 |
|
Neurocognitive Impairments in Nonpsychotic Parents of Children With Schizophrenia and Attention-Deficit/Hyperactivity Disorder
The University of California, Los Angeles Family Study
Robert F. Asarnow, PhD;
Keith H. Nuechterlein, PhD;
Kenneth L. Subotnik, PhD;
David L. Fogelson, MD;
Richard D. Torquato, MA;
Diana L. Payne, PhD;
Joy Asamen, PhD;
Jim Mintz, PhD;
Donald Guthrie, PhD
Arch Gen Psychiatry. 2002;59:1053-1060.
ABSTRACT
| |
Background We tested the hypotheses that certain neurocognitive impairments index genetic liability to schizophrenia and that childhood-onset schizophrenia (COS) is a variant of adult-onset schizophrenia (AOS) by determining whether parents of COS probands show the types of neurocognitive impairments found in relatives of AOS probands.
Methods Parents of COS probands (n = 79) were compared with parents of attention-deficit/hyperactivity disorder (ADHD; n = 190) and community control (CC; n = 115) probands on 3 neurocognitive tasks shown in previous research to detect impairments in patients with AOS and ADHD and in the relatives of patients with AOS. Parents with a diagnosis of psychosis were excluded from the study.
Results On the Degraded Stimulus–Continuous Performance Test and the Trail-Making Test B–Adolescent Version, the parents of COS probands performed significantly worse than the parents of CC and ADHD probands, who did not differ significantly from each other. On the Span of Apprehension, we found no significant group differences. Using rigorous cutoffs, a combination of scores on the 3 neurocognitive tests identified 16 (20%) of the mothers and fathers of COS probands compared with 0% of the mothers and fathers of CC probands. There was diagnostic specificity of the neurocognitive impairments. A combination of neurocognitive scores identified 6 (12%) of the mothers of COS probands vs 0% of the mothers of ADHD probands. A cutoff that identified 2 (2%) of the fathers of ADHD probands classified 5 (17%) of the fathers of COS probands. We found no significant differences in neurocognitive functions between the parents of ADHD and CC probands.
Conclusions The aggregation of neurocognitive impairments in the parents of COS probands provides further evidence of etiologic continuity between COS and AOS. A substantial subgroup of parents of COS probands had a worse neurocognitive performance than that of any of the parents of ADHD and CC probands. Receiver operating characteristic curves showed that when rigorous cutoffs define neurocognitive impairments, the combination of scores on certain neurocognitive tasks produced a level of diagnostic accuracy in the parents of COS probands that is sufficient for use in genetic linkage studies.
INTRODUCTION
COGNITIVE SYMPTOMS and neurocognitive impairments are core features of schizophrenia and attention-deficit/hyperactivity disorder (ADHD). Genetic predispositions appear to play an important role in the etiology of both schizophrenia1-3 and ADHD.4-5 The genetic predispositions may be reflected in subtle brain disorders.4, 6 The present study tests the hypothesis that certain neurocognitive impairments index liability to schizophrenia and ADHD.
First-degree relatives of probands with adulthood-onset schizophrenia (AOS) show subtle abnormalities on neurocognitive tasks that require mental encoding and control, abstraction, motor function, sustained attention, and rapid visual information processing.7-10 First-degree relatives of AOS probands showed deficits on the 3 neurocognitive tasks used in this study, the Degraded Stimulus–Continuous Performance Test (DS-CPT),11 the Trail-Making Test B–Adolescent Version (Trails B),12 and Span of Apprehension (Span).13 We determined whether the parents of probands with childhood-onset schizophrenia (COS), like the parents of AOS probands, showed neurocognitive impairments. Many neurocognitive,14 psychophysiologic,15-17 and structural brain abnormalities18-19 observed in AOS are present in COS, suggesting that COS and AOS have the same neurobiological substrate. If the parents of COS probands also have similar neurocognitive impairments, this will provide further evidence that COS is a variant of AOS and that these impairments index genetic liability to both COS and AOS.
Some previous studies that found neurocognitive impairments in relatives of schizophrenic patients did not exclude relatives with overt psychosis. We excluded parents with schizophrenia to determine the extent to which the presence of neurocognitive impairments is independent of the presence of schizophrenia.
Statistical power in linkage studies is directly related to the relative incidence of a phenotype in relatives of schizophrenic probands (true-positive findings) compared with its incidence in the general population (false-positive findings).20 Given the familial risk and population base rate of schizophrenia, to improve on the clinical diagnosis of schizophrenia as a phenotype, indicators must have ratios of true- to false-positive findings of at least 10:1.21-22 Therefore, to be useful in linkage studies, cutoff scores on neurocognitive tasks should produce low false-positive rates (high specificity).20 With a notable exception,8 previous studies reported group differences between the relatives of AOS and control probands and did not examine the diagnostic accuracy of neurocognitive measures. "For neurocognitive measures to be useful indices of a genetic liability to schizophrenia, group differences are not sufficient to show that a measure is useful for classification, rather what is needed are diagnostic accuracy analyses which specify the sensitivity and specificity . . . of measures."8 The present study performed diagnostic accuracy analyses of 3 neurocognitive tasks to determine whether they are useful indices of liability to schizophrenia in linkage studies.
Given the central role of cognitive impairments in the diagnosis and pathophysiology of ADHD and the growing evidence5-6 of a familial aggregation of ADHD, surprisingly little research has examined neurocognitive functioning in relatives of individuals with ADHD. Nonhyperactive siblings of ADHD probands have decreased intelligence and achievement test scores23-24 but are similar to control subjects on measures of executive functioning, attention, and verbal learning.25 The present study provides a further test of the hypothesis that certain neurocognitive impairments index liability in ADHD probands by examining neurocognitive functioning in their parents on 3 tasks that detect impairments in patients with ADHD.
METHODS
SUBJECTS
The University of California, Los Angeles Family Study consists of parallel studies of biological relatives of 3 child (principal investigator, R.F.A.) and 3 adult (principal investigator, K.H.N.) proband groups. This report presents data on the parents of the 2 child proband groups with COS and ADHD, and of a group of community control (CC) probands. We ascertained COS probands by screening 9 Los Angeles County treatment facilities and the University of California, Los Angeles Neuropsychiatric Hospital and school-based programs for seriously emotionally disturbed children in Los Angeles County. We ascertained ADHD probands through outpatient pediatric clinics and support groups for the families of children with ADHD located in Los Angeles County.
Lists of potential CC probands living in the same ZIP codes as COS probands were obtained from a scientific survey research firm (Survey Sampling, Inc, Fairfield, Conn). Parents of potential CC probands were contacted by telephone to explain the purpose of the study and to screen out probands who had schizophrenia or ADHD or were gifted or learning disabled. All participants in the study provided a written, informed consent and assent (if a minor).
A child psychologist used the Schedule for Affective Disorder and Schizophrenia for School-Aged Children–Epidemiologic Version (K-SADS)26 to interview the parent about the proband's psychiatric symptoms and then conducted a psychiatric interview of the proband. The psychologist then reviewed the medical and school records and K-SADS interviews with a child psychiatrist to reach consensus diagnoses.27 Fifty-one children with consensus diagnoses of DSM-III-R28 schizophrenia and with first onset of psychosis before 13 years of age constituted the schizophrenic probands. One hundred eighteen children with consensus diagnoses of DSM-III-R ADHD (at least 8 ADHD symptoms rated as definitely or probably present) constituted the ADHD probands. Of these, 106 had at least 8 ADHD symptoms rated as definitely present, whereas 12 children had at least 6 or 7 ADHD symptoms rated as definitely present and 1 or 2 ADHD symptoms rated as probably present.
Children were excluded from the CC group for a personal diagnosis of schizophrenia or ADHD, but not for any other psychiatric diagnoses. Nineteen (31%) of the 61 CC probands received at least 1 DSM-III-R diagnosis. The most frequent diagnoses were separation anxiety disorder (n = 5), adjustment disorder (n = 4), oppositional defiant disorder (n = 3), overanxious disorder (n = 3), and simple phobia (n = 3).
Children were excluded from all 3 proband groups if they had a Full-Scale IQ of less than 70 or a central nervous system disease or had taken psychoactive drugs that may have induced psychosis.
Parents of probands were given structured diagnostic interviews (the Diagnostic Interview Schedule29 supplemented with the psychosis section of the Present State Examination30) for Axis I disorders and the Structured Clinical Interview for DSM-III-R Personality Disorders31 for Axis II disorders by a different set of interviewers than the proband diagnostic team. Interviewers of parents were masked to the proband's diagnosis. Family history information was collected on all first- and second-degree relatives, usually from 2 informants using the Relative Family History Interview.32 Direct interviews, family history, and medical record information were reviewed to make consensus DSM-III-R diagnoses. Consensus diagnoses were reviewed by senior clinicians (R.F.A., K.H.N., and D.L.F.) who were always masked to the diagnoses of probands and relatives. Previous reports from our project33-34 described the background and training of family interviewers, the interrater reliability on key diagnoses, and the diagnostic process in more detail.
Five parents of COS probands and 1 parent of an ADHD proband were excluded from this report because they had personal diagnoses of schizophrenia, and 2 parents of COS probands were excluded because of personal diagnoses of schizoaffective disorder with depression. None of the parents of CC probands had diagnoses of schizophrenia or schizoaffective disorder.33
NEUROCOGNITIVE TASKS
The neurocognitive tasks were administered to participants in a fixed order of presentation by research assistants who were masked to the diagnoses of the probands and parents.
The Trails B, from the Halstead-Reitan Battery,12, 35 requires subjects to connect numbers and letters in sequence on a page as quickly as possible, alternating between the 2 sets of characters correctly. The dependent variable is the time to complete the page. This paper-and-pencil test requires efficient visual search, maintenance of the task set, and the ability to alternate between and maintain sequences for 2 series of stimuli.
In the DS-CPT, subjects monitor a rapidly presented series of degraded single digits and respond as quickly as possible by pressing a button each time they see a zero-target stimulus that occurs quasi-randomly on 25% of presentations. Stimuli were presented at 1 per second for 40 milliseconds. Degradation of the stimuli resulted in digits appearing extremely blurred and indistinct.36 Subjects were given 480 trials in an 8-minute vigilance period. The dependent variable was sensitivity (d1), a signal-detection theory measure of signal/noise discrimination. The DS-CPT provides a sensitive measure of subtle deficits in sustained visual signal detection without burdening short-term memory.11
For the Span, subjects report whether a T or F was flashed on a screen with other letters and are told to guess if in doubt.13 Stimuli were presented for 70 milliseconds. One target (T or F) and 0, 2, or 9 other letters were contained in each array. Each array size was presented for 40 trials. The dependent variable was the number of correct target detections for the 10-letter arrays. The Span provides an estimate of the rate of visual information processing.13
The Trails B, Span, and DS-CPT detect neurocognitive impairments in AOS11, 37-38 and ADHD probands36, 39-44 and the relatives of AOS probands,11, 45-48 whereas the Span also detects impairments in COS probands.14
EXPERIMENTAL DESIGN
This study used a case-control design in which CC probands with index disorders (schizophrenia or ADHD) were excluded, but those with any other psychiatric disorder were included. This resulted in the parents of CC probands being more representative of the general population than they would have been if all CC probands with any psychiatric diagnoses were excluded.49
DATA ANALYSIS
One-way analyses of variance (ANOVA) and  2 tests were used to test for group differences in demographic characteristics.
Scores on the 3 neurocognitive tasks were screened for validity by review of summaries of medical history and test results by 2 investigators (R.F.A. and K.H.N.) who were masked to proband and parent diagnoses. The Span data of 8 parents (of 3 COS, 2 CC, and 3 ADHD probands) and the DS-CPT data of 6 parents (of 4 COS, 1 CC, and 1 ADHD probands) were excluded because of poor visual acuity, medical conditions, or equipment failures or because the parents could not read English. Scores corrected for normal age effects were used to test for group differences on the 3 neurocognitive tests. Age corrections were developed by creating a best-fitting model of aging in parents of CC probands. Age-corrected scores were the residuals after the model of normal age effects was applied. We converted Trails B raw scores to reciprocals to normalize distributions, resulting in speed scores. We found a small, linear correlation between DS-CPT scores and test date associated with a subtle fading of the contrast of the stimulus slides. We regressed out this effect by entering test date as a covariate.
Because of prior evidence of sex effects on neurocognitive functioning,50 we evaluated sex effects in the parents of probands. Education was entered as a covariate to statistically control for group differences in education. A mixed-effects ANOVA51 with groups as an interfamily factor, parental sex as an intrafamily factor, and intrafamily education as a covariate was used to test the hypotheses that parents of COS and ADHD probands perform more poorly than parents of CC probands. Significant effects were followed up by means of pairwise post hoc t tests.
Logistic regression was used to evaluate the diagnostic accuracy of the neurocognitive tests in distinguishing between the parents of COS and CC probands and those of COS and ADHD probands. Separate logistic regression was performed for mothers and fathers because of significant sex differences on one of the neurocognitive tasks and because the data from mothers and fathers from the same family are not independent observations and therefore cannot be pooled. For each of the 2 contrasts (COS-CC and COS-ADHD), 3 univariate logistic regressions (the predictors were the individual neurocognitive tasks) and a multivariate logistic regression were conducted (the 3 neurocognitive variables were the predictors) with group as the criteria. These analyses were performed for mothers and fathers separately.
RESULTS
SUBJECT CHARACTERISTICS
We found small but statistically significant differences between the groups of parents in age at testing and the number of years of education. The parents of COS probands had completed fewer years of school than the parents of ADHD and CC probands, who did not differ from each other (Table 1). The parents of ADHD probands were younger than parents of CC probands, who did not differ from parents of COS probands (Table 1). The proportion of mothers interviewed did not significantly differ across the 3 groups.
|
|
|
|
Table 1. Characteristics of Parents of Probands*
|
|
|
GROUP MEAN DIFFERENCES
Because of group differences in the number of years of education, and because there is frequently a significant effect of education on neurocognitive tasks, education was entered as a covariate in the ANOVAs for the age-adjusted scores for the 3 neurocognitive tasks (Table 2). For the d1 index of the DS-CPT, we found statistically significant main effects for group and parental sex, but the group x sex interaction was not significant (Table 2). The fathers obtained significantly better scores on the DS-CPT than the mothers. The parents of COS probands obtained significantly poorer scores than the parents of CC (t228 = 3.33; P = .001) and ADHD (t228 = 2.99; P = .003) probands. The parents of ADHD and CC probands did not differ significantly from each other.
|
|
|
|
Table 2. Performance of Mothers and Fathers of 3 Proband Groups on Neurocognitive Tasks*
|
|
|
For Trails B, we found a significant main effect for group, but the main effect for sex and the group x sex interaction were not significant. The parents of COS probands were significantly slower completing Trails B than the parents of CC (t231 = 3.64; P<.001) or ADHD (t231 = 3.27; P = .001) probands. The parents of CC and ADHD probands did not differ significantly from each other.
For the Span, the main effects of group and sex and the group x sex interaction did not attain statistical significance.
DIAGNOSTIC ACCURACY ANALYSES
We tested the hypothesis that these neurocognitive tasks measure a schizophrenia endophenotype by using age-adjusted neurocognitive scores as predictors in a series of logistic regression analyses to test their ability to discriminate the parents of COS probands from the parents of CC and ADHD probands. Logits generated by the logistic regression analysis indicate the probability that a particular parent has a child with COS. Cutting points can be placed on logits to classify individuals as the parents of COS, CC, or ADHD probands. The effects of varying the cutoff on the sensitivity and specificity of the classification of individuals as parents of COS probands are depicted as receiver operating characteristic curves in Figure 1 and Figure 2. Sensitivity refers to the probability that a parent of a COS proband is classified correctly, whereas specificity is the probability that a parent of a CC proband is incorrectly classified as a parent of a COS proband. The cutoffs reported below optimize the ratio of specificity to sensitivity. Because we found no statistically significant differences on any of the neurocognitive tasks between parents of ADHD and CC probands, diagnostic accuracy analyses were not performed for this contrast.
|
|
|
|
Figure 1. Receiver operating characteristic curves of efficiency of different cutting scores on logits from logistic regression models of the Trail-Making Test B–Adolescent Version (Trails B), Span of Apprehension (Span), Degraded Stimulus–Continuous Performance Test (DS-CPT), and combined-measures discrimination (Combined) between the mothers of childhood-onset schizophrenia (COS) (n = 49) and community control (CC) probands (n = 67) (A) and the fathers of COS (n = 30) and CC probands (n = 48) (B). The straight lines indicate chance levels of discrimination of the parents of COS probands from the parents of CC and attention-deficit/hyperactivity probands. Values of the logit above the line exceed chance levels of discrimination.
|
|
|
|
|
|
|
Figure 2. Receiver operating characteristic curves of efficiency of different cutting scores from logistic regression models of the Trail-Making Test B–Adolescent Version (Trails B), Span, DS-CPT, and combined-measures discrimination between the mothers of COS (n = 49) and attention-deficit/hyperactivity disorder (ADHD) probands (n = 108) (A) and the fathers of COS (n = 30) and ADHD probands (n = 82) (B). The straight lines indicate chance levels of discrimination of the parents of COS probands from the parents of community control and ADHD probands. Values of the logit above the line exceed chance levels of discrimination. Other abbreviations are explained in the legend to Figure 1.
|
|
|
The logistic regression models testing the discrimination between the mothers of COS and CC probands were statistically significant for the DS-CPT (2 = 14.24; n = 110; P<.001) and Trails B (2 = 18.38; n = 112; P<.001), but not for the model for Span. For fathers, the logistic regression model for Trails B was a significant discriminator (2 = 5.53; n = 75; P = .02), whereas the models for the DS-CPT and Span were not. Multiple logistic regression models for the discrimination between the mothers (23 = 25.67; n = 106; P<.001) and fathers (23 = 13.13; n = 71; P = .004) of COS probands and CC probands were statistically significant.
The sensitivity and specificity of the cognitive measures for discriminating between the mothers of COS and CC probands are shown in Figure 1A. At 100.0% specificity, the sensitivity for the Trails B was 14.6% and for the DS-CPT was 11.1%. For the Span, a cutoff producing 97.0% specificity yielded 4.3% sensitivity. For fathers (Figure 1B), a Trails B cutoff that produced 97.9% specificity produced 10.7% sensitivity. For the DS-CPT, a cutoff producing 97.9% specificity yielded 3.4% sensitivity. For the Span, a cutoff that produced 100.0% specificity yielded 14.3% sensitivity.
For all 3 tasks, lowering the cutoff for designating neurocognitive impairment in both the mothers and the fathers decreased the ratio of specificity to sensitivity.
The multiple logistic model was more efficient than any single neurocognitive task for discriminating between the parents of COS and CC probands. For mothers, 20.0% sensitivity resulted in 100.0% specificity. Lowering the cutoff to 98.4% specificity increased sensitivity only slightly to 25.0%. For fathers, a cutoff that produced 100.0% specificity resulted in 20.0% sensitivity. Decreasing the cutoff to 93.4% specificity increased sensitivity to 32.0%.
For the discrimination between the mothers of COS and ADHD probands, the DS-CPT (21 = 12.21; n = 152; P<.001) and Trails B (21 = 15.01; n = 155; P<.001) were significant, whereas the Span was not. For fathers, the Trails B was a significant discriminator (21 = 5.42; n = 110; P = .02), whereas the DS-CPT and Span were not. The multiple logistic regression models for the discrimination between the mothers (23 = 21.60; n = 150; P<.001) and fathers (23 = 9.87; n = 105; P = .01) of COS and ADHD probands were both statistically significant.
For models testing the discrimination between the mothers of COS and ADHD probands (Figure 2A), the sensitivity for the Trails B was 14.6% and for the DS-CPT was 13.3% at 100.0% specificity. For the Span, a cutoff producing 99.1% specificity yielded 4.3% sensitivity. For fathers (Figure 2B), the sensitivity for the Trails B was 3.6% at 97.6% specificity, whereas the sensitivity for the DS-CPT was 3.4% at 98.7% specificity. For the Span, a cutoff producing 98.8% specificity yielded 3.6% sensitivity.
For each of the tasks, lowering the cutoff for cognitive impairment decreased the ratio of sensitivity to specificity for both mothers and fathers.
The multivariate models for the mothers (Figure 2A) produced 100.0% specificity with 12.2% sensitivity, whereas lowering the cutoff to produce 98.1% specificity resulted in 20.5% sensitivity. For the fathers (Figure 2B), the neurocognitive measures did not work quite as well. A cutoff producing 98.5% specificity had 16.6% sensitivity.
COMMENT
Consistent with previous studies of relatives of patients with AOS, the parents of COS probands showed neurocognitive impairments compared with the parents of CC probands. The presence of psychophysiologic and linguistic impairments and structural and functional brain abnormalities,18, 52-54 the response to typical and atypical antipsychotic medications,18, 52 and the aggregation of schizophrenia and schizophrenia spectrum personality disorders in parents25 is similar in COS and AOS. The results of the current study provide important additional evidence of etiologic continuity between COS and AOS.
The analyses that adjust for education effects could have underestimated the magnitude of the differences in neurocognitive functioning between parents of COS and CC probands. The lower educational attainments of the nonpsychotic parents of COS probands may partially reflect the subtle cognitive impairments tapped by the 3 neurocognitive tasks, thereby reversing the assumption about the direction of causality that underlies statistical correction for educational differences.55
By excluding parents with psychoses, this study demonstrated that neurocognitive impairments occur in the relatives of COS probands independent of the presence of schizophrenia.
The diagnostic accuracy analyses show that a combination of scores on neurocognitive tasks is better at differentiating between the parents of COS probands and those of CC and ADHD probands than scores from any single task. A useful indicator of the schizophrenia phenotype for linkage studies, an endophenotypic indicator, should have a ratio of at least 10:1 of true- to false-positive findings.8 A cutoff using a combination of scores across the 3 tasks correctly classifies 20.0% of the mothers and fathers of COS probands compared with 0% of the mothers and fathers of CC probands. Setting a less rigorous cutoff for neurocognitive impairment decreases the ratio of true- to false-positive findings.
Performance on the 3 neurocognitive tasks is somewhat less efficient in differentiating the parents of COS and ADHD probands, but it still shows diagnostic specificity.
A substantial subgroup of the parents of COS probands has worse neurocognitive performance than that of any of the parents of ADHD and CC probands. These results replicate and extend a previous report8 indicating that the combination of scores on certain neurocognitive tasks produces a level of diagnostic accuracy sufficient for use in genetic linkage studies. The levels of diagnostic accuracy in this study and the previous report8 are quite similar, despite the differences in neurocognitive tasks used in the studies.
In conjunction with the results of a previous study,8 these results have potentially important implications for genetic linkage studies. Including a rigorous definition of neurocognitive impairment as an index of the schizophrenia phenotype can add statistical power to linkage studies in 2 ways. First, when using models that dichotomize affected status, it can increase the number of individuals considered to be affected within a given family. In our study, even when neurocognitive impairment is rigorously defined so as to have very high specificity, the number of parents with neurocognitive impairments (20% of mothers and fathers) is much greater than the number of parents who obtain diagnoses of schizophrenia (4.9%).25 Second, neurocognitive measures yield continuous variables. When continuous scores of traits correlated with schizophrenia are used, pedigrees are more informative because "individuals can be assessed on a continuous scale that provides a greater range of information on ‘unaffected' individuals who are not yet through the risk period or who are near the liability threshold for affection."56(p170)
The cutting scores generated by logistic regression in the present study optimize the separation between groups. These results, therefore, must be replicated before firm conclusions can be made about the diagnostic accuracy of the neurocognitive tasks and before cutting scores can be established to identify the parents of COS probands carrying liability to schizophrenia.
The present study did not address the range of neurocognitive impairments and/or the specific nature of the cognitive processes that are impaired in the relatives of COS probands. The relatives of AOS probands are impaired on neurocognitive tasks that tap quite diverse cognitive processes.8, 57 The Span, Trails B, and DS-CPT also tap multiple cognitive processes. Identification of the specific cognitive processes that reflect liability to schizophrenia requires using tasks that tap more circumscribed cognitive processes than those used in the present and previous studies and the selection of appropriate control tasks.
Consistent with the results of the major, most rigorous study19 examining neurocognitive impairments in the siblings of ADHD probands, we found no significant differences in neurocognitive functioning between the parents of ADHD and CC probands. This was despite the fact that the tasks administered to the parents of ADHD probands had been demonstrated in previous research to distinguish between patients with ADHD and controls. In contrast to the role of neurocognitive impairments in schizophrenia, where neurocognitive impairments appear to index the liability to schizophrenia in nonpsychotic relatives, ADHD neurocognitive impairments are present primarily in individuals with a diagnosis of ADHD and are thus only part of the active disorder. Our finding that neurocognitive impairments in schizophrenia, but not ADHD, are an expression of liability parallels the findings of a study58 comparing neurocognition in relatives of schizophrenia and bipolar probands.
AUTHOR INFORMATION
Submitted for publication July 26, 2001; final revision received March 13, 2002; accepted March 15, 2002.
This study was supported by research grants MH 45112, MH 49716, MH 46981, MH 30911, and MH 14584 from the National Institute of Mental Health, Rockville, Md, and by the Della Martin Foundation, Los Angeles, Calif.
We acknowledge the invaluable assistance of the families who participated in this study and the agencies (Children and Adults with Attention Deficit Hyperactivity Disorder–CHADD, Los Angeles County Department of Mental Health, and Skip Baker, MD) that referred families to us. We also thank Carol Giannini, MPH, Tamsen Thorpe, PhD, Heidi Kuppinger, PhD, and Lisa Thomrongsith for their assistance in data collection and Beth Tang, MS, Debbie Weissman, Christie Ong, MPH, and Sun Hwang, MS, for assisting in data analysis and manuscript preparation.
Corresponding author and reprints: Robert F. Asarnow, PhD, Department of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles, Room 48-240C NPI, 760 Westwood Plaza, Los Angeles, CA 90024-1759 (e-mail: rasarnow{at}medmet.ucla.edu).
From the Neuropsychiatric Institute and the Department of Psychiatry and Biobehavioral Sciences (Drs Asarnow, Nuechterlein, Subotnik, Fogelson, Payne, Asamen, Mintz, and Guthrie and Mr Torquato), the Department of Psychology (Drs Asarnow and Nuechterlein), the Mental Retardation Research Center (Drs Asarnow and Guthrie), and the Department of Biostatistics (Dr Guthrie), University of California, Los Angeles; and the Graduate School of Education and Psychology, Pepperdine University, Malibu, Calif (Dr Asamen).
REFERENCES
| |
1. Tsuang MT, Gilbertson MW, Faraone SV. The genetics of schizophrenia: current knowledge and future directions. Schizophr Res. 1991;4:157-171.
FULL TEXT
|
ISI
| PUBMED
2. Kendler KS, Diehl SR. The genetics of schizophrenia: a current, genetic-epidemiologic perspective. Schizophr Bull. 1993;19:261-285.
3. Kendler KS. The genetics of schizophrenia. In: Kaplan HI, Sadock BJ, eds. Comprehensive Textbook of Psychiatry. Baltimore, Md: Williams & Wilkins; 1995: 942-957.
4. Faraone SV, Biederman J. Genetics of attention-deficit hyperactivity disorder. Child Adolesc Psychiatr Clin N Am. 1994;3:285-302.
5. Hechtman L. Families of children with attention deficit hyperactivity disorder: a review. Can J Psychiatry. 1996;41:350-360.
ISI
| PUBMED
6. Meehl PE. Schizotaxia, schizotypy, schizophrenia. Am Psychol. 1962;17:827-838.
FULL TEXT
7. Kremen WS, Seidman LJ, Pepple JR, Lyons MJ, Tsuang MT, Faraone SV. Neuropsychological risk indicators for schizophrenia: a review of family studies. Schizophr Bull. 1994;20:103-119.
8. Faraone SV, Seidman LJ, Kremen WS, Pepple JR, Lyons MJ, Tsuang MT. Neuropsychological functioning among the nonpsychotic relatives of schizophrenic patients: a diagnostic efficiency analysis. J Abnorm Psychol. 1995;104:286-304.
FULL TEXT
|
ISI
| PUBMED
9. Asarnow J. Children at risk for schizophrenia: converging lines of evidence. Schizophr Bull. 1988;14:613-631.
10. Cornblatt BA, Keilp JG. Impaired attention, genetics, and the pathophysiology of schizophrenia. Schizophr Bull. 1994;20:31-46.
11. Nuechterlein KH. Vigilance in schizophrenia and related disorders. In: Steinhauer SR, Gruzelier JH, Zubin J, eds. Neuropsychology, Psychophysiology and Information Processing. New York, NY: Elsevier Science Inc; 1991:397-433. Handbook of Schizophrenia; vol 5.
12. Reitan RM. Validity of the Trail-Making Test as an indicator of organic brain damage. Percept Mot Skills. 1958;8:271-276.
13. Asarnow RF, Granholm E, Sherman T. Span of Apprehension in schizophrenia. In: Steinhauer SR, Gruzelier JH, Zubin J, eds. Neuropsychology, Psychophysiology and Information Processing. New York, NY: Elsevier Science Inc; 1991:335-370. Handbook of Schizophrenia; vol 5.
14. Asarnow RF, Asamen J, Granholm E, Sherman T, Watkins J, Williams M. Cognitive/neuropsychological studies of children with a schizophrenic disorder. Schizophr Bull. 1994;20:647-670.
15. Strandburg RJ, Marsh JT, Brown WS, Asarnow RF, Guthrie D, Higa J. Continuity of information processing deficits across childhood and adult onset schizophrenia: ERP correlates. Schizophr Bull. 1994;20:685-696.
16. Zahn TP, Jacobsen LK, Gordon CT, McKenna K, Frazier JA, Rapoport JL. Autonomic nervous system markers of psychopathology in childhood-onset schizophrenia. Arch Gen Psychiatry. 1997;54:904-912.
ABSTRACT
17. Jacobsen LK, Hong WL, Hommer DW, Castellanos FX, Frazier JA, Giedd JN, Gordon CT, Karp BI, McKenna K, Rapoport JL. Smooth pursuit eye movements in childhood-onset schizophrenia: comparison with ADHD and normal controls. Biol Psychiatry. 1996;40:1144-1154.
FULL TEXT
|
ISI
| PUBMED
18. Jacobsen L, Rapoport J. Research update: childhood-onset schizophrenia: implications of clinical and neurobiological research. J Child Psychol Psychiatry. 1998;39:101-113.
FULL TEXT
|
ISI
| PUBMED
19. Sowell ER, Levitt J, Thompson PM, Holmes CJ, Blanton RE, Kornsand DS, Caplan R, McCracken J, Asarnow RF. Brain abnormalities in early onset schizophrenia spectrum disorder observed with statistical parametric mapping of structural magnetic resonance images. Am J Psychiatry. 2000;157:1475-1484.
FREE FULL TEXT
20. Faraone SV, Kremen WS, Lyons MJ, Pepple JR, Seidman LJ, Tsuang MD. Diagnostic accuracy and linkage analysis: how useful are schizophrenia spectrum phenotypes? Am J Psychiatry. 1995;152:1286-1290.
FREE FULL TEXT
21. Risch N. Linkage strategies for genetically complex traits, II: the power of affected relative pairs. Am J Hum Genet. 1990;46:229-241.
ISI
| PUBMED
22. Risch N. Linkage strategies for genetically complex traits, I: multilocus models. Am J Hum Genet. 1990;46:222-228.
ISI
| PUBMED
23. Welner Z, Welner A, Stewart M, Palkes H, Wish E. A controlled study of siblings of hyperactive children. J Nerv Ment Dis. 1977;165:110-117.
PUBMED
24. Faraone SV, Biederman J, Krifcher Lehman BL, Spencer T, Norman D, Seidman L, Krause I, Perrin J, Chen W, Tsuang MT. Intellectual performance and school failure in children with attention deficit hyperactivity disorder and their siblings. J Abnorm Psychol. 1993;102:616-623.
FULL TEXT
|
ISI
| PUBMED
25. Seidman LJ, Biederman J, Monuteaux MC, Weber W, Faraone SV. Neuropsychological functioning in non-referred siblings of children with attention deficit/hyperactivity disorder. J Abnorm Psychol. 2000;109:252-265.
FULL TEXT
| PUBMED
26. Orvaschel E, Puig-Antich J. Schedule for Affective Disorder and Schizophrenia for School Aged Children–Epidemiologic Version: (K-SADS-E). Philadelphia: Medial College of Pennsylvania; 1987.
27. Russell AT, Bott I, Sammons C. The phenomenology of schizophrenia occurring in childhood. J Am Acad Child Adolesc Psychiatry. 1989;28:399-407.
ISI
| PUBMED
28. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition. Washington, DC: American Psychiatric Association; 1987.
29. Robins LN, Helzer JE, Croughan J, Ratcliff KS. National Institute of Mental Health Diagnostic Interview Schedule: its history, characteristics and validity. Arch Gen Psychiatry. 1981;38:381-389.
ABSTRACT
30. Wing JK, Cooper JE, Sartorius N. The Measurement and Classification of Psychiatric Symptoms: An Instruction Manual for the PSE and CATEGO Programs. New York, NY: Cambridge University Press; 1974.
31. Spitzer RL, Williams JBW. Structured Clinical Interview for DSM-III-R Personality Disorders. New York: Biometrics Research Dept, New York State Psychiatric Institute; 1989.
32. Gershon ES. Relative Psychiatric History (RPH) Symptom Checklist Interview. Bethesda, Md: National Institute of Mental Health; 1985.
33. Asarnow RF, Nuechterlein KH, Fogelson D, Subotnik KL, Payne DA, Russell AT, Asamen J, Kuppinger H, Kendler KS. Schizophrenia and schizophrenia spectrum personality disorder in the first-degree relatives of children with schizophrenia: the UCLA Family Study. Arch Gen Psychiatry. 2001;58:581-588.
FREE FULL TEXT
34. Fogelson DL, Nuechterlein KH, Asarnow RF, Subotnik KL, Talovic SA. Interrater reliability of the Structured Clinical Interview for DSM-III-R, Axis II: schizophrenia spectrum and affective spectrum disorders. Psychiatry Res. 1991;39:55-63.
FULL TEXT
|
ISI
| PUBMED
35. Reitan RM. Investigation of the validity of Halstead's measures of biological intelligence. Arch Neurol Psychiatry. 1955;73:28-35.
36. Mariani MA, Barkeley RA. Neuropsychological and academic functioning in preschool boys with attention deficit hyperactivity disorder. Dev Neuropsychol. 1997;13:111-129.
37. Asarnow RF, MacCrimmon DJ |
