Exp Brain Res (2002) 147:45–52 DOI 10.1007/s00221-002-1208-4

RESEARCH ARTICLE

I. Evdokimidis · N. Smyrnis · T. S. Constantinidis · N. C. Stefanis · D. Avramopoulos · C. Paximadis · C. Theleritis · C. Efstratiadis · G. Kastrinakis · C. N. Stefanis

The antisaccade task in a sample of 2,006 young men I. Normal population characteristics Received: 30 January 2002 / Accepted: 14 June 2002 / Published online: 13 September 2002
Springer-Verlag 2002

Abstract A population of 2,075 young men aged 18– 25 years selected from the conscripts of the Greek Air Force performed an antisaccade task as part of a prospective study for the identification of risk factors in the development of psychoses. The aim of this study, which is ongoing, is to follow this population and investigate the possible predictive value of oculomotor, cognitive, and psychometric factors for the development of psychosis and other psychiatric conditions. In this report we present data concerning the antisaccade task in this population. We measured performance indices, including the percentage of errors (PE), the latencies of different eye movement responses (latency for correct antisaccades, errors, corrections), and performance in perseveration-prone trials. These indices were also evaluated with respect to IQ (measured by the Raven progressive matrices test) and educational level. Mean PE was 23%, with 17% variance. This large variance is of particular importance whenever the detection of a putative deviant behavior is explored. As mean latency of the first eye movement decreased, the PE increased, as did the latency variance. While the negative correlation between percentage of error and mean latency is well established, the relationship of the latency variability of the first response to error production has not been studied before. Thus, optimal performance appears to require both an intermediate mean latency and a small variability. Furthermore, performance seems to be affected by IQ (the higher the IQ score, the lower the percentage of errors). I. Evdokimidis ()) · N. Smyrnis · T.S. Constantinidis · C. Paximadis · C. Theleritis · C. Efstratiadis · G. Kastrinakis Cognition and Action Group, National University of Athens, Neurology Clinic, Aeginition Hospital, 72 Vas. Sofias Ave., Athens 11528, Greece e-mail: [email protected] Tel.: +30-1-7293244 Fax: +30-1-7216424 N. Smyrnis · T.S. Constantinidis · N.C. Stefanis · D. Avramopoulos · C.N. Stefanis University Mental Health Research Institute, National University of Athens, Athens, Greece

This report offers an analysis of the interindividual variation in the performance of the antisaccade task and discusses some of the sources of this variation. Keywords Antisaccade · Saccade latency · Latency variability · Frontal lobe · IQ · Human

Introduction In 1978, Hallet introduced a “novel task” in which subjects were instructed to perform eye movements in the opposite direction from the location of a stimulus that appeared in their right or left peripheral visual field while they were fixating on a central stimulus and, in particular, to saccade to the exact location symmetrical to that of the stimulus (mirror movement; Hallet 1978). This task became known in the literature as the “antisaccade task.” Until 1985, studies employing the antisaccade task were scarce. In 1985, a study on patients with frontal lobe lesions demonstrated that these individuals produced large numbers of errors in the antisaccade task (Guitton et al. 1985). In fact, some of these patients could not perform the task at all and they continuously made saccades toward the visual stimulus. Since then, the number of studies using the antisaccade task has increased significantly, reflecting the increasing interest in this test (for a comprehensive review of the relevant studies, see Everling and Fischer 1998). Most of these studies considered the increase in error rate in this task as a measure of frontal lobe function and thus explored the performance of neurological and psychiatric patients. High error rate in the antisaccade task has been reported for patients with Parkinson’s disease (Lueck et al. 1990; Vidaillet et al. 1994), Huntington’s disease (Lasker et al. 1987), amyotrophic lateral sclerosis (Shaunak et al. 1995), and dementia (Currie et al. 1991). A special interest in this task also developed in the biologically oriented research on schizophrenia and related disorders. Several studies have already reported that schizophrenics as well as subjects characterized as schizotypes demon-

46

strate high error rates in the antisaccade task (Fukushima et al. 1988; O’Driscoll et al. 1998; Sereno and Holzman 1995). The antisaccade error rate was additionally found to be elevated among the schizophrenic patients’ firstdegree relatives (Crawford et al. 1998). Thus the antisaccade error rate was considered as a trait or biological marker of schizophrenia, similar to the deficit observed in smooth eye pursuit (Sereno and Holzman 1995). A high error rate in the antisaccade task was also found in patients with obsessive-compulsive disorder (Tien et al. 1992) and depression (Katsanis et al. 1997). The plethora of clinical studies using the antisaccade task contrasts with the relatively few studies on basic neurophysiological aspects of the task (Everling and Munoz 2000; Everling et al. 1998, 1999; Schlag-Rey et al. 1997; Zhang and Barash 2000) and even fewer imaging studies in healthy humans (O’Driscoll et al. 1995; Sweeney et al. 1996). More disturbing is the fact that the error rate, which is the most widely used measure of performance in clinical studies using the antisaccade task, varies between 0 and 30 for normal individuals in different studies (see review by Everling and Fischer 1998), resulting in a lack of reliable normative data. The differences reported could reflect experimental paradigm differences, biased selection of subjects, or a large variation in the performance of the task in the population of normal control subjects. In addition to the error rate, several parameters, such as the latency of the antisaccades, the latency of the erroneous saccades toward the stimulus, the correction time for errors, and others, have been used in order to provide a more comprehensive description of the task and improve its clinical-diagnostic value, although for these variables normative data are also missing. The data presented in this study stem from the ASPIS project (Athens Study on Psychosis Proneness and Incidence of Schizophrenia) and originate from the analysis of 2,075 apparently normal, young male subjects. First the results obtained on the subject level are dealt with, then the results on a trial or task parameter level. Our results confirm several previously reported findings and add new observations on the interrelationships of different antisaccade task parameters. We also provide a set of normative data regarding several antisaccade parameters, although the population of young men used precludes a generalization to all ages and both sexes.

Materials and methods Subject population A large epidemiological study (ASPIS) was conducted in a sample of 2,130 young male subjects, 18–24 years of age, selected from the conscripts of the Greek Air Force. We tested a randomly selected group of 200–300 conscripts every 2 months for a total period of 15 months (8 sessions); each session lasted 6 days. During that time, the conscripts were tested individually in oculomotor tasks (smooth eye-pursuit, saccade, antisaccade, visual fixation) and cognitive tasks (continuous performance test, CPT-IP; verbal and

spatial working memory tests, N-back). Each individual gave written consent to participate in the study after being informed about the experimental procedures and study goals. The study protocol was approved by the ethics committee of the University Mental Health Research Institute. On the last day of each session, we administered a battery of questionnaires assessing general IQ (Raven Progressive Matrices test), current psychopathology (symptom checklist 90, SCL90-R), schizotypy (Perceptual Aberration Scale, PAS; Schizotypal Personality Questionnaire, SPQ), and personality profile (Temperament and Character Inventory, TCI140-R) to all conscripts in the session group available on that day. All conscripts were initially examined by an army medical committee and individuals with opthalmological problems such as very low visual acuity (below 20–40 best corrected) or strabismus, as well as individuals with other known major medical problems, had already been excluded from the population of conscripts from which our sample originated. We did not perform an optometric evaluation of the conscripts before testing them. Experimental setup All subjects were tested in the military establishment in which they were trained during the first 2 weeks of their service in the Greek Air Force. For this purpose, a specific room was transformed into a laboratory. We used five PC-driven setups for measuring eye movement performance from five individuals and five setups for measuring performance in cognitive tasks from another five individuals. Thus, in each session, ten individuals performed eye movement and cognitive tasks in two groups of five. Each eye movement setup consisted of an adjustable chair for the subject and a specially designed table that had a head holder adjusted at one end and a 17-inch computer monitor mounted at the other end (1 m distance from the head holder). A black cloth was draped over a steel frame, mounted at the edges of the table, to cover the whole area between the subject and the computer monitor thus creating a dark booth. The upper part of the monitor was outside the covered area and was visible to the experimenter. We used the lower part of the monitor to project stimuli that were visible to the subject and the upper part of the monitor to project the eye position data after each completed eye-movement trial so that the experimenter could have on-line control of the quality of the eye movement signal. Eye movements were recorded from the right eye using the IRIS scalar infrared device. Stimulus presentation and recording of the responses were accomplished with a program written in Turbo Pascal 7.0 for DOS. A 12-bit A/D converter was used for data acquisition (Advantech PC-Lab Card 818L). Eye movement data were sampled at 600 Hz and stored in each PC for off-line data processing. Antisaccade task design A population of 2,075 conscripts performed 90 trials of the antisaccade task. Before experiment initiation, a few trials were administered to familiarize the subjects with the task. Each trial started with the appearance of a central fixation stimulus (white cross 0.30.3 of visual angle). After a variable period of 1–2 s, the central stimulus was extinguished and a peripheral stimulus (same white cross) appeared randomly at one of nine distances (2–10 at 1 intervals) either to the left or to the right of the central fixation stimulus. The subjects were instructed to make an eye movement to the opposite direction from that of the peripheral stimulus as quickly as possible. A calibration procedure was performed before the task, using a sequence of four saccadic eye movements, two to the left and two to the right of the central fixation stimulus at an eccentricity of 10.

47 Data analysis We used an interactive program created using the TestPoint CEC software for detection and measuring of saccades on the eye movement record for each trial for each subject. The first derivative of the position record was used to calculate the instantaneous velocity curve. This curve in turn was used to detect the onset of saccadic eye movements using the criterion that 5 consecutive velocity values (8 ms duration at 600 Hz) were above a predefined noise level (the noise level was determined by taking the rootmean-square of the signal in 15 windows of 20 values each, covering the first 500 ms of the 1- to 2-s period of central fixation and then taking the median value of these 15 values). The end of the saccade was the return of the instantaneous velocity to the noise level. We excluded all trials with artifacts (blinks, etc.) in the analysis period or any type of eye movement in the period of 100 ms before the appearance of the peripheral stimulus (Fig. 1A). We further excluded all trials with a response latency that was not within the window of 80–600 ms. Based on these criteria, we excluded from our sample of 2,075 individuals a group of 69 subjects (3%) who did not perform at least 40 valid antisaccade trials. We measured, for each subject (N=2,006 subjects), the following indices of performance: 1. Percentage of errors (PE). A trial was coded as an error if the subject initially performed a movement toward the stimulus and then a corrective movement in the opposite direction. There were practically no cases in which a corrective movement was not performed after an error movement to the stimulus. 2. Mean latency of the first eye movement (Lm) regardless of whether this was an error or correct antisaccade eye movement. 3. Standard deviation of the latency of the first eye movement (Lsd). 4. Mean latency of correct antisaccades (LAm; Fig. 1C). 5. Standard deviation of the latency of correct antisaccades (LAsd). 6. Mean latency of error saccades (LEm; Fig. 1C). 7. Standard deviation of the latency of error saccades (LEsd). 8. Mean latency of corrections (LCm; Fig. 1C). This latency for each error trial was defined as the time it took the subject to initiate the corrective antisaccade from the last position of the eye in the error saccade direction. 9. Standard deviation of the latency of corrections (LCsd). Pearson correlations were performed among these variables. PE was transformed using an arcsine transformation for counts. Correlations of these variables with the years of formal education for each subject (EDU) and the score on the Raven Progressive Matrices test (RAVEN) were also investigated. All but five correlations (Table 1) gave significant effects at the corrected significance level of 0.05 (P=0.05/54=0.0009; Bonferroni correction for multiple comparisons). Thus we also considered the strength of the correlation and we used a correlation coefficient magnitude of more than 0.2 to indicate strong correlations, thus accepting a minimum of approximately 5% (0.22) of common variance between any two variables that were compared. Although this cutoff was arbitrarily chosen, it separates a group of interesting correlations among these variables.

Results Figure 1A shows examples of characteristic eye movement records that were included in the analysis. Trace 1 represents a correct antisaccade and traces 2–4 represent error prosaccades that were subsequently corrected. Figure 1B shows examples of trials that were excluded from further analysis. Trace 1 is a trial in which the subject made saccades during the 100 ms before stimulus

Fig. 1A–C Sample recordings and variable measurements. A The experimental paradigm and a group of typical eye movement traces: 1, the correct antisaccade; 2, an erroneous response starting with a prosaccade and followed by a corrective antisaccade; 3, same as 2, with the difference that there are two erroneous prosaccade movements; 4, an erroneous prosaccade followed immediately (zero correction time) by a corrective antisaccade eye movement. B Some examples of eye movement responses that were excluded from the analysis: 1, square-wave jerks prior to a correct antisaccade, 2 and 3, eye movement contaminated by blinks, 4, slow eye drifts. C Definitions of antisaccade latency variables superimposed on the eye movement traces in a correct trial and an error trial: LA latency to correct antisaccade, LE latency to error, and LC latency to correction

onset, traces 2 and 3 are trials in which a blink occurred in the analysis time, and trace 4 illustrates a slow drift, probably related to head movement. Figure 1C shows the four latency measures. In correct trials, latency to first movement (L) and latency of correct antisaccade (LA) were measured and are the same. In error trials, L and latency of error (LE) were measured and are the same, and latency of correction (LC) was measured in the case of a subsequent corrective antisaccade. Figure 2 shows the means, standard deviations, and histogram distribu-

48 Fig. 2 Distributions and summary statistics of antisaccade task indices. The histogram distribution of the percentage of errors (PE), as well as of the means (black bars) and standard deviations (empty bars) of the latency to the first response L, latency to correct antisaccades LA, latency of errors LE, and latency to corrections LC

49 Fig. 3A, B Relationship of PE to Lm and Lsd. A Scatter plots of the correlations between the percentage of error (PE) and the mean latency of the first response, Lm (empty circles), as well as the correlation of PE with the standard deviation of the latency of the first response, Lsd (filled circles). B A three variable contour plot of Lm, Lsd, and PE. The white-black gradient to the right represents the PE values. The asterisk indicates the area of no values

Table 1 Correlation coefficients among different antisaccade variables are presented on the left side of the table and their correlation with IQ score (RAVEN progressive matrices test) and level of education in years (EDU) are presented on the right side of the table. (Lsd mean latency and its SD for all responses, PE percent of LAm

LAsd 0.62

a

PE –0.25 0.29

LAm 0.87 0.77 0.15

Not significant at a corrected 0.05 level

LAsd 0.54 0.91 0.26 0.65

LEm 0.64 0.35 –0.16 0.44 0.37

errors, LAm, LAsd mean latency and its standard deviation for correct antisaccades, LEm, LEsd mean latency and its standard deviation for errors, LCm, LCsd mean latency and its standard deviation for corrections) LEsd 0.41 0.47 –0.01a 0.31 0.39 0.62

LCm 0.42 0.46 0.14 0.48 0.42 0.27 0.31

LCsd 0.20 0.28 0.16 0.21 0.28 0.23 0.31 0.63

EDU LAm LAsd PE LHm LHsd LEm LEsd LCm LCsd

RAVEN a

–0.03 –0.10 –0.14 –0.05a –0.13 –0.09a –0.11 –0.16 –0.14

–0.06a –0.19 –0.22 –0.11 –0.21 –0.13 –0.15 –0.18 –0.16

50 Fig. 4 Scatter plots and correlation values among the latency means. LA the latency of correct antisaccades, LE the latency of errors and LC the latency of correction

tions for all the antisaccade variables measured for each subject. As stated in the Materials and methods section, PE was transformed using an arcsine transformation for counts that resulted in a distribution closer to normality. For all statistical analyses, the arcsine transformed PE was used. Correlations Table 1 shows the coefficients for the correlations between the antisaccade variables with the years of formal education (EDU) and the RAVEN score of each subject. As stated previously, all but five correlations were significant at the corrected P<0.05, but only those with a correlation coefficient magnitude of more than 0.2 were considered in the following presentation of results. These correlations are: 1. PE with other performance indices.PE was correlated with mean latency of the first response Lm (–0.25), the standard deviation of the first response Lsd (0.29) and the standard deviation of the latency of the correct antisaccades LAsd (0.26). Figure 3A shows a scatter plot of the mean latency of the first response Lm and its standard deviation Lsd with the PE for all subjects. The PE decreased with increasing Lm and increased with increasing variability (larger standard deviation, Lsd) of this latency. This dual and opposite correlation of the percentage of error with the mean and the standard deviation of the latency of the first eye movement is clearly depicted in the three variable (Lm, Lsd and PE) contour map of Fig. 3B. When we employed the mean and standard deviation of the latency to first movement in a regression analysis to predict PE, the regression r was 0.62 and the betas for Lm and Lsd were –0.71 (t=–25, P<0.001) and 0.70 (t=25, P<0.001), respectively. Thus, when considered together, these two latency variables accounted for approximately 40% of the variance of PE in the population (regression r2=0.38). This analysis confirmed our hypothesis that the mean latency and its variance have separate effects on the PE of each individual. The PE was also positively correlated with the standard deviation of the latency of correct antisaccades LAsd, indicating that the PE increased with the

increase in variability of the latency of correct antisaccades. 2. The three mean latencies with each other. The mean latency for correction, LCm, the latency of correct antisaccade LAm, and the latency of error, LEm, were correlated with each other (Table 1, Fig. 4). 3. Each mean latency with its standard deviation. Each mean latency was positively correlated with its standard deviation, indicating that an increase in latency was accompanied by an increase in latency variability independent of whether the latency concerned a correct antisaccade, an error prosaccade, or a corrective antisaccade after an error. 4. All task performance indices with EDU and RAVEN score. The correlations of all antisaccade variables studied with the RAVEN score and EDU were rather weak, with the exception of the correlation of the PE and the RAVEN score, and that of LAsd with the RAVEN score (Table 1). Nevertheless, it should be noted that all correlations were negative, indicating that the general IQ and the education level resulted in a slightly better performance as well as a decrease in mean latency and the variability of the latency of responses (Table 1).

Discussion In this study, we provided data concerning the individual performance on the antisaccade task obtained from a large population of young men. The main results were as follows: (1) the mean error rate for the population was 23% and was unimodaly distributed; (2) the PE was larger for subjects who had shorter mean latencies to the first eye movement and for subjects who had a larger variability of this latency (latency standard deviation); (3) the various mean latencies (LA, LE, and LC) were positively correlated with one another and they were also positively correlated with their respective variances; (4) the error rate and the LAsd were negatively correlated with the general IQ score (RAVEN). Three types of responses are possible in the antisaccade task. Normal subjects typically respond with a correct antisaccade and less frequently with an erroneous saccade toward the stimulus followed by a corrective antisaccade. The third type of response consists of an erroneous saccade toward the stimulus with no correction. This response was related to severe frontal lobe lesions

51

(Guitton et al. 1985) or lack of understanding of the task instruction given (it can be observed occasionally in normal subjects at the beginning of the task). As this type of response was rarely observed in our sample, it is not discussed in the present study. Thus, each subject’s performance was characterized using the PE (the distractibility factor, DF, of other studies). The mean PE of the population was 23%, indicating that normal, young adult men make an error in one out of four responses. However, the intersubject variability of this performance index was very large (SD 17%). Several previous studies reported that the PE for normal adults varied between 0 and 29%, as summarized in a recent review article of the relevant literature (Everling and Fischer 1998). Given the age range of our sample, the range of error rate is of special interest for studies aiming at revealing differences between control and young schizophrenic patients. For example, in a study of antisaccade performance in schizophrenia (Sereno and Holzman 1995), the control population with a mean age of 32 years had a mean PE of 6%, while the schizophrenic patients had a mean of 24%, which is almost identical to the mean value we observed in our population of healthy young men. In addition to idiosyncratic differences in the PE among healthy individuals, differences in the PE could reflect different degrees of learning and motivation. In fact, daily practice improves performance in the antisaccade task (Fischer and Weber 1992). We also observed that some individual characteristics of each subject in the population could interfere with performance. The general IQ measured with the Raven Progressive Matrix score and to a lesser degree the educational level measured by the years of education showed a negative correlation with the PE. In addition the IQ score showed a negative correlation with all antisaccade latency variables, which was significant only for the variability of the latency of correct antisaccades. In other words, the antisaccade task followed a well-established fact in performance tasks, namely that performance is better when the individual responds quickly and with small variability of response time (Deary 2001). It is hard to say if this repetitively confirmed negative relation carries some specific information. It has been argued that subjects with high IQ are generally characterized by a faster “speed of information processing” and this relates to the putative crucial role of working memory (Neubauer 1997). Others point to the lack of specificity of this correlation (Deary 2001). The PE was inversely correlated with the latency of the first eye movement. Thus, individuals who tended to respond quickly on average also tended to make more errors. In a previous study, the latency of the erroneous saccades toward the stimulus was small, falling in the range of express saccades (80–120 ms; Fischer and Weber 1992). Although our data reveal significant differences between the latencies to correct antisaccades (270€39 ms) and the latencies to the errors (208€38 ms), the shorter error latencies are on average much longer than the expected express range saccadic latencies. This discrep-

ancy could be due to the fact that in their study Fischer and Weber used a gap paradigm where the central stimulus disappears before the appearance of the peripheral stimulus. This “gap” interval is known to increase the speed of response for visually guided saccades (Mayfrank et al. 1986). In addition to the well-studied relationship of error rate with mean latency to first response, we showed a positive correlation between error rate and variability of latency to the first response. Thus, subjects with a large variability in latency tended to produce more errors. The fact that the mean latency was positively correlated with latency variability in our population indicated to us that the effect of latency variability on error rate was independent of the effect of mean latency. The regression analysis confirmed our prediction and showed that both mean latency and latency variability (standard deviation of latency) were significant and independent factors predicting the individual’s performance. Moreover the regression model for predicting individual performance from these two factors was able to explain 40% of the variability in error rate in our population. We could then hypothesize the existence of particular subgroups of individuals in our population. One group could consist of the “fast responders,” with short mean latency and narrow latency variability, a combination leading to very high percentages of errors. The contour plot of Fig. 3B shows a group of subjects with a mean latency below 150 ms and a maximum standard deviation of latency of 60 ms, with a PE close or equal to 100%. The second group, “unstable responders,” could be those individuals with a mean latency close to the global mean of 253 ms and very large latency variability. This second group was also associated with high PE. In conclusion, the optimum antisaccade performance seems to require both the selection by the subject of a safe range of response latency and the maintenance of this efficient latency range throughout the task, a behavioral profile that leads to the minimum PE. The importance of response latency variability to PE necessitates further evaluation of the source of the variability of the response latency. It is well known that several task- and/or subject-related parameters affect the response latency, whether mean or variability. Some of the task parameters modulating response latency of eye movements are discussed in the second part of this study. In addition, several cognitive parameters such as attention, anxiety level, and depression influence response latency. In a recent study, age was positively correlated with mean and variance of saccade latency, depression decreased antisaccade latency, and anxiety marginally increased the PE in the antisaccade task (Shafiq-Antonacci et al. 1999). Children 4–5 years of age show highly variable latencies in the saccade task. In contrast, children of school age show similar latency variability to that found in adults, a finding implicating a developmental process that leads to a decrease in response latency variability for saccades (Kowler and Martins 1982). Also performance in the antisaccade task improves dramatically at the age of 9–12 years (Fischer et al. 1997). Thus,

52

it is possible that optimization of response latency variability could be a factor in the observed improvement of performance with age, leading to the adult level of performance at the age of 9–12 years. This role of response latency variability in PE could have interesting implications in the study of populations with a deficit in performance such as the one observed in schizophrenic patients. If an increase in error rate in schizophrenia is due to an underlying instability of response latency, this instability could point to subtle deficits in the developmental processes that lead to maturation of oculomotor behavior. In conclusion, performance indices in the antisaccade task vary greatly among healthy individuals and, thus, their effects should be separately weighted in any study that tries to identify differences in the antisaccade task between normal subjects and patient groups. Acknowledgements This work was supported by the grant EKBAN 97 to Professor C.N. Stefanis from the General Secretariat of Research and Technology of the Greek Ministry of Development. Intrasoft Co. provided technical support.

References Crawford TJ, Sharma T, Puri BK, Murray RM, Berridge DM, Lewis SW (1998) Saccadic eye movements in families multiply affected with schizophrenia:the Maudsley family study. Am J Psychiatry 155:1703–1709 Currie J, Ramsden B, McArthur C, Maruff P (1991) Validation of a clinical antisaccade eye movement test in the assessment of dementia. Arch Neurol 48:644–648 Deary IJ (2001) Human intelligence differences: a recent history. Trends Cognit Sci 4:127–130 Everling S, Fischer B (1998) The antisaccade: a review of basic research and clinical studies. Neuropsychologia 36:885–899 Everling S, Munoz DP (2000) Neuronal correlates for preparatory set associated with prosaccades and anti-saccades in the primate frontal eye field. J Neurosci 20:387–400 Everling S, Dorris MC, Munoz DP (1998) Reflex suppression in the anti-saccade task is dependent on prestimulus neural processes. J Neurophysiol 80:1584–1589 Everling S, Dorris MC, Klein RM, Munoz DP (1999) Role of primate superior colliculus in preparation and execution of antisaccades and prosaccades. J Neurosci 19:2740–2754 Fischer B, Weber H (1992) Characteristics of “anti” saccades in man. Exp Brain Res 89:415–424 Fischer B, Weber H (1997) Effects of stimulus conditions on the performance of antisaccades in man. Exp Brain Res 116:191– 200 Fischer B, Biscaldi M, Gezeck S (1997) On the development of voluntary and reflexive components in human saccade generation. Brain Res 754:285–297 Fukushima J, Fukushima K, Chiba T, Tanaka S, Yamashita I, Kato M (1988) Disturbances of voluntary control of saccadic

eye movements in schizophrenic patients. Biol Psychiatry 23:670–677 Guitton D, Buchtel HA, Douglas RM (1985) Frontal lobe lesions in man cause difficulties in suppressing reflexive glances and in generating goal-directed saccades. Exp Brain Res 58:455–472 Hallet PE (1978) Primary and secondary saccades to goals defined by instructions. Vision Res 18:1279–1296 Katsanis J, Kortenkamp S, Iacono WG, Grove WM (1997) Antisaccade performance in patients with schizophrenia and affective disorder. J Abnorm Psychol 106:468–472 Kowler E, Martins AJ (1982) Eye movements of preschool children. Science 215:997–999 Lasker AG, Zee DS, Hain TC, Folstein SE (1987) Saccades in Huntington’s disease: initiation defects and distractibility. Neurology 44:2285–2289 Lueck CJ, Tanyeri S, Crawford TJ, Henderson L (1990) Antisaccades and remembered saccades in Parkinson’ s disease. J Neurol Neurosurg Psychiatry 53:284–288 Mayfrank L, Mobashery M, Kimmig H, Fischer B (1986) The role of fixation and visual attention in the occurence of express saccades in man. Eur Arch Psychiatry Neurol Sci 235:269–275 Neubauer AC (1997) The mental speed approach of the intelligence. In: Kingma J, Tomic N (eds) Advance in cognition and education: reflections on the concept of intelligence. JAI Press, pp149–173 O’Driscoll GA, Alpert NM, Matthysse SW, Levy DL, Rauch SL, Holzman PS (1995) Functional neuroanatomy of antisaccade eye movements investigated with positron emission tomography. Proc Natl Acad Sci USA 92:925–929 O’Driscoll GA, Lenzenweger MF, Holzman PS (1998) Antisaccades and smooth pursuit eye tracking and schizotypy. Arch Gen Psychiatry 55:837–843 Schlag-Rey M, Amador N, Sanchez H, Schlag J (1997) Antisaccade performance predicted by neuronal activity in the supplementary eye field. Nature 390:398–401 Sereno AB, Holzman PS (1995) Antisaccades and smooth pursuit eye movements in schizophrenia. Biol Psychiatry 37:394–401 Shafiq-Antonacci R, Maruff P, Whyte S, Tyler P, Dudgeon P, Currie J (1999) The effects of age and mood on saccadic function in older individuals. J Gerontol B Psychol Sci Soc Sci 54:361–8 Shallice T (1988) From neuropsychology to mental structure. Cambridge University Press, Cambridge Shaunak S, Orrell RW, O’Sullivan E, Hawken MB, Lane RJ, Henderson L, Kennard C (1995) Oculomotor function in amyotrophic lateral sclerosis, evidence for frontal impairment. Ann Neurol 38:38–44 Sweeney JA, Mintun MA, Kwee S, Wiseman MB, Brown DL, Rosenberg DR, Carl JR (1996) Positron emission tomography study of voluntary saccadic eye movements and spatial working memory. J Neurophysiol 75:454–468 Tien AY, Pearlson GD, Machlin SR, Bylsma FW, Hoehn-Saric R (1992) Oculomotor performance in obsessive-compulsive disorder. Am J Psychiatry 149:641–646 Vidaillet M, Rivaud S, Gouider-Khouja N, Pillon B, Bonnet A, Gaymard B (1994) Eye movements in parkinsonian syndromes. Ann Neurol 35:420–426 Zhang M, Barash S (2000) Neuronal switching of sensorimotor transformations for antisaccades. Nature 408:971–975