–6.82, P<0.001).
treated chronic patients, two antipsychotic-free chronic patients
and one treated first-episode patient were unavailable. For the pur-
poses of the ANCOVA, these patients were assigned SAPS and
SANS scores equal to their respective group means. The results
Comparisons between first-episode patients,
were unchanged when the analysis was repeated excluding these chronic patients and controls
Finally, in the first-episode patients, the effects of conventional The smooth pursuit velocity gain for the three groups is
versus atypical antipsychotic treatment were explored using inde- displayed in Fig. 1. Velocity gain decreased with increas-
pendent t-tests.
ing target speed for all groups [F(3,459)=10.01,
P<0.001]. There was also a significant main effect of
Group [F(2,153)=21.8, P<0.001] and a Group×Speed in-
Subject characteristics

The groups were matched for sex ratio (male/female –

control: 36/18; first episode: 55/12; chronic schizophre-
nia: 25/11, chi square=4.16, P=0.12) Table 1 shows that
the first-episode patients showed a greater severity of
positive and negative symptoms compared with the
chronic patients (first-episode SAPS total score=37.2,
chronic SAPS total score=25.9; t=2.77, P<0.01; first-epi-
sode SANS total score=24.4, chronic SANS total
score=14.2; t=2.63, P<0.05]. The groups also differed in
age [F(2,154)=30.3, P<0.001]. Chronic patients were
significantly older than the first-episode patients
(t=–6.62, P<0.001) and control subjects (t=–5.9,
P<0.001), who did not differ (t=1.4, P=0.17). Fig. 1 Smooth pursuit velocity gain across four target speeds for
There was a trend for the antipsychotic-naive first-ep- controls (circles), first-episode schizophrenic patients (squares)
isode patients to be older than their treated counterparts and chronic schizophrenic patients (triangles)

Fig. 3 Interaction between chronicity (first-episode schizophrenia,

squares versus chronic schizophrenia, triangles) and antipsychotic
drug treatment (antipsychotic-naive/antipsychotic-free AN/AF,
versus antipsychotic-treated AT) for smooth pursuit velocity gain
Fig. 2 LAG across the four target speeds for controls (circles),
first-episode patients (squares) and chronic patients (triangles)
P<0.001], who did not differ [t(117)=1.17, P=0.24].
Table 2 Mean (SD) corrective and intrusive saccades. CUS catch- There were also differences in the number of BUS made
up saccades, BUS back-up saccades, AS anticipatory saccades, [F(2,142)=11.55, P<0.001]. Both chronic [t(79)=3.51,
SWJ square wave jerks P=0.001] and first-episode patients [t(117)=3.75,
P<0.01] made fewer BUS than controls but there was no
difference between the two patient groups [t(90)=0.68,
Controls 83.7 (22.1) 11.4 (7.3) 2.7 (3.7) 5.9 (6.9) P=0.5]. There were no differences between the three
First-episode 79.8 (27.0) 6.8 (5.7)b 3.2 (4.0) 3.1 (5.3)c groups in the number of AS made [F(2,142)=0.23,
Chronic 116.4 (32.1)a 6.0 (4.5)b 3.1 (3.9) 6.2 (7.4) P=0.79]. Finally, there were differences in the number of
a Chronic patients significantly different to first-episode patients
SWJ made [F(2,142)=5.33, P<0.01]. First-episode pa-
and controls tients made fewer SWJ than either controls (t=2.9,
b Chronic and first episode patients significantly different to con- P<0.005) or chronic patients (t=2.66, P<0.01) who did
trols not differ (t=–0.19, P=0.85).
c First-episode patients significantly different to controls and
In summary, chronic schizophrenic patients had sig-
chronic patients
nificantly reduced velocity gain, took longer to respond
to the change in target direction and made more catch up
saccades than both first-episode schizophrenic patients
teraction [F(6,459)=10.5, P<0.001]. The main effect of and controls. First-episode patients were impaired rela-
the covariate, Age, was also significant [F(1,153)=12.4, tive to controls only on the measure of velocity gain.
P=0.01]. Analysis of the main effect of Group demon-
strated that both the first-episode (t=4.51, P<0.001) and
chronic (t=7.19, P<0.001) patients were impaired rela- Interactions between illness chronicity
tive to controls, and chronic patients were additionally and treatment status
impaired relative to the first-episode patients (t=4.32,
P<0.01). The interaction reflects the greater effect of tar- The effects of illness chronicity and antipsychotic treat-
get speed on the chronic patients compared to the first- ment on GAIN are displayed in Fig. 3. Chronic patients
episode patients and controls. had generally more impaired velocity gain than first-epi-
LAG times are displayed in Fig. 2. LAG times de- sode patients [F(1,96)=7.61, P<0.01], and treated patients
creased with increasing target speed for all groups had generally more impaired velocity gain than antipsy-
[F(3,441)=12.75, P<0.001]. The ANCOVA also revealed chotic na / antipsychotic free patients [F(1,96)=
a significant main effect of Group [F(2,147)=5.53, 14.29, P=0.001]. However, these main effects were modi-
P=0.005]. The Group×Speed interaction was not signifi- fied by a significant Chronicity×Treatment interaction
cant [F(6,441)=1.22, P=0.29]. Analysis of the main ef- [F(1,99)=5.22, P<0.05]. The main effect of the covariate
fect indicated that chronic patients had significantly lon- Age was significant [F(1,96)=10.79, P=0.001]. The main
ger LAG times than first-episode patients (t=–5.71, effects of the covariates SAPS and SANS were not sig-
P=0.001) and controls (t=–5.48, P=0.001), who did not nificant [SAPS: F(1,96)=1.15, P=0.29; SANS: F(1,96)=
differ (t=0.15, P=0.89). 0.156, P=0.69]. The difference between antipsychotic-na-
The mean numbers of intrusive and corrective sac- ive and antipsychotic-treated first-episode schizophrenic
cades are displayed in Table 2. The groups differed in the patients was not significant (t=0.91, P=0.37), whereas an-
number of CUS made [F(2,142)=8.75, P<0.001]: chronic tipsychotic-treated chronic patients were significantly
patients made more CUS than first-episode patients more impaired than antipsychotic-free chronic patients
[t(90)=–5.96, P<0.001] and controls [t(79)=–5.38, (t=–3.20, P<0.005). The difference between the antipsy-
Table 3 Oculomotor perfor-
mance in first-episode patients Group GAIN LAG CUS BUS AS SWJ
treated with conventional and
atypical antipsychotics: mean Conventional 0.85 (0.14) 91.9 (27.5) 80.6 (27.8) 7.7 (6.3) 3.7 (4.2) 3.8 (5.4)
(SD) Atypical 0.90 (0.05) 97.5 (25.9) 72.1 (25.8) 5.9 (4.6) 1.9 (3.0) 2.3 (5.7)

jects can switch direction in keeping with the target,

which was also slower in the chronic patients.

Comparison of conventional and atypical antipsychotics

in treated first-episode patients

The effects of conventional versus atypical antipsychotic

medication on smooth pursuit function in treated first-
episode patients are displayed in Table 3. There was a
trend (t=–1.9, P=0.07) for patients taking atypical anti-
psychotic medication at the time of the first episode to
Fig. 4 Interaction between chronicity (first-episode schizophren- be less impaired on the measure of GAIN than the pa-
ics, squares versus chronic schizophrenics, triangles) and antipsy- tients taking conventional antipsychotic medications. No
chotic drug treatment (antipsychotic-naive/antipsychotic-free
AN/AF, versus antipsychotic-treated AT) for LAG other comparisons were significant (P>0.15 in all cases).

chotic-naive first-episode and antipsychotic-free chronic
patients approached significance (t=1.87, P=0.07). The main findings of this study are that smooth pursuit
The effects of Chronicity and Treatment on LAG are eye tracking is significantly more impaired in chronic
displayed in Fig. 4. Chronic patients were generally schizophrenic patients than in first-episode schizophren-
slower than first-episode patients to change the direction ic patients and that this difference is mainly mediated by
of their eye movements in response to the change in tar- the effects of long-term treatment with antipsychotic
get direction [F(1,96)=6.47, P<0.02]. However the main medication. Thus, smooth pursuit velocity gain was sig-
effect of Treatment [F(1,96)=0.94, P=0.33] and Chronic- nificantly better in chronic schizophrenic patients who
ity×Treatment interaction [F(1,96)=0.17, P=0.68] were had been free of antipsychotic medication for at least
not significant. The effects of the covariates age and 6 months, compared to those who had been treated con-
SANS were significant [age: F(1,96)=10.96, P<0.001; tinuously, and was similar in degree to that of first epi-
SANS F(1,96)=5.36, P<0.05]. Older subjects and sub- sode patients.
jects with fewer negative symptoms had larger LAG val- Our results are consistent with those of Bartfai and
ues. The effect of the covariate SAPS was not significant colleagues (1983), who found less smooth pursuit im-
[F(1.96)=0.01, P=0.91]. pairment in a group of never medicated patients com-
Chronic schizophrenic patients made more catch-up pared to a group of previously medicated patients, and
saccades than first episode patients [F(1,85)=8.86, those of Kufferle and colleagues (1992), who found that
P>0.005], but the main effect of Treatment [F(1,85)= smooth pursuit velocity gain was significantly less im-
1.47, P=0.23] and Chronicity×Treatment interaction paired in a group of chronic patients who had been medi-
[F(1,85)=0.15, P=0.70] were not significant. The effect cation free for at least 2 years compared to a group who
of the covariate Age was significant [F(1,85)=6.99, had remained treated. Together with our findings, these
P=0.01]. The effects of the covarites SAPS [F(1,85)= suggest that long-term treatment with antipsychotic med-
0.06, P=0.80] and SANS [F(1,85)=0.24, P=0.63] were ication is associated with a decrease in velocity gain,
not significant. which reverses significantly after a prolonged antipsy-
In summary, although chronic schizophrenic patients chotic-free period.
have reduced velocity gain relative to first-episode pa- One criticism of studies comparing two separate
tients, this effect is mainly due to the particularly poor groups of patients, one treated and one drug-free, is that
performance of the treated chronic patients. The velocity there may be factors explaining the differences in smooth
gains of antipsychotic-free chronic patients showed were pursuit performance other than medication status. Indeed
similar in degree to those of both treated and antipsy- the fact that some patients remain untreated for lengthy
chotic-naive first-episode patients. These results were periods may suggest that they have a milder illness. In
significant when we controlled statistically for differ- our study, care was taken to recruit patients from the
ences due to age and symptom severity. Although dura- same population. Thus most patients were recruited from
tion of treatment influenced velocity gain, this was not a depot clinic, untreated patients being those who had
the case for LAG, a measure of the speed at which sub- failed to attend for at least 6 months. They were also

matched for symptom severity, duration of illness and in- cuits occurs over the course of schizophrenia. Evidence
dices of premorbid educational attainment, socio-eco- for such a deterioration comes from neuropsychological
nomic status and social function, all of which may reflect studies which have demonstrated that chronic schizo-
the severity of illness. To preclude more subtle group dif- phrenic patients are significantly more impaired than
ferences explaining the results, however, a longitudinal first-episode schizophrenic patients on a test of attention-
within-subjects design would be necessary. al set shifting sensitive to frontal function (Elliot et al.
Several studies have addressed the effects of antipsy- 1995; Hutton et al. 1998b; Pantelis et al. 1999).
chotic treatment on smooth pursuit function by compar- An alternative explanation is that the small residual
ing chronically treated patients before and after a medi- smooth pursuit impairment in antipsychotic-free chronic
cation-free period (Holzman et al. 1974; Shagass et al. compared to antipsychotic-naive first-episode patients
1974; Spohn et al. 1988) or have tested medication-free may reflect irreversible changes in brain morphology
patients and then again after medication has been re- following long-term administration of antipsychotic
introduced (Levy et al. 1983; Siever et al. 1986). None drugs. Certainly the most consistent effects of long-term
found changes in smooth pursuit consistent with drug ef- antipsychotic drug treatment are seen in striatum and in-
fects. However in all of these studies the antipsychotic- clude changes in neuronal size and synaptic morphology
free/antipsychotic-treated intervals were short. Most (see Harrison 1999 for a review). These would therefore
were in the order of 2–4 weeks and the longest was of 10 be expected to affect the function oculomotor circuitry
weeks. Our results with first-episode patients are consis- but whether they are irreversible or not is still under de-
tent with these findings. We found no differences in bate (Jeste 1998).
smooth pursuit velocity gain in antipsychotic-naive and The conclusion that long-term treatment with antipsy-
antipsychotic-treated first-episode patients matched for chotic medication worsens smooth pursuit performance
age, symptom severity at presentation and duration of ill- can only be made with respect to the conventional anti-
ness. As our antipsychotic-treated first-episode patients psychotic drugs as no chronic patient was receiving one
had been receiving medication for an average of 7.3 of the newer “atypical” drugs. Indeed we have prelimi-
weeks, this supports the proposal that patients with nary evidence to suggest that atypical drugs may be less
schizophrenia have an intrinsic impairment in smooth detrimental, as within the treated first-episode patient
pursuit performance, which is not further exacerbated by group, we found a trend for the patients taking the atypi-
short-term treatment with conventional antipsychotics. cal antipsychotics olanzapine and risperidone to have
This is also supported by Sweeney and colleagues less impaired velocity gain than patients taking conven-
(1992), who found greater smooth pursuit impairments tional antipsychotic medication. While previous research
in long-term treated chronic patients compared to short- has suggested that atypical antipsychotics actually im-
term treated first-episode patients. pair smooth pursuit performance rather than improve it
Although our results suggest that a considerable com- (Friedman et al. 1992b; Litman et al. 1994), these studies
ponent of the reduced velocity gain commonly observed only addressed the effects of clozapine. Atypical antipsy-
in chronic schizophrenic patients may be due to the effects chotics are defined by their relative lack of extrapyrami-
of long-term antipsychotic medication, the difference in dal side effects rather than by any specific pharmacologi-
velocity gain between the antipsychotic-free chronic and cal property. Compared to risperidone and olanzapine,
antipsychotic-naive first-episode patients approached sig- clozapine has relatively high affinity for histamine re-
nificance. Further, other indices of smooth pursuit perfor- ceptors. This results in significant sedative qualities, par-
mance, unaffected by medication status, also indicate that ticularly early in the course of treatment. As other seda-
illness chronicity may also influence smooth pursuit. We tive medications are known to impair smooth pursuit
have replicated the findings of Sweeney and colleagues performance (Masson et al. 2000), the detrimental ef-
(1994), who compared chronic patients who had been an- fects of clozapine may reflect its histaminergic proper-
tipsychotic-free for at least 28 days with antipsychotic-na- ties. Further research is required in order to replicate the
ive first-episode patients and found a non-significant dec- finding of comparatively unimpaired smooth pursuit per-
rement in velocity gain in the chronic patients and in- formance in patients receiving olanzapine and risperi-
creased rate of catch-up saccades compared to antipsy- done, and explore possible mechanisms for this effect.
chotic-naive first-episode patients. In addition, we found In conclusion, we have demonstrated that smooth pur-
that chronic, but not first-episode, schizophrenic patients suit dysfunction is more severely impaired in chronic pa-
demonstrated a significant impairment on the measure of tients with schizophrenia compared to first-episode pa-
predictive pursuit unrelated to treatment factors. tients. The severity of the dysfunction in chronic pa-
The precise role of predictive processes in smooth tients, at least in terms of the reduction in velocity gain,
pursuit dysfunction is not clear. However, other studies appears to be due mainly to the consequences of long-
confirm that chronic patients with schizophrenia have term antipsychotic medication. Further longitudinal re-
difficulties with this aspect of pursuit (Trillenberg et al. search is needed to fully elucidate the influence of medi-
1998) and have further demonstrated that predictive pur- cation on smooth pursuit performance in schizophrenia.
suit correlates to some extent with measures sensitive to Acknowledgements This research was supported by The Well-
frontal lobe dysfunction (Radant et al. 1997). One hy- come Trust. We are grateful to B. Puri, M. Chapman, S. Mutsatsa,
pothesis is that some deterioration in certain frontal cir- B Haeger and M. Reveley for patient assessments.
Levy DL, Lipton RB, Holzman PS, Davis JM (1983) Eye tracking
References dysfunction unrelated to clinical state and treatment with halo-
peridol. Biol Psychiatry 18:813–819
Abel LA, Friedman L, Jesberger J, Malki A, Meltzer HY (1991) Levy DL, Holzman PS, Matthysse S, Mendell NR (1994) Eye
Quantitative assessment of smooth pursuit gain and catch-up tracking and schizophrenia – a selective review. Schizophr
saccades in schizophrenia and affective-disorders. Biol Psy- Bull 20:47–62
chiatry 29:1063–1072 Lieberman JA, Alvir JMJ, Woerner M, Degreef G, Bilder R,
Alexander GE, DeLong MR, Strick PL (1986) Parallel organiza- Ashatri M, Bogerts B, Mayerhoff D, Geisler S, Loebel A,
tion of functionally segregated circuits linking basal ganglia Levy D, Hinrichsen G, Szymanski S, Chakos M, Koreen A,
and cortex. Annu Rev Neurosci 9:357–381 Borenstein M, Kane J (1992) Prospective study of psychobiol-
Andreasen NC (1984a) Schedule for the Assessment of Negative ogy in first episode schizophrenia at Hillside Hospital. Schiz-
Symptoms (SANS). University of Iowa Press, Iowa City ophr Bull 18:351–371
Andreasen NC (1984b) Schedule for the Assessment of Positive Litman RE, Hommer DW, Clem T, Rapaport MH, Pato CN, Pickar
Symptoms (SAPS). University of Iowa Press, Iowa City D (1989) Smooth pursuit eye-movements in schizophrenia –
Bartfai A, Levander SE, Sedvall G (1983) Smooth pursuit eye- effects of neuroleptic treatment and caffeine. Psychopharma-
movements, clinical symptoms, CSF metabolites, and skin- col Bull 25:473–478
conductance habituation in schizophrenic-patients. Biol Psy- Litman RE, Hommer DW, Radant A, Clem T, Pickar D (1994)
chiatry 18:971–987 Quantitative effects of typical and atypical neuroleptics on
Berg AV van den (1988) Human smooth pursuit during transient smooth pursuit eye tracking in schizophrenia. Schizophr Res
perturbations of predictable target movement. Exp Brain Res 12:107–120
72:95–108 Malaspina D, Colemann EA, Quitkin M, Amador XF, Kaufmann
Berman RA, Colby CL, Genovese CR, Voyvodic JT, Luna B, CA, Gorman JM, Sackeim HA (1994) Effects of pharmacolog-
Thulborn KR, Sweeney JA (1999) Cortical networks subserv- ical catecholamine manipulation on smooth-pursuit eye-move-
ing pursuit and saccadic eye movements in humans: an FMRI ments in normals. Schizophr Res 13:151–159
study. Human Brain Mapping 8:209–225 Masson GS, Mestre DR, Martineau F, Soubrouillard C, Brefel C,
Campion D, Thibaut F, Denise P, Courtin P, Pottier M, Levillain D Rascol O, Blin O (2000) Lorazepam induced modifications of
(1992) SPEM impairment in drug-naive schizophrenic-patients saccadic and smooth pursuit eye movements in humans: atten-
– evidence for a trait marker. Biol Psychiatry 32:891–902 tional and motor factors. Behav Brain Res 108:169–180
Clementz BA (1998) Psychophysiological measures of (dis)inhibi- Nelson HE (1982) The National Adult Reading Test (NART): test
tion as liability indicators for schizophrenia. Psychophysiolo- manual. NFER Nelson, Windsor
gy 35:648–668 O’Driscoll GA, Wolff A-LV, Benkelfat C, Florencio PS, Lal S,
Crawford TJ, Lawden MC, Haegar B, Henderson L, Kennard C Evans AC (2000) Functional neuroanatomy of smooth pursuit
(1995) Smooth pursuit eye movement abnormalities in patients and predictive saccades. NeuroReport 11:1335–1340
with schizophrenia and focal cortical lesions. In: Findlay JM Pantelis C, Barber FZ, Barnes TRE, Nelson HE, Owen AM,
(ed) Eye movement research. Elsevier, Amsterdam, pp 281–289 Robbins TW (1999) Comparison of set-shifting ability in pa-
Creese I, Burt DR, Snyder SH (1976) Dopamine receptor binding tients with chronic schizophrenia and frontal lobe damage.
predicts clinical and pharmacological potencies of antischizo- Schizophr Res 37:251–270
phrenic drugs. Science 192:481–483 Radant AD, Claypoole K, Wingerson DK, Cowley DS, RoyByrne
Elliott R, McKenna PJ, Robbins TW, Sahakian BJ (1995) Neuro- PP (1997) Relationships between neuropsychological and ocu-
psychological evidence for frontostriatal dysfunction in lomotor measures in schizophrenia patients and normal con-
schizophrenia. Psychol Med 25:619–630 trols. Biol Psychiatry 42:797–805
Friedman L, Abel LA, Jesberger JA, Malki A, Meltzer HY Ross RG, Olincy A, Harris JG, Radant A, Adler LE, Compagnon
(1992a) Saccadic intrusions into smooth pursuit in patients N, Freedman R (1999a) The effects of age on a smooth pursuit
with schizophrenia or affective-disorder and normal controls. tracking task in adults with schizophrenia and normal subjects.
Biol Psychiatry 31:1110–1118 Biol Psychiatry 46:383–391
Friedman L, Jesberger JA, Meltzer HY (1992b) Effect of typical an- Ross RG, Olincy A, Radant A (1999b) Amplitude criteria and an-
tipsychotic medications and clozapine on smooth pursuit perfor- ticipatory saccades during smooth pursuit eye movements in
mance in patients with schizophrenia. Psychiatry Res 41:25–36 schizophrenia. Psychophysiology 36:464–468
Gibson JM, Pimlott RM, Kennard C (1987) Oculomotor and man- Shagass C, Amadeo M, Overton DA (1974) Eye tracking perfor-
ual tracking in Parkinson’s disease and the effect of treatment. mance in psychiatric patients. Biol Psychiatry 9:245–260
J Neurol Neurosurg Psychiatry 50:853–860 Siever LJ, Vankammen DP, Linnoila M, Alterman I, Hare T,
Harrison PJ (1999) The neuropathological effects of antipsychotic Murphy DL (1986) Smooth pursuit eye-movement disorder
drugs. Schizophr Res 40:87–99 and its psychobiologic correlates in unmedicated schizophren-
Holzman PS, Proctor LR, Levy DL, Yasillo NJ, Meltzer HY, Hirt ics. Biol Psychiatry 21:1167–1174
SW (1974) Eye tracking dysfunctions in schizophrenic pa- Spohn HE, Coyne L, Spray J (1988) The effect of neuroleptics and
tients and their relatives. Arch Gen Psychiatry 33:1415–1420 tardive-dyskinesia on smooth-pursuit eye-movement in chron-
Hutton SB, Kennard C (1998) Oculomotor abnormalities in ic-schizophrenics. Arch Gen Psychiatry 45:833–840
schizophrenia: critical review. Neurology 50:604–609 Sweeney JA, Haas GL, Li S (1992) Neuropsychological and eye
Hutton SB, Crawford TJ, Duncan L-J, Puri BK, Kennard C Barnes movement abnormalities in first-episode and chronic schizo-
TRE, Joyce EM (1998a) Smooth pursuit and saccadic abnormal- phrenia. Schizophr Bull 18:283–293
ities in first-episode schizophrenia. Psychol Med 28:685–692 Sweeney JA, Haas GL, Li S, Weiden PJ (1994) Selective effects
Hutton SB, Puri BK, Duncan L-J, Robbins TW, Barnes TRE, of antipsychotic medications on eye tracking performance in
Joyce EM (1998b) Executive function in first-episode schizo- schizophrenia. Psychiatry Res 54:185–198
phrenia. Psychol Med 28:463–473 Trillenberg P, Heide W, Junghanns K, Blankenburg M, Arolt V,
Jeste DV, Lohr JB, Eastham JH, Rockwell E, Caligiuri MP (1998) Kompf D (1998) Target anticipation and impairment of
Adverse neurobiological effects of long-term use of neurolep- smooth pursuit eye movements in schizophrenia. Exp Brain
tics: human and animal studies. J Psychiatr Res 32:201–214 Res 120:316–324
King DJ (1994) Psychomotor impairment and cognitive distur- Waterston JA, Barnes GR, Grealy MA, Collins S (1996) Abnor-
bances induced by neuroleptics. Acta Psychiatr Scand 89:53–58 malities of smooth eye and head movements control in Parkin-
Kufferle B, Friedmann A, Topitz A, Foldes P, Anderer P, Kutzer son’s disease. Ann Neurol 39:749–760
M, Steinberger K (1990) Smooth pursuit eye-movements in White OB, Saint-Cyr JA, Tomlinson, RD, Sharpe JA (1983) Ocu-
schizophrenia – influences of neuroleptic treatment and the lomotor deficits in Parkinson’s disease 2. Control of the sac-
question of specificity. Psychopathology 23:106–114 cadic and smooth pursuit systems. Brain 106:571–587