Abnormal-Pursuit in Schizophrenia Evidence for
a
Eye Movements
Genetic Indicator
Philip S. Holzman, PhD; Einar Kringlen, MD; Shelby J. Haberman, PhD; Nicholas J. Yasillo
Deborah L.
Levy, PhD; Leonard R. Proctor, MD;
\s=b\ Disordered smooth-pursuit eye movements occur in a high percentage of schizophrenic patients and their first-degree relatives. A test of the hypothesis that these disorders represent a genetic indicator of schizophrenia was undertaken by testing pursuit eye movements in a sample of monozygotic and dizygotic twins discordant for clinical schizophrenia. Deviant eye tracking is significantly concordant within monozygotic twin pairs, and less so within dizygotic twin pairs discordant for schizophrenia. A genetic interpretation is consistent with these results. (Arch Gen Psychiatry 34:802-805, 1977)
of schizophrenia still Although Thetheetiology prevalence schizophrenic disorganization is
of
unknown.
is
high among the biological relatives of a schizophrenic patient, either an environmentalist or a genetic interpreta¬
tion would fit such data. Pooled results from several studies place concordance rates in monozygotic (MZ) and dizygotic (DZ) twins between 25% and 45% and 5% and 15%, respectively, and provide firmer support for a genetic contribution to schizophrenic predisposition.1'- Recent cross-fostering studies also buttress the hypothesis that genetic factors play a role in the etiology of manifest schizophrenic illness.'"s Since it is difficult to specify a transmission mode that will fit all the data, it is possible that schizophrenia, as diagnosed on the basis of behavioral manifestations, may be genetically heterogeneous. The existence of a characteristic that expresses a genetic predisposition to schizophrenia would be helpful in defin¬ ing a genetically more homogeneous group and ultimately in solving the problem of the mode of transmission. This study offers evidence that disorders of smooth-pursuit eye movements may qualify as such a genetic indicator. In previous reports"7 we described a disorder of smoothpursuit eye movements that occurred in 65% to 80% of diagnosed schizophrenics and in about 45% of their firstAccepted
for publication April 14, 1977. From the Departments of Psychiatry (Drs Holzman and Levy) and Statistics (Dr Haberman), and the Franklin McLean Memorial Research Institute (Mr Yasillo), University of Chicago; the Department of Psychiatry, University of Oslo (Dr Kringlen); and the Department of Otolaryngology, Johns Hopkins Hospital, Baltimore (Dr Proctor). Reprint requests to Department of Psychiatry, University of Chicago, 950 E 59th St, Chicago, IL 60637 (Dr Holzman).
but in only 7% of normal subjects. These results do not reflect medication effects.89 This finding indicated a familial association between eye-tracking dysfunction and schizophrenia, and suggested that impair¬ ments of smooth pursuit may be under genetic control and may reflect a genetic predisposition to schizophrenia. To test this hypothesis, we examined eye-tracking perform¬ ance in sets of MZ and DZ twins in whom one of the pair had a psychiatric condition diagnosed as schizophrenic but most of the co-twins were clinically free of the disorder.
degree relatives,
SUBJECTS The MZ and DZ twin pairs, one of whom was schizophrenic, were selected on the basis of availability from Kringlen's 1967 Norwe¬ gian twin sample.2 The psychological conditions of these twins had been carefully studied and diagnosed ten years ago by one of the authors (E. K.).2 At that time the concordance rate in the MZ sample was 25% to 38% and 4% to 12% in DZ twins, depending on whether a narrow or wide diagnostic criterion of schizophrenia was used. Ten sets of MZ twins and 15 sets of DZ twins were tested in Norway in 1975. For the most part, we limited the sample to those twins who lived in the southern half of Norway, although several of the twins traveled from the extreme northern reaches of the country to Oslo for the testing. Most subjects were tested in their homes; some came to the University Psychiatric Clinic in Oslo; a few probands who were still hospitalized were tested in hospitals in various parts of the southern half of Norway. All subjects were reinterviewed at the time of testing to confirm the diagnosis. To determine zygosity of twin pairs, the Institute of Medical Genetics, University of Oslo, had previously performed blood typing for the following groups: ABO, MNS, rhesus, Lewis, P, Duffy, Kell, Gm, and Gc. Similarity on all systems determined monozygosity. In several twin pairs, the Institute performed retyping and typed several pairs who had not been typed earlier. Of the MZ pairs, two sets were concordant for schizophrenia, two sets were partially concordant, and six sets were discordant. One set of the MZ twins discordant for schizophrenia and tested in the United States was added to the group, making 11 MZ pairs. One pair of DZ twins was partially concordant and 14 pairs were discordant.
METHOD The method for testing pursuit movements, described in previous publications,67 employs electronystagmography.
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of California - San Diego User on 01/07/2016
Table 1.—Sex,
Age, and Eye-Tracking Performance of Monozygotic and 15 Dizygotic Twins*
Twin Pairs
Sex
Age
Pursuit Arrest Score
11
2.—Frequency of Good and Bad Eye Tracking in First-Degree Relatives of Schizophrenic, Nonschizophrenic Psychotic, and Nonpsychotic Patients Table
Eye-Tracking Type
and in Normal Controls
Monozygotic
1af 1bt
M
2a
62
1.9096 1.9782 1.5748 1.4620
47
2bf 3a
55
3bî 4a
61
4bt 5a 5b
55
6a 6b 7a 7b 8a 8b 9a 9b 10a 10b 11a 11b
73 56 35 59 56 21
Patient Group Schizophrenia
1.5908 1.3748 1.7160 1.5739 1.3007 1.2374
Normal controls
1.5674 1.7383 1.7657 1.7403 1.4294 1.4725 2.0109 1.6984 1.5205 1.6037 1.7959 1.7294
M
47
2a 2b
1.8415 M
54 63
6a 6b 7a 7b 8a 8b 9a 9b 10a 10b 11a 11b 12a 12b 13a 13b 14a 14b 15a 15b
56
"Pursuit arrest
1.4204 1.4073
3a 3b 4a 4b 5a 5b
1.9174 1.7156 1.6365 1.5144
59
1.7989 1.6918
55 55
62 52
73 58
45 53
scores are
1.7662 1.8001 1.8968 1.7550 2.1042 1.4851 1.7631 1.4356 1.8115 1.8354 1.3567 1.5529 1.8707 1.9867 1.7703 1.5091 1.6170 1.6292 1.9463 2.0329
expressed
as
the
Members With
A
Tracking 14
26 10 0
Tracking 14 28
25 0
by the eyes. In this study eye movements were recorded using a differential preamplifier driving an FM reel-toreel magnetic recorder. The head was held still by a headrest attached to a chair. The eye-tracking task was administered four times to each subject during a single
1.8654
1bt
in Patients
Nonschizophrenic psychotic Nonpsychotic
Dizygotic 1a
Type of Eye Tracking
Frequency of 1st-0egree Family
mean
log of four tracking
trials, five cycles per trial. The eye-tracking type, whether good (A) or bad (B), is the average judgment of the four trials. In each pair, the proband is labeled a and the co-twin b. tConcordant for schizophrenia. tPartially concordant for schizophrenia.
The corneoretinal potential was recorded from two elec¬ trodes placed at the outer canthus of the eyes and a ground at mid-forehead. A pendulum oscillating at 0.4 Hz subtend¬ ing 20° of visual angle was the stimulus, and in good, intact pursuit movements the sinusoidal motion of the pendulum should be reflected in corresponding motion and velocity
session. The testers were unaware of the identities of any of the subjects with regard to zygosity of twin pair or psychiatric diagnosis. In order to ensure that eye-tracking records would be evaluated blindly, without knowledge of the twinship relationship and the kind of eye tracking in the pairs, each subject was assigned a random number as he was tested. The identities of the subjects were retained in Norway by one of the authors (E. K.), who kept the others uninformed as to the twinships, diagnoses, and relation¬ ships among the people we had tested. After all data had been collected, the tapes were sent to the United States where the signals were displayed on paper for visual analysis and blind scoring. Quality of the tracing, deter¬ mined to be good (A) or bad (B), and the number of pursuit arrests for each of the four trials were scored indepen¬ dently by three of us and then cross-checked. These scores, described in our previous reports, are the principal depen¬ dent variables. Reliability of the scoring was 95% for the qualitative score and .97 for the pursuit arrests. On a prearranged date the scores, names, and relationships of subjects, and zygosity were exchanged between Oslo and Chicago. The data could not be evaluated before this mutual exchange of scores and codes had occurred. In a small experiment with ten subjects, we recorded pursuit movements by electro-oculography (EOG) and by photodiode reflection simultaneously. Both methods distin¬ guished the same five good and five bad eye trackers, indicating that the qualitative score is not an artifact of recording method. The reflection technique showed consistently fewer pursuit arrests than did the EOG, although subjects with tracking had significantly higher numbers of pursuit arrests on both techniques. The quan¬ titative differences in pursuit arrests between the two methods suggest that there is ambiguity concerning the precise meaning of a pursuit arrest and that caution should be observed in interpreting this measure. More than eye movement is represented by the pursuit arrests on EOG, which naturally include other biopotential signals that may
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of California - San Diego User on 01/07/2016
Smooth-pursuit eye movement records of three sets of twins illustrate concordance within proband; bottom record is that of clinically discordant co-twin. nevertheless be systematically implicated in the process. We are continuing our studies into the nature of pursuit arrests.
RESULTS With Previous Schizophrenic
Groups contains the data for each twin with respect
Comparison Table 1 sex, age, and
to
eye-tracking performance. The pursuit arrest score is expressed as the log of the mean number of pursuit arrests per five pendulum cycles for all four trials. The qualitative score is the mean "goodness" rating of the four trials. There were 26 twin probands whose conditions were
twin
pairs. Top record of each pair is that of
as schizophrenic. Two MZ co-twins had at one time also been hospitalized with a schizophrenic condition. There were thus 28 subjects in this sample who had conditions that warranted hospitalization for schizophre¬ nia, either at present or in the past. Nineteen of the 26 probands (73%) had bad eye tracking, and of the two concordant MZ co-twins, one had bad tracking. Twenty of 28 schizophrenic persons (71%) in this sample had eyetracking dysfunctions. This prevalence is within the range of our previous findings7 for schizophrenic patients (65% to 80%) and thus confirms the association of disorderedpursuit eye movements with schizophrenia.
diagnosed
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of California - San Diego User on 01/07/2016
Within the nonschizophrenic co-twins (N 24), 13 (54%) had disordered smooth pursuit eye movements. This is consistent with our previous report of prevalence of bad eye tracking in first-degree relatives.7 =
Concordance Data
quantitative index of pursuit tracking performance, the product-moment correlation in MZ twin pairs was .77 (P < .005) and in DZ pairs it was .40 (P < .15). These values remain the same when corrected For the
arrest score, the
for age and attenuation and are about 80% of the theoret¬ ically expected values of 1.00 and .50 when a trait is under fully additive polygenic control. Given the small sample size, these correlations, although promising, should be viewed as preliminary. Table 1 shows that using pair-wise concordance, five of seven (71%) of the MZ sets and seven of 13 (54%) of the DZ sets were concordant for bad eye tracking. Since twin pairs concordant for good eye tracking were of necessity omitted in this calculation, the sample is too small to assess whether MZ twin pairs were more concordant for poor eye tracking than DZ twin pairs. Within the MZ group, however, MZ probands with good or bad tracking tend to have co-twins with the same kind of tracking (Fisher's exact test: .045, one-tailed), but within the DZ sets, the level of concordance is no greater than could be expected by chance. The concordance rates for schizophrenia in these samples were 18% to 36% for MZ and 0% to 13% for DZ twins, depending on the strictness of the diagnostic criteria. The data thus show a considerably higher concordance for bad eye tracking (71% and 54%) than for clinical schizophrenia, a finding consistent with the previous evidence of poor eye tracking in nonschizophrenic first-degree relatives of schizophrenic probands with bad eye tracking.7 In this context it is instructive to consider other data we have gathered on eye-tracking patterns in family members of schizophrenic patients. The data collected to date in this ongoing study (Table 2) show a marked prevalence of tracking in first-degree family members of schizophrenics, whether or not the proband has tracking. The absence of 100% concordance in the MZ and in family data suggests that whatever the genetic mechanism of poor eye tracking may be, it is not completely penetrant and that some environmental factors are involved. The fact that of five MZ twin pairs concordant for poor eye tracking only one was concordant for schizophrenia is consistent with the suggestion that bad eye tracking indicates a genetic predisposition for schizophrenia in the absence of a clinical diagnosis of this disorder. The two DZ probands (twins 9a and 14a) who had good tracking, but whose nonschizophrenic co-twins had bad tracking, could be interpreted as indicating that the traits are under independent genetic control. However, the evidence of a strong association between poor eye tracking and schizo¬ phrenia suggests in general that bad eye tracking is related to the etiology of schizophrenia and is not the result of a schizophrenic process or an effect of treatment. Inspection of the tracking patterns themselves of some of the twin pairs is especially instructive. The high degree of similarity within twin pairs stamps itself on these =
records, and indicates that where there is tracking concor¬ dance, the tracking deviations by the twins are astonish¬
ingly alike.
COMMENT These results support a conclusion of nonrandom associa¬ tion of eye-tracking patterns within this sample of twins. The correlations of pursuit arrest scores between MZ and DZ twin pairs are close to the theoretically expected values of 1.00 and .50 and therefore suggest that additive genetic effects are more important than random environmental effects and nonadditive genetic effects in determining bad eye tracking. The pairwise concordance rates for the qualitative (good-bad) score are in the correct direction for a genetic interpretation of the results, but the small number of twins in this study limits generalizability at
present.
The finding that a substantial number of schizophrenics have good eye tracking is of interest. It is possible that schizophrenics with good eye tracking have a disorder that is genetically differentiable from those with poor eye tracking. Thus, it would be very important in future studies to ascertain whether or not these two groups differ on biochemical factors or clinical features that are known to vary among schizophrenics. We have collected monoamine oxidase and creatine phosphokinase samples from these twins. The data are being analyzed and will be reported separately. Similarly, it would be important to evaluate the possible modes of transmission in families in which the probands are schizophrenics with good or bad eye track¬
ing.
While we have emphasized the possible genetic implica¬ tions of these findings, we cannot rule out a possible explanation of the data involving toxic, viral, or other environmental causes that could result in impaired eye tracking in a number of members of a family, with the schizophrenic members having the greatest vulnerability. This investigation was supported in part by grants from the Benevolent Foundation of the Scottish Rite, Northern Masonic Jurisdiction; the Howard Pack Fund for Psychiatry; the Forest Hospital Foundation; and US Public Health Service grant MH-19477. Dr Holzman is a recipient of Public Health Service Research Scientist Award MH-70900. Kenneth Kidd and Elizabeth Dorus provided helpful comments.
References 1. Gottesman II, Shields J: Schizophrenia and Genetics. New York, Academic Press Inc, 1972. 2. Kringlen E: Heredity and Environment in the Functional Psychoses. Oslo, Universitetsforlaget, 1967. 3. Heston LL: Psychiatric disorders in foster home reared children of schizophrenic mothers. Br J Psychiatry 112:819-825, 1966. 4. Kety SS, Rosenthal D, Wender PH, et al: Mental illness in the biological and adoptive families of adopted individuals who have become schizophrenic: A preliminary report based on psychiatric interviews, in Fieve RR, Rosenthal D, Brill H (eds): Genetic Research in Psychiatry. Baltimore, Johns Hopkins University Press, 1975, pp 147-165. 5. Wender PH, Rosenthal D, Kety SS, et al: Crossfostering. Arch Gen Psychiatry 30:121-128, 1974. 6. Holzman PS, Proctor LR, Hughes DW: Eye tracking patterns in schizophrenia. Science 181:179-181, 1973. 7. Holzman PS, Proctor LR, Levy DL, et al: Eye tracking dysfunctions in schizophrenic patients and their relatives. Arch Gen Psychiatry 31:143-151, 1974. 8. Holzman PS, Levy DL, Uhlenhuth EH, et al: Smooth-pursuit eye movements and diazepam, CPZ, and secobarbital. Psychopharmacologia 44:111-115, 1975. 9. Shagass C, Amadeo M, Overton DA: Eye tracking performance in psychiatric patients. Biol Psychiatry 9:245-260, 1974
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of California - San Diego User on 01/07/2016
a
Eye Movements
Genetic Indicator
Philip S. Holzman, PhD; Einar Kringlen, MD; Shelby J. Haberman, PhD; Nicholas J. Yasillo
Deborah L.
Levy, PhD; Leonard R. Proctor, MD;
\s=b\ Disordered smooth-pursuit eye movements occur in a high percentage of schizophrenic patients and their first-degree relatives. A test of the hypothesis that these disorders represent a genetic indicator of schizophrenia was undertaken by testing pursuit eye movements in a sample of monozygotic and dizygotic twins discordant for clinical schizophrenia. Deviant eye tracking is significantly concordant within monozygotic twin pairs, and less so within dizygotic twin pairs discordant for schizophrenia. A genetic interpretation is consistent with these results. (Arch Gen Psychiatry 34:802-805, 1977)
of schizophrenia still Although Thetheetiology prevalence schizophrenic disorganization is
of
unknown.
is
high among the biological relatives of a schizophrenic patient, either an environmentalist or a genetic interpreta¬
tion would fit such data. Pooled results from several studies place concordance rates in monozygotic (MZ) and dizygotic (DZ) twins between 25% and 45% and 5% and 15%, respectively, and provide firmer support for a genetic contribution to schizophrenic predisposition.1'- Recent cross-fostering studies also buttress the hypothesis that genetic factors play a role in the etiology of manifest schizophrenic illness.'"s Since it is difficult to specify a transmission mode that will fit all the data, it is possible that schizophrenia, as diagnosed on the basis of behavioral manifestations, may be genetically heterogeneous. The existence of a characteristic that expresses a genetic predisposition to schizophrenia would be helpful in defin¬ ing a genetically more homogeneous group and ultimately in solving the problem of the mode of transmission. This study offers evidence that disorders of smooth-pursuit eye movements may qualify as such a genetic indicator. In previous reports"7 we described a disorder of smoothpursuit eye movements that occurred in 65% to 80% of diagnosed schizophrenics and in about 45% of their firstAccepted
for publication April 14, 1977. From the Departments of Psychiatry (Drs Holzman and Levy) and Statistics (Dr Haberman), and the Franklin McLean Memorial Research Institute (Mr Yasillo), University of Chicago; the Department of Psychiatry, University of Oslo (Dr Kringlen); and the Department of Otolaryngology, Johns Hopkins Hospital, Baltimore (Dr Proctor). Reprint requests to Department of Psychiatry, University of Chicago, 950 E 59th St, Chicago, IL 60637 (Dr Holzman).
but in only 7% of normal subjects. These results do not reflect medication effects.89 This finding indicated a familial association between eye-tracking dysfunction and schizophrenia, and suggested that impair¬ ments of smooth pursuit may be under genetic control and may reflect a genetic predisposition to schizophrenia. To test this hypothesis, we examined eye-tracking perform¬ ance in sets of MZ and DZ twins in whom one of the pair had a psychiatric condition diagnosed as schizophrenic but most of the co-twins were clinically free of the disorder.
degree relatives,
SUBJECTS The MZ and DZ twin pairs, one of whom was schizophrenic, were selected on the basis of availability from Kringlen's 1967 Norwe¬ gian twin sample.2 The psychological conditions of these twins had been carefully studied and diagnosed ten years ago by one of the authors (E. K.).2 At that time the concordance rate in the MZ sample was 25% to 38% and 4% to 12% in DZ twins, depending on whether a narrow or wide diagnostic criterion of schizophrenia was used. Ten sets of MZ twins and 15 sets of DZ twins were tested in Norway in 1975. For the most part, we limited the sample to those twins who lived in the southern half of Norway, although several of the twins traveled from the extreme northern reaches of the country to Oslo for the testing. Most subjects were tested in their homes; some came to the University Psychiatric Clinic in Oslo; a few probands who were still hospitalized were tested in hospitals in various parts of the southern half of Norway. All subjects were reinterviewed at the time of testing to confirm the diagnosis. To determine zygosity of twin pairs, the Institute of Medical Genetics, University of Oslo, had previously performed blood typing for the following groups: ABO, MNS, rhesus, Lewis, P, Duffy, Kell, Gm, and Gc. Similarity on all systems determined monozygosity. In several twin pairs, the Institute performed retyping and typed several pairs who had not been typed earlier. Of the MZ pairs, two sets were concordant for schizophrenia, two sets were partially concordant, and six sets were discordant. One set of the MZ twins discordant for schizophrenia and tested in the United States was added to the group, making 11 MZ pairs. One pair of DZ twins was partially concordant and 14 pairs were discordant.
METHOD The method for testing pursuit movements, described in previous publications,67 employs electronystagmography.
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of California - San Diego User on 01/07/2016
Table 1.—Sex,
Age, and Eye-Tracking Performance of Monozygotic and 15 Dizygotic Twins*
Twin Pairs
Sex
Age
Pursuit Arrest Score
11
2.—Frequency of Good and Bad Eye Tracking in First-Degree Relatives of Schizophrenic, Nonschizophrenic Psychotic, and Nonpsychotic Patients Table
Eye-Tracking Type
and in Normal Controls
Monozygotic
1af 1bt
M
2a
62
1.9096 1.9782 1.5748 1.4620
47
2bf 3a
55
3bî 4a
61
4bt 5a 5b
55
6a 6b 7a 7b 8a 8b 9a 9b 10a 10b 11a 11b
73 56 35 59 56 21
Patient Group Schizophrenia
1.5908 1.3748 1.7160 1.5739 1.3007 1.2374
Normal controls
1.5674 1.7383 1.7657 1.7403 1.4294 1.4725 2.0109 1.6984 1.5205 1.6037 1.7959 1.7294
M
47
2a 2b
1.8415 M
54 63
6a 6b 7a 7b 8a 8b 9a 9b 10a 10b 11a 11b 12a 12b 13a 13b 14a 14b 15a 15b
56
"Pursuit arrest
1.4204 1.4073
3a 3b 4a 4b 5a 5b
1.9174 1.7156 1.6365 1.5144
59
1.7989 1.6918
55 55
62 52
73 58
45 53
scores are
1.7662 1.8001 1.8968 1.7550 2.1042 1.4851 1.7631 1.4356 1.8115 1.8354 1.3567 1.5529 1.8707 1.9867 1.7703 1.5091 1.6170 1.6292 1.9463 2.0329
expressed
as
the
Members With
A
Tracking 14
26 10 0
Tracking 14 28
25 0
by the eyes. In this study eye movements were recorded using a differential preamplifier driving an FM reel-toreel magnetic recorder. The head was held still by a headrest attached to a chair. The eye-tracking task was administered four times to each subject during a single
1.8654
1bt
in Patients
Nonschizophrenic psychotic Nonpsychotic
Dizygotic 1a
Type of Eye Tracking
Frequency of 1st-0egree Family
mean
log of four tracking
trials, five cycles per trial. The eye-tracking type, whether good (A) or bad (B), is the average judgment of the four trials. In each pair, the proband is labeled a and the co-twin b. tConcordant for schizophrenia. tPartially concordant for schizophrenia.
The corneoretinal potential was recorded from two elec¬ trodes placed at the outer canthus of the eyes and a ground at mid-forehead. A pendulum oscillating at 0.4 Hz subtend¬ ing 20° of visual angle was the stimulus, and in good, intact pursuit movements the sinusoidal motion of the pendulum should be reflected in corresponding motion and velocity
session. The testers were unaware of the identities of any of the subjects with regard to zygosity of twin pair or psychiatric diagnosis. In order to ensure that eye-tracking records would be evaluated blindly, without knowledge of the twinship relationship and the kind of eye tracking in the pairs, each subject was assigned a random number as he was tested. The identities of the subjects were retained in Norway by one of the authors (E. K.), who kept the others uninformed as to the twinships, diagnoses, and relation¬ ships among the people we had tested. After all data had been collected, the tapes were sent to the United States where the signals were displayed on paper for visual analysis and blind scoring. Quality of the tracing, deter¬ mined to be good (A) or bad (B), and the number of pursuit arrests for each of the four trials were scored indepen¬ dently by three of us and then cross-checked. These scores, described in our previous reports, are the principal depen¬ dent variables. Reliability of the scoring was 95% for the qualitative score and .97 for the pursuit arrests. On a prearranged date the scores, names, and relationships of subjects, and zygosity were exchanged between Oslo and Chicago. The data could not be evaluated before this mutual exchange of scores and codes had occurred. In a small experiment with ten subjects, we recorded pursuit movements by electro-oculography (EOG) and by photodiode reflection simultaneously. Both methods distin¬ guished the same five good and five bad eye trackers, indicating that the qualitative score is not an artifact of recording method. The reflection technique showed consistently fewer pursuit arrests than did the EOG, although subjects with tracking had significantly higher numbers of pursuit arrests on both techniques. The quan¬ titative differences in pursuit arrests between the two methods suggest that there is ambiguity concerning the precise meaning of a pursuit arrest and that caution should be observed in interpreting this measure. More than eye movement is represented by the pursuit arrests on EOG, which naturally include other biopotential signals that may
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of California - San Diego User on 01/07/2016
Smooth-pursuit eye movement records of three sets of twins illustrate concordance within proband; bottom record is that of clinically discordant co-twin. nevertheless be systematically implicated in the process. We are continuing our studies into the nature of pursuit arrests.
RESULTS With Previous Schizophrenic
Groups contains the data for each twin with respect
Comparison Table 1 sex, age, and
to
eye-tracking performance. The pursuit arrest score is expressed as the log of the mean number of pursuit arrests per five pendulum cycles for all four trials. The qualitative score is the mean "goodness" rating of the four trials. There were 26 twin probands whose conditions were
twin
pairs. Top record of each pair is that of
as schizophrenic. Two MZ co-twins had at one time also been hospitalized with a schizophrenic condition. There were thus 28 subjects in this sample who had conditions that warranted hospitalization for schizophre¬ nia, either at present or in the past. Nineteen of the 26 probands (73%) had bad eye tracking, and of the two concordant MZ co-twins, one had bad tracking. Twenty of 28 schizophrenic persons (71%) in this sample had eyetracking dysfunctions. This prevalence is within the range of our previous findings7 for schizophrenic patients (65% to 80%) and thus confirms the association of disorderedpursuit eye movements with schizophrenia.
diagnosed
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of California - San Diego User on 01/07/2016
Within the nonschizophrenic co-twins (N 24), 13 (54%) had disordered smooth pursuit eye movements. This is consistent with our previous report of prevalence of bad eye tracking in first-degree relatives.7 =
Concordance Data
quantitative index of pursuit tracking performance, the product-moment correlation in MZ twin pairs was .77 (P < .005) and in DZ pairs it was .40 (P < .15). These values remain the same when corrected For the
arrest score, the
for age and attenuation and are about 80% of the theoret¬ ically expected values of 1.00 and .50 when a trait is under fully additive polygenic control. Given the small sample size, these correlations, although promising, should be viewed as preliminary. Table 1 shows that using pair-wise concordance, five of seven (71%) of the MZ sets and seven of 13 (54%) of the DZ sets were concordant for bad eye tracking. Since twin pairs concordant for good eye tracking were of necessity omitted in this calculation, the sample is too small to assess whether MZ twin pairs were more concordant for poor eye tracking than DZ twin pairs. Within the MZ group, however, MZ probands with good or bad tracking tend to have co-twins with the same kind of tracking (Fisher's exact test: .045, one-tailed), but within the DZ sets, the level of concordance is no greater than could be expected by chance. The concordance rates for schizophrenia in these samples were 18% to 36% for MZ and 0% to 13% for DZ twins, depending on the strictness of the diagnostic criteria. The data thus show a considerably higher concordance for bad eye tracking (71% and 54%) than for clinical schizophrenia, a finding consistent with the previous evidence of poor eye tracking in nonschizophrenic first-degree relatives of schizophrenic probands with bad eye tracking.7 In this context it is instructive to consider other data we have gathered on eye-tracking patterns in family members of schizophrenic patients. The data collected to date in this ongoing study (Table 2) show a marked prevalence of tracking in first-degree family members of schizophrenics, whether or not the proband has tracking. The absence of 100% concordance in the MZ and in family data suggests that whatever the genetic mechanism of poor eye tracking may be, it is not completely penetrant and that some environmental factors are involved. The fact that of five MZ twin pairs concordant for poor eye tracking only one was concordant for schizophrenia is consistent with the suggestion that bad eye tracking indicates a genetic predisposition for schizophrenia in the absence of a clinical diagnosis of this disorder. The two DZ probands (twins 9a and 14a) who had good tracking, but whose nonschizophrenic co-twins had bad tracking, could be interpreted as indicating that the traits are under independent genetic control. However, the evidence of a strong association between poor eye tracking and schizo¬ phrenia suggests in general that bad eye tracking is related to the etiology of schizophrenia and is not the result of a schizophrenic process or an effect of treatment. Inspection of the tracking patterns themselves of some of the twin pairs is especially instructive. The high degree of similarity within twin pairs stamps itself on these =
records, and indicates that where there is tracking concor¬ dance, the tracking deviations by the twins are astonish¬
ingly alike.
COMMENT These results support a conclusion of nonrandom associa¬ tion of eye-tracking patterns within this sample of twins. The correlations of pursuit arrest scores between MZ and DZ twin pairs are close to the theoretically expected values of 1.00 and .50 and therefore suggest that additive genetic effects are more important than random environmental effects and nonadditive genetic effects in determining bad eye tracking. The pairwise concordance rates for the qualitative (good-bad) score are in the correct direction for a genetic interpretation of the results, but the small number of twins in this study limits generalizability at
present.
The finding that a substantial number of schizophrenics have good eye tracking is of interest. It is possible that schizophrenics with good eye tracking have a disorder that is genetically differentiable from those with poor eye tracking. Thus, it would be very important in future studies to ascertain whether or not these two groups differ on biochemical factors or clinical features that are known to vary among schizophrenics. We have collected monoamine oxidase and creatine phosphokinase samples from these twins. The data are being analyzed and will be reported separately. Similarly, it would be important to evaluate the possible modes of transmission in families in which the probands are schizophrenics with good or bad eye track¬
ing.
While we have emphasized the possible genetic implica¬ tions of these findings, we cannot rule out a possible explanation of the data involving toxic, viral, or other environmental causes that could result in impaired eye tracking in a number of members of a family, with the schizophrenic members having the greatest vulnerability. This investigation was supported in part by grants from the Benevolent Foundation of the Scottish Rite, Northern Masonic Jurisdiction; the Howard Pack Fund for Psychiatry; the Forest Hospital Foundation; and US Public Health Service grant MH-19477. Dr Holzman is a recipient of Public Health Service Research Scientist Award MH-70900. Kenneth Kidd and Elizabeth Dorus provided helpful comments.
References 1. Gottesman II, Shields J: Schizophrenia and Genetics. New York, Academic Press Inc, 1972. 2. Kringlen E: Heredity and Environment in the Functional Psychoses. Oslo, Universitetsforlaget, 1967. 3. Heston LL: Psychiatric disorders in foster home reared children of schizophrenic mothers. Br J Psychiatry 112:819-825, 1966. 4. Kety SS, Rosenthal D, Wender PH, et al: Mental illness in the biological and adoptive families of adopted individuals who have become schizophrenic: A preliminary report based on psychiatric interviews, in Fieve RR, Rosenthal D, Brill H (eds): Genetic Research in Psychiatry. Baltimore, Johns Hopkins University Press, 1975, pp 147-165. 5. Wender PH, Rosenthal D, Kety SS, et al: Crossfostering. Arch Gen Psychiatry 30:121-128, 1974. 6. Holzman PS, Proctor LR, Hughes DW: Eye tracking patterns in schizophrenia. Science 181:179-181, 1973. 7. Holzman PS, Proctor LR, Levy DL, et al: Eye tracking dysfunctions in schizophrenic patients and their relatives. Arch Gen Psychiatry 31:143-151, 1974. 8. Holzman PS, Levy DL, Uhlenhuth EH, et al: Smooth-pursuit eye movements and diazepam, CPZ, and secobarbital. Psychopharmacologia 44:111-115, 1975. 9. Shagass C, Amadeo M, Overton DA: Eye tracking performance in psychiatric patients. Biol Psychiatry 9:245-260, 1974
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of California - San Diego User on 01/07/2016
