BIOL PSYCHIATRY 1990;28:705-720
705
Smooth Pursuit Eye Movements of Normal and Schizophrenic Subjects Tracking an Unpredictable Target John S. Allen, Katsuya Matsunaga, Selim Hacisalihzade, and Lawrence Stark
An experimental paradigm employed by several workers in the field of schizophrenic eye movements has involved finding sequences of stimuli that induce saccadic smooth pursuit in the eye movements of normal individuals. It is hoped that the identification of such stimuli will lead to clues concerning the etiology or nature of eye tracking dysfunction in schizophrenia. In this study, the pursuit eye movements of normal and schizophrenic subjects tracking an unpredictable target (composed of summed sine waves) were examined. Eye tracking performance was evaluated both qualitatively and quantitatively using percent root-mean-square (%RMS) error and pursuit gain scores. Schizophrenics are capable of tracking an unpredictable target. Th;,s finding has implications for our understanding of schizophrenic information processing during visual tracking.
Introduction An experimental paradigm employed by several workers in the field of schizophrenic eye movements has involved finding sequences of stimuli that induce saccadic smooth pursuit in the eye movements of normal individuals (Brezinova and Kendell 1977; Lipton et al 1980; Mather and Puchat 1983; Mather 1986; Kaufman and Abel 1986). It is hoped that the identification of such stimuli will lead to clues concerning the etiology or nature of eye tracking dysfunction in schizophrenia. One way to induce low gain smooth pursuit and to increase the number of saccades in a normally tracking irldividual is to use an unpredictable target (Stark et al 1962; Michael and Jones 1966; Bahill, et al 1980). The smooth pursuit system employs "adaptive predictor that allows the system to overcome its innate delays upon exposure to a regular input pattern." (Stark et al 1962, p. 52). According to the Billheimer-Stark model (Stark 1971), learning a predictable targe~'s path involves a three-stage process: (1) a nonpredictive stage, during which the target is first encountered; (2) a transient
From the Department of Anthropology, University of California at Berkeley, Berkeley, California (J.S.A.), the Depadment of Psychology, Faculty of Literature, Kyushu Universiq', Fukuoka, Japan (K.M.), the Department of Control Engineering, ETH, Zurich, Switzerland (S.H.), and the Telembotics and Neurology Unit, University of California at Berkeley, Berkeley, California (S.H.; L.S.) Address reprint requests to D,'. John S. Allen, Pathology Research (i51B), Palo Alto VA Medical Center, Palo Alto, CA 94304. Received November 7, 1989; re~sed March 31, 1990. © 1990 Society of Biological Psychiatry
0006-3223/90/$03.50
706
BIOL PSYCHIATRY 1990:28:705-720
J.S. Allen
stage, during which learning takes place; and (3) a final steady stage, wherein the target motion is being tracked in a predictive fashion. The progress of the subject through these three stages can be followed and quantified using a five-state, partitioned Markov matrix. Bahill and McDonald (1983) found that humans can learn and accurately track any predictable waveform that is smooth and periodic. With an unpredictable target, the clues by which a tracking subject learns to make accurate predictions are by definition absent. Stark et al (1962) noted that the decrease in phase lag seen in subjects tracking a predictab!e rather than an unpredictable target was evidence for the existence of an "adaptive predictor." Although learning undoubtedly plays a role in the accurate anticipation of future target motion of a predictable target, for an unpredictable target, previous target motion is used without conscious thought or effort by subjects to generate expectations concerning the future motion of the target. Kowler et al (1984) found that through the use of a different finite-state Markov model, they could characterize the anticipatory eye movements of a subject who is track:~ng an unpredictable target. Thus it appears that there are two levels of prediction generation: during the tracking of a predictable target, the overall, periodic pattern of the target is learned, and accurate predictions of future target motion are made based on the learned pattern; for an unpredictable target, predictions (which will ultimately be inaccurate), or amicipations, are made based on immediate prior target motion, since there is no general pattern to learn. In this article, the eye movements of normal and schizophrenic subjects tracking an unpredictable target composed of six randomly summed sine waves will be described. The smooth movement of the unpredictable target used in this enperiment (see Figures 1 and 2 for examples of the target and typical subject responses) might be termed "continuous," as opposed to the unpredictable step-ramp target used by Levin et al (1988). Implications of these results for the more general problem of smooth pursuit dysfunction in schizophre, ia, with special reference to the issue of deficits in preregistration and/or postregistration information processing (Braff 1981; Braff and Saccuzzo 1985; Miller et al 1979), are examined in the discussion section.
Materials and M e t h o d s
Subjects Subjects consisted of 14 normal individuals [9 men, 5 women, mean age (_+ SD) 26.1 _ 7.3 years, range 20-47 years] and 16 chronic schizophrenics (11 men, 5 women, 36.9 -+ 6.1 years, range 29-46 years). Schizophr¢iii¢ :~.~bjects were recruited and examined at Motobu Kinen Psychiatric Hospital in Okinawa Prefecture, Japan. The diagnosis of schizophrenia was made by the hospital's clinical psychiatric staff and was based on Schneiderian criteria. Primary symptoms included auditory hallucinations (10 subjects), paranoid delusions (7 subjects), and hebephrenia (4 subjects). Normal subjects were recruited from the hospital staff and among the student population of Kyushu University, Fukuoka, Japan; they were screened for personal and familial histories of mental illness. Although the schizophrenic subjects in this experiment were older than the normal subjects, and the smooth pursuit system has been described as an "age-dependent motor system" (Sharpe and Sylvester 1978), no previous study has demonstrated a significant age-dependent decrease in smooth pui'suit ability in subjects below the age of 50 years. The oldest psychiatric subject in this study was only 46 years old, and thus age was
Eye Movements of Normal and Schizophrenic Subjects
BIOL PSYCHIATRY
1990;28:705-720
707
unlikely to have been an important contributor to observed differences in smooth pursuit performance between the two subject populations. All schizophrenic subjects were receiving standard (phenothiazines and/or haloperidol) antipsychotic medication. It has been demonstrated in numerous studies that these drugs do not affect smooth pursuit eye movements. In addition, patients received a small dose (30-40 mg) of phenobarbital with their bedtime medication. At least 12 hr passed between the administration of the bedtime medication and the examination of eye movements. Given the time course of the effects of barbiturates on smooth pursuit eye movements (Norris 1968), it is unlikely that these drugs had a significant effect on the sclofizophrenics' eye movements. In a larger study (Allen et al 1990) done in conjunction with the present experiment, a comparison between barbiturate-free schizophrenics and those who had received a small dose of the drug with their bedtime medication showed no significant differences in smooth pursuit performance between the two groups.
Experimental Task vnd Data Acquisition Subjects were required to track with their eyes a sinusoidal target moving across +_ 15 degrees of the visual field. The target, a white cursor (approximately 2 para across) of light against a dark background, was displayed using an oscilloscope driven by a wave generator. Four target frequencies were used: 0.2, 0.4, 0.6, and 0.8 Hz, plus th~ unpredictable target. These frequencies correspond to average velocities of 12, 24, 36, and 48 degrees/sec, respectively. The random target was generated by randomly summing six sine curves (0.2, 0.4, 0.6, 0.8, 1.0, and 1.2 Hz) on a computer, followed by digitalto-analog conversion of this file to tape. The tape was then input to the oscilloscope, where it would drive the cursor ir~ an unpredictable (to the subject) fashion. The target was actually pseudo-random, and it had a periodicity of about l0 sec. Subjects tracked for at least 30 sec at each target frequency. Subject performance was continuously monitored and verbal prompts (e.g., urging the subject not to blink and to attend to the tracking task) were made as nccessary. Frequent calibration trials were made using a slow (0.2Hz) square-wave target. Aftqr tracking the four predictable sinusoidal targets, the subject tracked the unpredictable target for approximately 30 sec. Eye movements were measured monocularly (left eye) using the high-resolution photoelectric method (infrared reflectometry) to directly record eye position (Stark et al 1962). Photocells were mounted on lensless eyeglass frames, and a chin rest was used to restrain head movement. Target and eye position signals were recorded simultaneously on a multitrack cassette data recorder. Chart recordings of the data were examined, and those portions of the eye movement signal without artifacts (blinks, etc.), along with the corresponding target signal, were digitized over 12 bits at 60 Hz and stored on disk for processing and quantitative analysis.
Data Analysis After linear transformation of the eye signal based on calibration results, the root-meansquare (RMS) error between target and eye position was determined. For the predictable targets, the RMS error was determined by overlapping 10-sec periods at l-sec intervals over the entire subject run (up to 30 sec) at each target frequency. The 10-sec period with the lowest position RMS score was saved on disk. For the unpredictable target, RMS position error scores were calculated over the entire 30-sec trials, rather than
708
BIOL PSYCHIATRY 1990;28:705-720
J.S. Allen
sampling the ten seconds with the lowest score. RMS scores were normalized by assigning a zero-eye-movement run (eye position fixed at zero degrees) an ewor score of 100%. Calculation of the gain was the same for predictable and unpredictable targets. A Fast Fourier Transform (FFT) was performed on the time functions of the stimulus, s(t), and the subject's response, r(t). This yielded the stimulus, S(w), and response, R(w), in the comr~lex frequency,domain. The gain in decibels at a given frequency, fl, was then defined as G(II) = 20 log
with IR(~)I and IS(fl) I denoting the amplitudes of the complex valued response and stimulus functions at the given frequency II (Oppenheim et al 1983). Before the gain was calculated, saccades were identified using a velocity threshold of 42 degrees/see and removed from the eye movement sig~.al; they were replaced by linear interpolations between the points immediately preceding and following the deleted saccades. At this velocity threshold, some small saccades (< 1 degree in amplitude) could have been missed by the filter, and thus higher velocity epochs may have been included :,n the calculation of gain. The eye movement signal was then smoothed, using a three-point moving average. The resulting gain score is perhaps better described as the "single-mode tracking gain" (Bahill et al 1980). Qualitative analyses of the predictable eye movements only were performed by two blind raters examining the complete (up to 30 see) 0.4-Hz chart recordings. A five-point scale (1 = good, 3 borderline, 5 -= poor) was used to score eye tracking performance. A subject's tracking was considered abnormal if it received an average score greater than 3. In the larger study (Allen et al 1990) done in conjunction with the present experiment, the eye movements of 88 schizophrenics and 37 normal subjects were qualitatively scored. Interrater reliability was high (r 2 = 0.81), and in 93.2% of the cases, the two raters gave identical scores or scores within 1 point of each other. The unpredictable eye movements were not qualitatively scored. Results Position plots of two normal subjects tracking the unpredictable stimulus are presented in Figures la and lb. Similar plots for two schizophrenic subjects are presented in Figures 2a and 2b. Note in all cases the saccadic nature of the pursuit. The normal subject in Figure l a demonstrates excellent tracking of the unpredictable target, although he overshoots the target consistently during rapid target direction reversals. The normal subject in Figure lb scored close to the average for %RMS error; her performance is characterized by an abundance of "catchup" and "look-ahead" (perhaps anticipatory) saccades. The eye movements of the schizophrenic subjects in Figures 2a and 2b are characterized by an abundance of catchup saccades and a relative lack of look-ahead saccades. Qualitative scoring on the basis of the 0.4-Hz chart record showed that 2 of i4 (14%) of the normal subjects and 14 of 16 (88%) of the schizophrenic subjects had abnormal smooth pursuit for a predictable target. The results of the %RMS score analysis are presented in Table 1. Schizophrenics scored significantly worse than normal (Mann-Whitney test, p < 0.01) at all target frequencies and for the unpredictable target. Note that the ratio of the schizophrenic to the normal score (Sc/N) decreased for the unpredictable target relative to the predictable
BIOL PSYCHIATRY 1990;28:705-720
Eye Movements of Normal and Schizophrenic Subjects
709
,..:, .o O t~ o)
N
c~J (~ (M ..
oem
tr) N
"o
0J
c3
o o --
or)
0J 0J
~°
: 0J
.-
O ~J
--.
OI
._
(]~
A
'0 0
O 0 --
cO
..
In
_.
(~)
W
O t~ t~ f~ W
Z
C) --
ID
.o ~o
O
O~
W
OO
c~ c ~ w
L__~r ~ U'I
0
1 In
~
I~
O
In
C)
.~.=
=
710
J.S.
BIOL PSYCHIATRY 1990;28:705-720
fr~
O1
--
(M
cl
tM •.--
t",,.
,.--
tO ~t~
..--
trj
Q t
...= ¢r~
W
....
F"
0d t'M
CO --.. O t'M .--.
01 O
.~
v'q
e~
tO ....
If}
q,}
ILl m
t.-
.~ 4~
.....q.
Z U
...
i'M M M
--.O t,U ---
01
O
fq
--{0 a.
O e~
"-ID
4.,~ t-,~tU ,.U a,-
.-
,,,q.
O N
,~ ..=,
'--" t'M .6 "
b. ~
|
_ It)
.
l'--.."!l_'-.'~l~-'~'-~=~ I Q
In
O
I/~
~
In
~3
Allen
Eye Movements of Normal and Schizophrenic Subjects
BIOL ~'S¥CH~,~RV
1990;28:705-720
.o
Lr} tM
CE
rJ~ ~E CE
._...~ N ._. c::)
---
(~
¢{I
°~
U
E!
I,IJ
~-,
{I] _
...... -.___._~
_
O
~f
I,! C::I
...,..
(Zl
z
O
o
~1
E - ' - - - ' ~ j
-
.....
u~ O i.b
O
c~.C:
-"
~==
.......
0Dm
In
c~
El
El
t~ !/1
El
~0
E.
711
7 !2
J.S. Allen
BIOL PSYCHIATRY 1990;28:705-720
c~ ¢J
e~
•. - .
O
ed
LLI
.=
¢q i'M
t~
r'~ .._oi
,..
c
{~
~
c O o
~=~ ..--
~2
.== if} M
,.= t'O
O
Z =..
{M M
..==, M
W
.==.
{~
•'-
O}
•= =
{~
•. = -
p,,
,A
.q. u
z
'== {D •" -
ill O O
M
r"'
705
Smooth Pursuit Eye Movements of Normal and Schizophrenic Subjects Tracking an Unpredictable Target John S. Allen, Katsuya Matsunaga, Selim Hacisalihzade, and Lawrence Stark
An experimental paradigm employed by several workers in the field of schizophrenic eye movements has involved finding sequences of stimuli that induce saccadic smooth pursuit in the eye movements of normal individuals. It is hoped that the identification of such stimuli will lead to clues concerning the etiology or nature of eye tracking dysfunction in schizophrenia. In this study, the pursuit eye movements of normal and schizophrenic subjects tracking an unpredictable target (composed of summed sine waves) were examined. Eye tracking performance was evaluated both qualitatively and quantitatively using percent root-mean-square (%RMS) error and pursuit gain scores. Schizophrenics are capable of tracking an unpredictable target. Th;,s finding has implications for our understanding of schizophrenic information processing during visual tracking.
Introduction An experimental paradigm employed by several workers in the field of schizophrenic eye movements has involved finding sequences of stimuli that induce saccadic smooth pursuit in the eye movements of normal individuals (Brezinova and Kendell 1977; Lipton et al 1980; Mather and Puchat 1983; Mather 1986; Kaufman and Abel 1986). It is hoped that the identification of such stimuli will lead to clues concerning the etiology or nature of eye tracking dysfunction in schizophrenia. One way to induce low gain smooth pursuit and to increase the number of saccades in a normally tracking irldividual is to use an unpredictable target (Stark et al 1962; Michael and Jones 1966; Bahill, et al 1980). The smooth pursuit system employs "adaptive predictor that allows the system to overcome its innate delays upon exposure to a regular input pattern." (Stark et al 1962, p. 52). According to the Billheimer-Stark model (Stark 1971), learning a predictable targe~'s path involves a three-stage process: (1) a nonpredictive stage, during which the target is first encountered; (2) a transient
From the Department of Anthropology, University of California at Berkeley, Berkeley, California (J.S.A.), the Depadment of Psychology, Faculty of Literature, Kyushu Universiq', Fukuoka, Japan (K.M.), the Department of Control Engineering, ETH, Zurich, Switzerland (S.H.), and the Telembotics and Neurology Unit, University of California at Berkeley, Berkeley, California (S.H.; L.S.) Address reprint requests to D,'. John S. Allen, Pathology Research (i51B), Palo Alto VA Medical Center, Palo Alto, CA 94304. Received November 7, 1989; re~sed March 31, 1990. © 1990 Society of Biological Psychiatry
0006-3223/90/$03.50
706
BIOL PSYCHIATRY 1990:28:705-720
J.S. Allen
stage, during which learning takes place; and (3) a final steady stage, wherein the target motion is being tracked in a predictive fashion. The progress of the subject through these three stages can be followed and quantified using a five-state, partitioned Markov matrix. Bahill and McDonald (1983) found that humans can learn and accurately track any predictable waveform that is smooth and periodic. With an unpredictable target, the clues by which a tracking subject learns to make accurate predictions are by definition absent. Stark et al (1962) noted that the decrease in phase lag seen in subjects tracking a predictab!e rather than an unpredictable target was evidence for the existence of an "adaptive predictor." Although learning undoubtedly plays a role in the accurate anticipation of future target motion of a predictable target, for an unpredictable target, previous target motion is used without conscious thought or effort by subjects to generate expectations concerning the future motion of the target. Kowler et al (1984) found that through the use of a different finite-state Markov model, they could characterize the anticipatory eye movements of a subject who is track:~ng an unpredictable target. Thus it appears that there are two levels of prediction generation: during the tracking of a predictable target, the overall, periodic pattern of the target is learned, and accurate predictions of future target motion are made based on the learned pattern; for an unpredictable target, predictions (which will ultimately be inaccurate), or amicipations, are made based on immediate prior target motion, since there is no general pattern to learn. In this article, the eye movements of normal and schizophrenic subjects tracking an unpredictable target composed of six randomly summed sine waves will be described. The smooth movement of the unpredictable target used in this enperiment (see Figures 1 and 2 for examples of the target and typical subject responses) might be termed "continuous," as opposed to the unpredictable step-ramp target used by Levin et al (1988). Implications of these results for the more general problem of smooth pursuit dysfunction in schizophre, ia, with special reference to the issue of deficits in preregistration and/or postregistration information processing (Braff 1981; Braff and Saccuzzo 1985; Miller et al 1979), are examined in the discussion section.
Materials and M e t h o d s
Subjects Subjects consisted of 14 normal individuals [9 men, 5 women, mean age (_+ SD) 26.1 _ 7.3 years, range 20-47 years] and 16 chronic schizophrenics (11 men, 5 women, 36.9 -+ 6.1 years, range 29-46 years). Schizophr¢iii¢ :~.~bjects were recruited and examined at Motobu Kinen Psychiatric Hospital in Okinawa Prefecture, Japan. The diagnosis of schizophrenia was made by the hospital's clinical psychiatric staff and was based on Schneiderian criteria. Primary symptoms included auditory hallucinations (10 subjects), paranoid delusions (7 subjects), and hebephrenia (4 subjects). Normal subjects were recruited from the hospital staff and among the student population of Kyushu University, Fukuoka, Japan; they were screened for personal and familial histories of mental illness. Although the schizophrenic subjects in this experiment were older than the normal subjects, and the smooth pursuit system has been described as an "age-dependent motor system" (Sharpe and Sylvester 1978), no previous study has demonstrated a significant age-dependent decrease in smooth pui'suit ability in subjects below the age of 50 years. The oldest psychiatric subject in this study was only 46 years old, and thus age was
Eye Movements of Normal and Schizophrenic Subjects
BIOL PSYCHIATRY
1990;28:705-720
707
unlikely to have been an important contributor to observed differences in smooth pursuit performance between the two subject populations. All schizophrenic subjects were receiving standard (phenothiazines and/or haloperidol) antipsychotic medication. It has been demonstrated in numerous studies that these drugs do not affect smooth pursuit eye movements. In addition, patients received a small dose (30-40 mg) of phenobarbital with their bedtime medication. At least 12 hr passed between the administration of the bedtime medication and the examination of eye movements. Given the time course of the effects of barbiturates on smooth pursuit eye movements (Norris 1968), it is unlikely that these drugs had a significant effect on the sclofizophrenics' eye movements. In a larger study (Allen et al 1990) done in conjunction with the present experiment, a comparison between barbiturate-free schizophrenics and those who had received a small dose of the drug with their bedtime medication showed no significant differences in smooth pursuit performance between the two groups.
Experimental Task vnd Data Acquisition Subjects were required to track with their eyes a sinusoidal target moving across +_ 15 degrees of the visual field. The target, a white cursor (approximately 2 para across) of light against a dark background, was displayed using an oscilloscope driven by a wave generator. Four target frequencies were used: 0.2, 0.4, 0.6, and 0.8 Hz, plus th~ unpredictable target. These frequencies correspond to average velocities of 12, 24, 36, and 48 degrees/sec, respectively. The random target was generated by randomly summing six sine curves (0.2, 0.4, 0.6, 0.8, 1.0, and 1.2 Hz) on a computer, followed by digitalto-analog conversion of this file to tape. The tape was then input to the oscilloscope, where it would drive the cursor ir~ an unpredictable (to the subject) fashion. The target was actually pseudo-random, and it had a periodicity of about l0 sec. Subjects tracked for at least 30 sec at each target frequency. Subject performance was continuously monitored and verbal prompts (e.g., urging the subject not to blink and to attend to the tracking task) were made as nccessary. Frequent calibration trials were made using a slow (0.2Hz) square-wave target. Aftqr tracking the four predictable sinusoidal targets, the subject tracked the unpredictable target for approximately 30 sec. Eye movements were measured monocularly (left eye) using the high-resolution photoelectric method (infrared reflectometry) to directly record eye position (Stark et al 1962). Photocells were mounted on lensless eyeglass frames, and a chin rest was used to restrain head movement. Target and eye position signals were recorded simultaneously on a multitrack cassette data recorder. Chart recordings of the data were examined, and those portions of the eye movement signal without artifacts (blinks, etc.), along with the corresponding target signal, were digitized over 12 bits at 60 Hz and stored on disk for processing and quantitative analysis.
Data Analysis After linear transformation of the eye signal based on calibration results, the root-meansquare (RMS) error between target and eye position was determined. For the predictable targets, the RMS error was determined by overlapping 10-sec periods at l-sec intervals over the entire subject run (up to 30 sec) at each target frequency. The 10-sec period with the lowest position RMS score was saved on disk. For the unpredictable target, RMS position error scores were calculated over the entire 30-sec trials, rather than
708
BIOL PSYCHIATRY 1990;28:705-720
J.S. Allen
sampling the ten seconds with the lowest score. RMS scores were normalized by assigning a zero-eye-movement run (eye position fixed at zero degrees) an ewor score of 100%. Calculation of the gain was the same for predictable and unpredictable targets. A Fast Fourier Transform (FFT) was performed on the time functions of the stimulus, s(t), and the subject's response, r(t). This yielded the stimulus, S(w), and response, R(w), in the comr~lex frequency,domain. The gain in decibels at a given frequency, fl, was then defined as G(II) = 20 log
with IR(~)I and IS(fl) I denoting the amplitudes of the complex valued response and stimulus functions at the given frequency II (Oppenheim et al 1983). Before the gain was calculated, saccades were identified using a velocity threshold of 42 degrees/see and removed from the eye movement sig~.al; they were replaced by linear interpolations between the points immediately preceding and following the deleted saccades. At this velocity threshold, some small saccades (< 1 degree in amplitude) could have been missed by the filter, and thus higher velocity epochs may have been included :,n the calculation of gain. The eye movement signal was then smoothed, using a three-point moving average. The resulting gain score is perhaps better described as the "single-mode tracking gain" (Bahill et al 1980). Qualitative analyses of the predictable eye movements only were performed by two blind raters examining the complete (up to 30 see) 0.4-Hz chart recordings. A five-point scale (1 = good, 3 borderline, 5 -= poor) was used to score eye tracking performance. A subject's tracking was considered abnormal if it received an average score greater than 3. In the larger study (Allen et al 1990) done in conjunction with the present experiment, the eye movements of 88 schizophrenics and 37 normal subjects were qualitatively scored. Interrater reliability was high (r 2 = 0.81), and in 93.2% of the cases, the two raters gave identical scores or scores within 1 point of each other. The unpredictable eye movements were not qualitatively scored. Results Position plots of two normal subjects tracking the unpredictable stimulus are presented in Figures la and lb. Similar plots for two schizophrenic subjects are presented in Figures 2a and 2b. Note in all cases the saccadic nature of the pursuit. The normal subject in Figure l a demonstrates excellent tracking of the unpredictable target, although he overshoots the target consistently during rapid target direction reversals. The normal subject in Figure lb scored close to the average for %RMS error; her performance is characterized by an abundance of "catchup" and "look-ahead" (perhaps anticipatory) saccades. The eye movements of the schizophrenic subjects in Figures 2a and 2b are characterized by an abundance of catchup saccades and a relative lack of look-ahead saccades. Qualitative scoring on the basis of the 0.4-Hz chart record showed that 2 of i4 (14%) of the normal subjects and 14 of 16 (88%) of the schizophrenic subjects had abnormal smooth pursuit for a predictable target. The results of the %RMS score analysis are presented in Table 1. Schizophrenics scored significantly worse than normal (Mann-Whitney test, p < 0.01) at all target frequencies and for the unpredictable target. Note that the ratio of the schizophrenic to the normal score (Sc/N) decreased for the unpredictable target relative to the predictable
BIOL PSYCHIATRY 1990;28:705-720
Eye Movements of Normal and Schizophrenic Subjects
709
,..:, .o O t~ o)
N
c~J (~ (M ..
oem
tr) N
"o
0J
c3
o o --
or)
0J 0J
~°
: 0J
.-
O ~J
--.
OI
._
(]~
A
'0 0
O 0 --
cO
..
In
_.
(~)
W
O t~ t~ f~ W
Z
C) --
ID
.o ~o
O
O~
W
OO
c~ c ~ w
L__~r ~ U'I
0
1 In
~
I~
O
In
C)
.~.=
=
710
J.S.
BIOL PSYCHIATRY 1990;28:705-720
fr~
O1
--
(M
cl
tM •.--
t",,.
,.--
tO ~t~
..--
trj
Q t
...= ¢r~
W
....
F"
0d t'M
CO --.. O t'M .--.
01 O
.~
v'q
e~
tO ....
If}
q,}
ILl m
t.-
.~ 4~
.....q.
Z U
...
i'M M M
--.O t,U ---
01
O
fq
--{0 a.
O e~
"-ID
4.,~ t-,~tU ,.U a,-
.-
,,,q.
O N
,~ ..=,
'--" t'M .6 "
b. ~
|
_ It)
.
l'--.."!l_'-.'~l~-'~'-~=~ I Q
In
O
I/~
~
In
~3
Allen
Eye Movements of Normal and Schizophrenic Subjects
BIOL ~'S¥CH~,~RV
1990;28:705-720
.o
Lr} tM
CE
rJ~ ~E CE
._...~ N ._. c::)
---
(~
¢{I
°~
U
E!
I,IJ
~-,
{I] _
...... -.___._~
_
O
~f
I,! C::I
...,..
(Zl
z
O
o
~1
E - ' - - - ' ~ j
-
.....
u~ O i.b
O
c~.C:
-"
~==
.......
0Dm
In
c~
El
El
t~ !/1
El
~0
E.
711
7 !2
J.S. Allen
BIOL PSYCHIATRY 1990;28:705-720
c~ ¢J
e~
•. - .
O
ed
LLI
.=
¢q i'M
t~
r'~ .._oi
,..
c
{~
~
c O o
~=~ ..--
~2
.== if} M
,.= t'O
O
Z =..
{M M
..==, M
W
.==.
{~
•'-
O}
•= =
{~
•. = -
p,,
,A
.q. u
z
'== {D •" -
ill O O
M
r"'
