Aims: Previous family, adoption and twin studies of schizophrenia have shown that genetic factors contribute significantly to the risk of schizophrenia. The aim of the present study was therefore to investigate whether exploratory eye movement (EEM) abnormalities are related to the genetic markers linked to schizophrenia.
Methods: Twenty-three probands with schizophrenia, 23 of their healthy siblings (23 proband–sibling pairs), and 43 unrelated normal controls performed EEM tasks. Two parameters were measured: (i) number of eye fixations in responsive search (NEFRS) and (ii) responsive search score (RSS).
Results: Abnormalities in NEFRS and RSS were more frequent in schizophrenia probands than in their unaffected siblings and in normal controls, and were also more frequent in the healthy siblings than in normal controls. Thus, the EEM test performances of the healthy siblings were intermediate between those of the probands with schizophrenia and those of normal controls.
Conclusion: Abnormalities of the EEM test parameters may be related to the genetic etiology of schizophrenia. The use of EEM parameters as an endophenotype for schizophrenia may facilitate linkage and association studies in schizophrenia.
PREVIOUS FAMILY, ADOPTION and twin studies of schizophrenia have indicated that genetic components contribute significantly to the development of schizophrenic disorder. The mode of inheritance in schizophrenia, however, is complex. In addition, schizophrenia probably has etiologic heterogeneity, including locus heterogeneity, in genetic-associated cases of schizophrenia.1–4 The conflicting results of recent linkage studies involving schizophrenia as the phenotype may be due to the complexity of genetic transmission.5,6 Current findings of genetic studies in schizophrenia cannot completely account for the genetic factors of schizophrenia. One approach to resolving this issue is to search for a biological marker that fulfils the following criteria: (i) characteristic of schizophrenia; and (ii) related to the genetic predisposition to schizophrenia. Such an indicator may facilitate linkage analysis of schizophrenia.7 Linkage analysis with such a biological marker of schizophrenia may lead to identification of chromosomal loci for susceptibility to schizophrenia.
Our group previously developed a method to study eye movements while subjects viewed geometric figures, called the exploratory eye movement (EEM) test.8–10 We have obtained responsive search scores (RSS) for the EEM test. In previous studies we did not identify any patients with psychiatric diseases in whom the RSS was similar to that of schizophrenia patients. RSS abnormalities were found only in schizophrenia patients.8–10 Moreover, we conducted a worldwide collaborative EEM study to analyze the stability of parameters of EEM. The EEM tests were performed at seven World Health Organization collaborative centers in six countries. The RSS of patients with schizophrenia were significantly lower than those of depressed patients or healthy controls in all centers.10 Thus, we believe that RSS may be a candidate indicator of schizophrenia.
The aim of the present study was to investigate whether EEM abnormalities are related to genetic vulnerability to schizophrenia. For that purpose, this project was designed to compare EEM test data between schizophrenia probands, their healthy siblings, and normal controls. We investigated the possibility that the EEM test can assist with the clarification of genetic components in schizophrenia.
Twenty-three probands with schizophrenia, 23 of their healthy siblings (23 proband–sibling pairs), and 43 unrelated normal controls participated in this study. All probands met the DSM-IV criteria for schizophrenia. The schizophrenia probands (14 men and nine females) had a mean age of 29.3 ± 9.1 years; mean duration of illness was 5.2 ± 4.6 years; mean age at onset was 24.0 ± 5.8 years. All probands were receiving an average daily dosage of 9.3 ± 7.1 mg of a neuroleptic medication equivalent to haloperidol, and were also taking anti-cholinergic drugs. The probands were 10 inpatients and 13 outpatients at Nihon University Hospital in Tokyo or one of three affiliated hospitals (two in Tokyo; one in Chiba Prefecture close to Tokyo). The schizophrenia probands were subclassified into DSM-IV categories: disorganized type (n = 3), paranoid type (n = 15), residual type (n = 2), and undifferentiated type (n = 3). We performed the EEM test on the probands during a period when they were not suffering from acute symptoms. All probands in the present study cooperated with the tests and understood the investigator's instructions clearly.
The normal siblings (10 men and 13 women) had a mean age of 30.9 ± 12.3 years. The goal of this project was to research one non-psychotic sibling for each proband. Whenever possible, the healthy sibling chosen was of the same sex and nearest in age to the proband from each family. The unrelated normal controls (22 men and 21 women) had a mean age of 34.7 ± 12.2 years. The controls were selected from healthy volunteers among hospital staff, students from Nihon University, and members of Tokyo-based drug companies. The healthy siblings and normal controls had no specific history of mental illness according to DSM-IV criteria and had never received psychiatric medications. In addition, the normal controls had no history of psychotic illness in their first-degree family members.
The schizophrenia probands, their healthy siblings, and the normal controls were matched for age and sex. None of the probands, their healthy siblings, or the normal controls had evidence of substance or alcohol abuse or organic brain pathology. The diagnosis of the probands, their healthy siblings, and the normal controls was based on structured clinical interviews for DSM-IV. Each face-to-face interview was conducted by two experienced interviewers. After the nature of the study had been fully explained, written informed consent was obtained from the probands, their siblings, and the normal controls.
Exploratory eye movement
The EEM procedure followed that used by Kojima et al.8 The subjects were asked to sit on a stool equipped with a nac VIII-type Eye Mark Recorder (nac, Tokyo, Japan), a device that detects corneal reflection of infrared light. Three repeats of an original horizontal S-shaped motifs (Fig. 1a,c,e) and two S-shaped motifs that differed slightly from the original one (Fig. 1b,d) were projected individually onto a screen positioned 1.5 m directly in front of the subject's eyes. The width of each of these projected geometric figures was 90 cm, and the height was 75 cm (angle of sight was 33° horizontally and 27.5° vertically). Figure 1 illustrates the sequence of events in the EEM test, which was done in the following steps. First, each subject was directed to view the motif carefully because he/she would be asked to draw it later. The subject was then shown the original S-shaped motif (original motif: OM, Fig. 1a) for 15 s. Immediately after viewing it, the subject was asked to draw the OM from memory. Second, the subject was instructed to compare the OM (Fig. 1a) with a subsequent motif and was then shown a slightly different motif with one bump in a different position (bump in different position motif: BDPM, Fig. 1b) for 15 s; after 15 s had elapsed and with the BDPM still visible, the subject was asked whether it differed from the OM and if it did, how it differed; after the subject had replied and while the BDPM was still being shown, he/she was asked, ‘Are there any other differences?’ This question was repeated until the subject stated there were no further differences. Step 2 was repeated with the OM (Fig. 1c) and with a motif without bumps (no bump motif: NBM, Fig. 1d). Third, the subject was told to look at a projection of the OM (Fig. 1e) again for 15 s and to draw it again.
EEM tests during all steps were recorded on videotape with the eye mark recorder. These tapes were analyzed with a computerized system (eye movement analyzing software for Windows developed by our group). Eye fixations that focused on the same position for at least 200 ms were taken as real eye fixations. Movements of two degrees or more of sight were considered eye movements. In the present study we ascertained the following two measures: number of eye fixations in responsive search (NEFRS) and responsive search score (RSS). The actual NEFRS and RSS of a normal control subject are presented in Fig. 2.
Number of eye fixations in responsive search
The NEFRS is the number of eye fixations during the first 5 s immediately after the final question (‘Are there any other differences?’) when the subjects look at the BDPM (Fig. 1b) and the NBM (Fig. 1d). The NEFRS is the total number: BDPM result (Fig. 2a) + NBM result (Fig. 2b). In Fig. 2 the NEFRS of one control subject is shown: 30 (15 + 15).
Responsive search score
The BDPM and NBM were each divided into seven sections. Figure 2 shows the seven sections relevant to RSS scoring. The number of sections upon which the subject's eye fixed at least once was counted during the first 5 s immediately after the final question while the subjects looked at the BDPM and the NBM. The RSS is the total score: BDPM result (Fig. 2a) + NBM result (Fig. 2b). As shown in Fig. 2, the RSS of one of the controls was 11 (5 + 6).
The NEFRS is a new parameter in the EEM test. The RSS has been developed by our group. The RSS is not raw eye movement data. Hence, it can be suggested that we are not able to obtain comprehensive information from the data of eye movements when we use only the RSS as the EEM parameter. For this reason, in the present study we added the NEFRS as a new item in the EEM test.
Based on the distribution of scores, the present data did not meet the criteria for normality. Therefore, comparisons of the three groups were performed using Wilcoxon matched-pair signed-ranks test for proband group versus sibling group pairwise comparisons of each EEM parameter, and the Mann-Whitney U-test for comparisons of proband group versus normal control group and sibling group versus normal control group according to previous studies.11,12 An association between the two EEM test parameters was investigated using Spearman rank-order correlational test. Statistical significance was set at P < 0.05 (two-tailed). Statistical analysis was carried out with SPSS for Windows, version 14.0 (SPSS, Chicago, IL, USA).
Group comparisons (probands, siblings, controls) based on the EEM test parameters
For visualization of data, boxplots (sometimes called box-and-whiskers plots) of the NEFRS and RSS are presented in Fig. 3. The boxplot describes the distribution and dispersion of a variable, showing its median, quartiles and outliers. The box shows the quartiles; and a line in the box is the median. Whiskers at the ends of the box present the distance from the end of the box to the largest and smallest observed values that are <1.5 box lengths from either end of the box (SPSS manual). As shown in the boxplots, the NEFRS and RSS are lower in schizophrenia probands than in their unaffected siblings or in normal controls, and are also lower in healthy siblings than in normal controls. The scores of the healthy siblings were intermediate between those of the probands with schizophrenia and those of normal controls.
Table 1 shows the results of EEM tests for the three groups. NEFRS was significantly lower in the schizophrenia probands than in their healthy siblings (z = −3.09, P = 0.0020) or in the unrelated normal controls (z = −5.40, P < 0.0001). Moreover, the NEFRS was significantly lower in the healthy siblings than in the normal controls (z = −2.47, P = 0.0137). The probands had significantly lower RSS than that of their siblings (z = −2.38, P = 0.0173) or that of the normal controls (z = −5.39, P < 0.0001). In addition, the siblings had significantly lower RSS than that of the normal controls (z = −3.44, P = 0.0006). There were significant differences between the probands, their siblings, and the normal controls in the NEFRS and the RSS.
Table 1. EEM test parameters (mean ± SD)
Schizophrenia probands (n = 23)
Healthy siblings (n = 23)
Normal, unrelated controls (n = 43)
Probands vs controls
Probands vs siblings
Siblings vs controls
Probands vs controls, Mann–Whitney U-test probands vs siblings, Wilcoxon matched-pair signed-ranks test; siblings vs controls, Mann–Whitney U-test.
EEM, exploratory eye movement; NEFRS, number of eye fixations in responsive search; RSS, responsive search score.
21.4 ± 3.8
25.8 ± 4.4
28.5 ± 3.6
7.5 ± 1.7
8.9 ± 2.0
10.7 ± 1.7
Relationship between the two parameters of the EEM test
Figure 4 illustrates the Spearman correlations between the NEFRS and the RSS. The NEFRS were significantly positively correlated with the RSS in all groups (ρ = 0.53, n = 23, P = 0.0095 in probands; ρ = 0.62, n = 23, P = 0.0016 in siblings; ρ = 0.34, n = 43, P = 0.025 in controls).
Relationship between NEFRS, RSS, and medication
Relationship between NEFRS, RSS, and the dosage of a haloperidol-equivalent neuroleptic medication were examined on Spearman rank-order correlational test to investigate medication effects. There were no significant correlations between NEFRS, RSS, and dosage (NEFRS, ρ = −0.28, n = 19, P = 0.37; RSS, ρ = 0.06, n = 19, P = 0.80).
The principal findings of the present study are that abnormalities of EEM test parameters are more frequent in schizophrenia probands than in their unaffected siblings or in normal controls, and are also more frequent in healthy siblings than in normal controls. The EEM test performances of the healthy siblings were intermediate between those of the probands with schizophrenia and those of the normal controls.
EEM studies of schizophrenia patients have indicated consistent disturbances. In our previous and present investigation we did not identify any normal individuals or patients with other psychiatric diseases in whom the RSS was similar to that of schizophrenia patients. Not only chronic and acute schizophrenia patients but also those in remission can be distinguished on RSS from patients with depression, neurosis, methamphetamine psychosis, temporal lobe epilepsy, and frontal lobe lesions, and from normal controls.8–10,13,14 The present findings are consistent with those of previous studies in that we were able to replicate abnormalities in the EEM test in schizophrenia patients. Thus, we believe that the RSS in the EEM test may be specific to schizophrenia and may be a predictor for schizophrenia.
Because the NEFRS is a new parameter, there are no previous studies that have investigated differences of the NEFRS between schizophrenia patients, non-schizophrenic psychosis patients and normal controls. Thus, the present results do not prove that the NEFRS is specific to schizophrenia. In the present study, we did confirm that there is a significant difference between schizophrenia patients and normal controls. Further investigation is needed to examine the possible presence of NEFRS abnormalities in non-schizophrenic psychosis. If NEFRS is not specific to schizophrenia, it cannot be presumed to be an indicator of genetic vulnerability to schizophrenia. RSS, however, is scored from the NEFRS (Fig. 2), and there were significant correlations between NEFRS and RSS in all groups. The correlation coefficient of the control group was lower than that of the proband or sibling group, but there was a marginal correlation between the NEFRS and the RSS even though the correlation coefficient was low in the control group. Therefore, based on the evidence that the RSS may be specific to schizophrenia, it is possible that the NEFRS may also be one of the characteristics of schizophrenia.
From the fact that siblings share 50% of their genes on average, the present findings indicate that the NEFRS and the RSS may relate to genetic liability to schizophrenia. But siblings also share many environmental features with the schizophrenia probands. Therefore, it is possible that the NEFRS and the RSS may reflect environmental factors. From our previous data and the present study, however, we propose that each of the EEM test parameters may be a trait indicator.8,9,13,15 It seems likely that genetic factors influence the NEFRS and the RSS more potently than do environmental factors.
According to these discussions, the NEFRS and the RSS may be an intermediate phenotype of schizophrenia, and may be useful for linkage studies of schizophrenia. We found a significant linkage to chromosome 22q11.2-q12.1 in our previous linkage study using the NEFRS as an endophenotype for schizophrenia.16 Chromosome 22q11 is one of the most interesting regions for schizophrenia. Several studies have found that adults with 22q11 microdeletions have a high risk of schizophrenia, and suggested linkage between 22q11 and schizophrenia.17,18 Moreover, there are several candidate genes for schizophrenia, for example COMT, PRODH and ZDHHC8 and so on, in this area.17,18 Therefore, based on the fact that the NEFRS is linked to 22q11, we also consider that the NEFRS may be characteristic of schizophrenia, and be related to genetic predisposition to schizophrenia.
In the light of abnormalities of brain function in schizophrenia, we investigated brain activation during a visual exploration task that was similar to the EEM task using functional magnetic resonance imaging (fMRI) in schizophrenia patients and normal controls. The normal control subjects had activations at the bilateral thalamus and the left anterior medial frontal cortex. In contrast, the schizophrenia subjects had activations at the right anterior cingulate gyrus, but no activations at the thalamus and the left anterior medial frontal cortex.19 These findings indicate that the RSS abnormality of schizophrenia may be associated with dysfunctions of the thalamus, frontal cortex or cingulate gyrus.
In conclusion, we suggest that the present EEM test parameters may be markers of genetic predisposition to schizophrenia. In the future, the EEM test may facilitate advances in linkage and association studies of schizophrenia. Mapping EEM abnormalities to a specific chromosome, and finding an association between EEM deficits and a candidate gene for schizophrenia may yield further knowledge concerning genetic influences on schizophrenia.