Eye Movement Task Related to Frontal Lobe Functioning in Children with Attention Deficit Disorder

Eye Movement Task Related to Frontal Lobe Functioning in Children with Attention Deficit Disorder RANDAL G. ROSS, M.D., DANIEL HOMMER, M.D., DAVID BREIGER, PH.D., CHRISTOPHER VARLEY, M.D., AND ALLEN RADANT, M.D.

ABSTRACT Objective: Attention-deficit hyperactivity disorder (ADHD) has been postulated to be related to dysfunction of the

prefrontal cortex. In the oculomotor delayed response task, a subject is cued as to where he or she should look (shift visual gaze to) but must delay a short period and then shift gaze to the location where the cue previously existed but no longer exists (a memory-guided saccade). Dependent measures from this task provide information on three functions tentatively tied to prefrontal cortex functioning: the ability to inhibit response (during the delay period), preparation of motor response (inversely tied to the latency of shifting visual gaze), and accuracy of working visuospatial memory (accuracy of the memory-guided saccade). Method: Thirteen children with ADHD and 10 normal controls, aged 9 to 12 years, were tested using an 800-msec delay period. Results: Children with ADHD showed, relative to normal controls, deficits on inhibiting response during the delay period but no differences in latency (preparation of motor response) or accuracy of visuospatial memory. Conclusions: These results support the hypothesis that the primary deficit in ADHD is difficulty in inhibition of response. This deficit may be associated with pathology located outside the dorsolateral prefrontal cortex. J. Am. Acad. Child Ado/esc. Psychiatry, 1994, 33, 6:869-874. Key Words: eye movements, children, saccades, attention-deficit hyperactivity disorder, visuospatial working memory.

There has been increasing interest in the relationship between prefrontal cortex functioning and the deficits found in children with attention-deficit hyperactivity disorder (ADHD). Children with frontal lobe lesions show impulsive hyperactive behavior (Benton, 1991; Grattan and Eslinger, 1991), and adolescents with ADHD show decreased anterior frontal lobe activity on positron emission tomography (Zametkin et al., 1993). Performance on neuropsychological tests purported to test frontal cortex functioning is deficient in children with ADHD (Barkley et aI., 1992; Grodzinsky Accepted February 3, 1994. Drs. Ross, Breiger, Varley, and Radantarewith the University o/Washington and Dr. Hommer is at the NationalInstitute 0/Alcohol Abuseand Alcoholism, NationalInstitutes ofHealth. Dr. Ross is currently at the University ofColorado Health Sciences Center in Denver. This research was supported, in part, by a Holmes' Scholarship Award. Special thanks to Judith Driscoll, R.N., M.S., C.S., A.R.N.P.; Patti Leger, M.N., c.N.S., A.R.N.P.; and Ellyn Cavanaugh, M.N., A.R.N.P., for their help in subject recruitment. Reprintrequests to Dr. Ross, Department 0/Psychiatry, Campus Box C26871, University ofColorado Health Sciences Center, 4200 E. 9th Ave., Denver, CO 80262; telephone: (303) 270-5368; fax: (303) 270-5347. 0890-8567/94/3306-0869$03.00/0©1994 by the American Academy of Child and Adolescent Psychiatry.

and Diamond, 1992), although tests of response inhibition (e.g., go-no go tasks) seem the most consistently sensitive to ADHD-control differences (Barkley et aI., 1992). Recently, oculomotor delayed response tasks have received attention as measures of prefrontal cortical function. This type of task consists of three sequential steps. First, the subject is asked to fixate his or her gaze on a specific spatial location while a cue stimulus is briefly displayed in another location. The subject then must maintain his or her gaze on the fixation point (and inhibit eye movement to the cued location) during a brief delay period (during which the cue stimulus is no longer present but the fixation stimulus remains on). Finally, when the fixation point disappears, the subject is to move his or her gaze to the spatial location at which the cue stimulus was previously present. The accuracy of the saccade to the remembered location (a memory-guided saccade) is used to assess the accuracy of working visuospatial memory. Versions of this task have been used in monkeys in combination with single cell recording techniques (Barone and Joseph, 1989; Funahashi et aI., 1989) and lesion studies

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 33:6, JULY/AUGUST 1994

869

ROSS ET AL.

(Funahashi et al., 1993) to demonstrate the necessity of intact functioning of the dorsolateral prefrontal cortex in this type of working visuospatial memory. Fuster (1989) has argued that the prefrontal cortex is involved, in addition to visuospatial working memory, in both the inhibition of response and in preparing the motor system to respond (preparatory set). Visuospatial memory and preparatory set may be related to functioning in the dorsolateral prefrontal cortex (Barone and Joseph, 1989; Funahashi et al., 1989, 1993; Fuster, 1989) and its thalamic, basal ganglia, and temporal lobe connections (Fuster, 1989). Inhibition of response may involve the dorsolateral prefrontal cortex (Pierrot-Deseilligny et al., 1991a) and may also involve other cortical areas (Fuster, 1989; Grattan and Eslinger, 1991). Hommer and Radant (unpublished data) have noted that the oculomotor delayed response task actually provides data on each of the three primary prefrontal functions. In addition to the accuracy of the delayed oculomotor response, the ability to inhibit response during the delay period (inhibition of response) and latency of the delayed response (inversely related to successat formation of a preparatory set) are dependent measures provided by an oculomotor delayed response task that may measure different aspects of prefrontal cortical functioning. In humans, adult patients with dorsolateral prefrontal (on either side) lesions (due to cerebral infarction) show decreased accuracy, increased latency of memoryguided saccades, and impaired saccadic inhibition, whereas patients with dorsomedial frontal cortex lesions do not show such deficits (Pierrot-Deseilligny et al., 1991b). Similarly, schizophrenia has been proposed to be a syndrome characterized by prefrontal cortex dysfunction (see Park and Holzman, 1992, for a review), and adult patients with schizophrenia show impairment on all three dependent measures: decreased accuracy of memory-guided saccades, increased latency of memory-guided saccades (Hommer and Radant, unpublished data; Park and Holzman, 1992), and decreased ability to inhibit (Hommer and Radant, unpublished data). Although ADHD has also been proposed to be a psychopathological state related to dysfunction in the prefrontal cortex, to our knowledge children with ADHD have not been examined with oculomotor delayed response tasks. The purpose of this study is

870

to use an oculomotor delayed response task to assess prefrontal cortical functioning in children with ADHD.

METHOD Subjects Subjects were 13 children with AOHO and 10 normal controls. Children were recruited from the University of Washington Outpatient Pediatric Psychopharmacology Clinics and were about equally divided between children in ongoing treatment with stimulant medication and those who had just been diagnosed and had just completed a double-blind, placebo-controlled trial with methylphenidate (a standard clinical tool in our clinic; Varley and Trupin, 1983). To be included in the AOHO group, children had to (1) meet DSM-III-R criteria for AOHO based on clinical interview and school information; (2) score 15 or greater on a parent-report checklist of Conners (1973); (3) demonstrate methylphenidate responsiveness, as defined by a 25% or greater decrease on either the teacher-report or the parent-report checklist, during a 3-week double-blind, placebo-controlled methylphenidate trial (patterned after Varley and Trupin, 1983); and (4) have an IQ estimate of greater than 75. Mean daily methylphenidate dose (::!::SO) was 35.8 (::!:: 17.8) mg. Control children had, by parent report, no history of attentional or learning disabilities and no history of major psychiatric or neurological disorders. All children were between 110.9 months (9 years, 2.9 months) and 155.8 months (12 years, 11.8 months) of age. Age was not significantly different between the groups (control mean age ::!:: SO = 138.3 ::!:: 12.4 months versus AOHD mean age::!:: SO = 134.5 ::!:: 15.9 months; t = 0.62, not significant [NS]); IQ (as extrapolated from the Vocabulary and Block Design subtests of the WISC-III; Weschler, 1991) also did not significantlydiffer between the groups (control mean extrapolated IQ ::!:: SO = 116 ::!:: 16 versus AOHD extrapolated IQ = 107 ::!:: 12; t = 1.5, NS). Because we have previouslyfound no effectsof gender on any eye movement measure (Ross et al., 1993, in press, a.b), we did not match for gender in this initial study. The AOHD group consisted of 13 males; the control group, 5 males and 5 females. All subjects were tested on delayed oculomotor response tasks twice, with testing sessions approximately 1 week apart. Time of day was not controlled for in this initial study, although the majority of children were tested between 9:00 and 11:30 in the morning. For children in the AOHO group, one testing session occurred while on methylphenidate, one session after methylphenidate had been temporarily discontinued for approximately 1 week (minimum period of discontinuation, 4 days). When on stimulant medication, children took medication as prescribed (generally 1 to 3 hours before testing). Children in the AOHO group were randomly assigned as to whether the first session was conducted while they were taking medication or were medication free. All legal guardians gave informed consent, and all children had the procedure explained to them and gave verbal or written assent.

Eye Movement Procedure Children were allowed to tour the eye movement laboratory before initiating the procedure. All children were allowed to ask any questions and to take as much time as needed between aspects of the procedure. Parents could remain with the child at the

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 33:6, JULY/AUGUST 1994

MEMORY-GUIDED SACCADES IN ADHD

child's request. Children appeared relaxed and not stressed by the experience. Subjects were seated 43 cm in front of a video monitor on which a small target was displayed against a black background in an otherwise dark room. The subject's head was stabilized with a bite bar and head rest. Horizontal eye movements were recorded using an infrared photoelectric limbus detection eye-tracking device (Eye-trac model 210, Applied Sciences Laboratories, Waltham, MA), which is accurate to within 0.25° of visual angle and has a time constant of 4 msec. The analog output of the device is sampled at 1000 Hz using a 12-bit analog-to-digital converter. Data were collected only from the eye for which the most rapid and accurate calibration could be obtained.

Delayed Oculomotor Response Task The delayed oculomotor task is displayed in Figure 1. A fixation point is present for 1500 msec. Then for 100 msec, a cue stimulus is presented concurrently with the fixation point. This is followed by an 800-msec period with the fixation stimulus presented alone

11 y/o male normal control

.

Cue

Memory-guided saccade gain = 0.87 latency = 207 ms

9 y/o male with ADHD

..

Cue

Premature saccade gain = 1.11 latency = 394 ms

~I

.,

! ~iiistiiiim-uiiluliisilllll@!!l--. 1 second Fig. 1 Delayed oculomotor response task. Note the l l-year-old normal controlchildexhibits a memory-guided saccade that occurs afterthe fixation stimulushasbeenextinguished. The 9-year-old with attention-deficit hyperactiviry disorder (ADHD) exhibits a saccade during the delay period (a premature saccade). Gain = saccadic amplitudeltarget step size. Latency is the amount of time from fixation offset (memory-guided saccade) or cue onset (premature saccade).

(total time of continuous fixation point presentation is 2.4 seconds). This 800 msec is the delay period, during which the location of the cue stimulus must be maintained in working visuospatial memory. After the delay period, the fixation point extinguishes, and this is the signal to make a memory-guided saccade. There is a 600-msec period of blank screen during which the memoryguided saccade should occur, followed by the cue light reappearing. This cue light then becomes the new fixation point for the next trial. Cue stimuli could occur to either the left or the right of the fixation stimulus at distances of 3.75°, 7.5°, 15°, 22.5°, or 30°. Stimuli were presented in an identical pseudorandom order to all subjects. Task length was 31.5 seconds. Children completed two similar versions of the task on each visit. Instructions to the child consisted of explaining the task ("You watch the first dot. When a second dot appears, don't look at it, but remember where it was. When the first dot disappears, look where the second dot was." The instructions are repeated to explain the repetitive nature of the task.) The child was then asked to explain the procedure back to the experimenter to ensure understanding. Children seemed to easily comprehend task instructions.

Eye Movement Analyses All eye movement data were analyzed with a computerized pattern recognition program developed by one of the authors (A.R.). This analysis system has been described elsewhere (Ross et al., in press, a) and will be briefly described here. Raw data consist of eye position and target position for each millisecond of recorded tracking. Eye movements are divided into discrete segments. Saccades are identified on the basis of peak velociry (greater than 30 0/second), initial acceleration (greater than 2000 0/second 2 ) , and minimum duration (9 msec). Arrifactual segments caused by eye blinks and head movements show distinct morphology and are removed from the analysis by the pattern recognition software. Output from the software package includes, for each saccade, latency from the most recent target change and accuracy of the saccade (calculated both as position error and saccadic gain = saccade size/target step size). The process the brain uses to calculate direction and amplitude for a visually triggered saccade is thought to require at least 80 to 90 msec (Becker, 1989), and therefore any saccade must be considered to be representing brain processes that occurred at least 80 msec earlier. For purposes of these analyses, the first saccade that occurred at least 80 msec after cue onset, that was in the same direction as the cue, and that was at least 50% of the distance to the cue was considered the primary saccade for that cue presentation. Primary saccades were also classified based on time relative to cue activiry. Those primary saccades that occurred from 80 msec after cue onset to 80 msec after fixation point extinction (i.e., during the time when the subject was supposed to maintain gaze on the fixation stimulus) were classified as premature saccades. Those saccades that occurred from 80 msec after fixation point extinction to 80 msec after cue reappearance (i.e., those primary saccades that represented successful completion of the task) were defined as memory-guided saccades. Finally, when the primary saccade did not occur until more than 80 msec after cue reappearance, the primary saccade was classified as visually guided.

Data Analyses There are three dependent measures of interest: 1. The percentage of trials with premature saccades is a measure of the ability to inhibit responding to the remembered location until cued by fixation offset.

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 33:6, JULY/AUGUST 1994

871

ROSS ET AL.

2. The latency of memory-guided saccades is a measure of the abiliry to prepare the motor system for saccade generation. Shorrer latency indicates increased preparedness of the motor system. 3. Accuracy of the memory-guided saccade is a measure of the accuracy of cue location in working memory. The dependent measure is saccadic gain (saccadic amplitude/rarget step amplitude). The effects of group membership (ADHD versus control), medication status (for the ADHD children), and visit number (to assess for possible learning effects) are examined.

RESULTS

Normal controls showed no differences across testing sessions on any dependent measure (t = 0.4 to 0.8, NS) (Table 1). Children with ADHD showed no differences based on medication status on any dependent measure (t = 0.2 to 1.1, NS) (Table 1). All trials are collapsed across testing session and medication status (by combining all data points from all trials) to facilitate comparison between children with ADHD and normal controls. Because the variability is significantly different between groups, dependent measures between groups are compared with nonparametric tests. Table 2 shows the data for each of these three dependent measures. Children with ADHD were significantly more likely to exhibit a premature saccade than were normal control children (Mann-Whitney U = 31.5, P < .04), but the two groups showed no significant differences on either the latency or saccadic gain of the memory-guided saccade (U = 64.0 and 43.0, NS, respectively). After further examination of the premature saccades, we found that the delay period was divided in half (latencies of 80 to 530 msec after cue onset, and latencies of 53 1 to 980 msec after cue onset). Children with ADHD exhibited 46.8 ± 32.2% of premature

saccades in the first half of the delay period compared to 54.4 ± 38.0% for normal control children (t = 0.5, NS). DISCUSSION

Children with ADHD exhibited more premature saccades than did normal control children. However, 22 of the 23 children in this study (all except one normal control) exhibited premature saccades, suggesting that some difficulty with inhibiting response is normal in this age group. Premature saccades may represent either a reflexive saccade toward the cue immediately after the cue appears or an inability to inhibit responding to the remembered location of the cue stored in working memory. If most premature saccades are reflexive, premature saccades should have latencies close to visually guided saccades (normal visually guided saccadic latencies in this age range are 203 ± 27 msec; Ross et al., in press, a). If most premature saccades represent failure to inhibit response to information coded in working memory, one would expect most saccades to have longer latencies. In this study, both children with ADHD and normal control children had close to 50% ofpremature saccades with latencies of greater than 530 msec, suggesting that many premature saccades represent an inability to inhibit response to information coded in working memory. Of interest is the finding that the methylphenidateresponsiveADHD children showed no effect ofmethylphenidate on their rate of premature saccades (their ability to inhibit saccades during the delay period). It may be possible that this type of inhibition is different from that involved with behavioral inhibition, that this measure is more sensitive to ADHD pathology and may not be helped by stimulant medication, or that

TABLE 1 An Oculomotor Delayed Response Task: Comparison across Visit Number and Medication Status

% of trials with premature saccades Latency of memory-guided saccades (msec) Gain of memory-guided saccades

AOHO Medicated

AOHO Un medicated

t

0.8

28.0 ± 24.1

30.6 ± 23.8

0.4

289 ± 52

0.4

277 ± 55

281 ± 62

0.2

0.96 ± 0.04

0.6

0.89 ± 0.07

0.95 ± 0.15

1.1

Control Visit 1

Control Visit 2

17.0 ± 14.7

13.2 ± 13.2

283 ± 42 0.94 ± 0.09

Note: Results are means ± SO. t is the result of a repeated-measures t test. All comparisons have probabilities greater than .28 and are considered nonsignificant. ADHD = attention-deficit hyperactivity disorder.

872

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 33:6, JULY/AUGUST 1994

MEMORY-GUIDED SACCADES IN ADHD

TABLE 2 An Oculomotor Delayed Response Task: Comparison of Children with ADHD and Normal Controls Children with ADHD

Normal Controls

U

p

31.7 ± 23.8

15.2 ± 11.8

31.5

.04

281 ± 45

286 ± 44

64.0

NS

0.91 ± 0.08

0.95 ± 0.04

43.0

NS

% of trials with

premature saccades Latency of memory-guided saccades (msec) Gain of memory-guided saccades

Note: Results are means ± SD. U is the result of a Mann-Whitney comparison. p is the probability of that comparison. Probabilities less than .05 are considered significant. ADHD = attention-deficit hyperactivity disorder; NS = not significant.

the effect is dose sensitive and may respond to higher doses of methylphenidate than those required for behavioral improvement. If this finding is reproducible, it may be possible to test for problems in inhibition even in medicated children. Although children with ADHD had more difficulty than normal control children at inhibiting a saccade to the cue stimulus, children with ADHD appeared to be similar to control children in their ability to prepare the motor system to respond and in the accuracy (as measured by saccadic gain) of the memoryguided saccade. This pattern of deficits is very different from the pattern of deficits demonstrated in adults with either schizophrenia (Hommer and Radant, unpublished data; Park and Holzman, 1992) or dorsolateral prefrontal cortex lesions (Pierrot-Desceilligny et al., 1991b). (Adults with schizophrenia and adults with dorsolateral prefrontal cortex lesions demonstrate deficits on all three delayed oculomotor response measures.) There are at least three explanations for this discrepancy in findings: developmental changes, differences in the psychopathological underpinnings of the disorders, and task differences. One could argue that developmental differences may account for the differences in results found here for children with ADHD here as compared to those previously reported for adults with dorsolateral prefrontal cortex lesions or with schizophrenia. The prefrontal cortex reportedly undergoes major developmental changes at least through the mid-lOs (Weinberger, 1987), and the systems responsible for motor preparations and visuospatial working memory may be insufficiently developed to show deficits in any psychopathological group. However, incomplete development would predict poor performance, and children in this

study performed well on motor preparation and visuospatial memory portions of the task. However, further research exploring the full normal developmental spectrum is indicated. If developmental changes do not explain the discrepancy in results between children with ADHD and adults with schizophrenia, differences in the underlying psychopathology might. Fuster (1989) has postulated that inhibition of response, motor preparation, and visuospatial working memory are located in different parts of the prefrontal cortex. If there is anatomical and physiological separation of these three functions, it may be possible to show a deficit on only one function. Children with ADHD who show methylphenidate responsiveness may have their deficit primarily limited to inhibition of response, a finding consistent with other neuropsychological test results (Barkley et al., 1992) and suggestive of dysfunction in brain regions other than dorsolateral prefrontal cortex (Fuster, 1989). Finally, children with ADHD may have impairments on visuospatial memory and motor preparation, but this version of an oculomotor delayed response task may be insensitive to this difference. Although this oculomotor delayed response task is patterned after the findings of Hommer and Radant (unpublished data), Hommer and Radant used a 1200-msec delay period as compared to 800 msec in this protocol. In lesioned primates, delay differences as small as 1000 msec can significantly alter the degree of deficit (Funahashi et al., 1993), and so ADHD children may show more widespread deficits if the delay period is increased. However, in this comparatively simple delayed oculomotor response task, inhibition of response is the primary deficit. This suggests that inhibition of re-

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 33:6, JULY/AUGUST 1994

873

ROSS ET AL.

sponse is the greatest impairment of the three functions in children with ADHD. These conclusions must be considered tentative given the small sample size, the failure to assess subjects over a broader age range (to address issues of development), and the failure to concurrently test children with other presumed frontal lobe-mediated psychopathology (e.g., children who, as offspring of schizophrenic parents, are at increased risk to later develop schizophrenia). Additionally, future studies should consider the use of a variety of delay lengths to assess the possible deterioration of working visuospatial memory over time, and/or an interaction between this delay and diagnostic group membership. The frontal cortex is proposed to be intricately involved in multiple higher cortical functions, including intention to inhibit response, directing attention, determining goals, setting patterns of priority, and creating and executing a plan to reach goals (Mateer and Williams, 1991; Welsh and Pennington, 1988). Multiple child and adolescent psychopathological conditions have been postulated to be related to dysfunction of the frontal cortex, including ADHD (Barkley et al., 1992), schizophrenia (Park and Holzman, 1992), and autism (Rogers and Pennington, 1991). Because portions of the frontal cortex may not complete development until the mid-20s or later (Weinberger, 1987), frontal cortex-mediated psychopathology symptom presentation may be difficult to distinguish from normal developmental changes, or, because of incomplete development of the brain, may look similar in different diagnostic groups. It would be premature at this point to formulate a role for this eye movement task in clinical care. However, by combining performance on this task with other eye movement tasks (e.g., smooth pursuit; Ross et aI., 1993), it may eventually be possible to better describe the deficits in a variety of child and adolescent psychopathological disorders. This may eventually lead to better diagnostic strategies and more specific and effective therapeutic interventions. Despite the limitations of the current study, the results presented support further investigation into the use of oculomotor delayed response tasks in children with presumed frontal lobe-mediated psychopathology.

874

REFERENCES Barkley RA, Grodzinsky G, DuPaul GJ (1992), Frontal lobe functions in attention deficit disorder with and without hyperactivity: a review and research report. J Abnorm Child PsychoI20:163-188 Barone P, Joseph JP (1989), Role of the dorsolateral prefrontal cortex in organizing visually guided behavior. Brain Behav EvoI33:132-135 Becker W (1989), Metrics. In: TheNeurobiology ofSaccadic EyeMovements, Wurtz RH, Goldberg ME, eds. Amsterdam: ElsevierScience Publishers BV, pp 13-68 Benton A (1991), Prefrontal injury and behavior in children. Developmental Neuropsychology 7:275-28 I Conners C (1973), Rating scales for use in drug studies with children. Psycbopbarmacol Bull (Special Issue: Pharmacotherapy of Children) 9:24-84 Funahashi S, Bruce CJ, Goldrnan-Raki PS (1989), Mnemonic coding of visual space in the monkey's dorsolateral prefrontal cortex. J Neuropbys-

ioI61:331-349 Funahashi S, Bruce CJ, Goldrnan-Rakic PS (1993), Dorsolateral prefrontal lesions and oculomotor delayed-response performance: evidence for mnemonic "scotomas." J Neurosci 13:1479-1497 Fuster JM (1989), The Prefrontal Cortex, New York: Raven Press Grattan LM, Eslinger PJ (1991), Frontal lobe damage in children and adults: a comparative review. Developmental Neuropsychology 7:283-326 Grodzinsky G, Diamond R (1992), Frontal lobe functioning in boys with attention-deficit hyperactivity disorder. Developmental Neuropsychology 8:427-445 Mateer CA, Williams 0 (1991), Effects offrontallobe injury in childhood.

Developmental Neuropsychology 7:359-376 Park S, Holzman PS (1992), Schizophrenics show spatial working memory deficits. Arch Gen Psychiatry 49:975-982 Pierror-Deseilligny C, Rivaud S, Gaymard B, Agid Y (1991a), Cortical control of reflexive visually-guided saccades. Brain 114:1473-1485 Pierrot-Deseilligny C, Rivaud S, Gaymard B, Agid Y (1991b), Cortical control of memory-guided saccades in man. Exp BrainRes 83:607-617 Rogers SJ, Pennington BB (1991), A theoretical approach to the deficits in infantile autism. Development and Psychopathology 3:137-162 Ross RG, Radant AD, Hommer OW (1993), A developmental study of smooth pursuit eye movements in normal children from 7 to 15 years of age. JAm Acad ChildAdo/esc Psychiatry 32:783-791 Ross RG, Radant AD, Hommer OW (in press, a), Saccadic eye movements in normal children from 8 to 15 years of age: a developmental study of visuo-spatial attention. J Autism Dev Disord Ross RG, Radant AD, Hommer OW (in press, b), Open- and closed-loop smooth pursuit eye movements in normal children: analyses of a stepramp task. Developmental Neuropsychology Varley CK, Trupin EW (1983), A model for clinical application of double blind assessment of stimulant medication in children with attention deficit 'disorder. Am J Orthopsychiatry 53:542-547 Weinberger DR (1987), Implications of normal brain development for the pathogenesis of schizophrenia. Arch Gen Psychiatry 44:660-669 Welsh MC, Pennington BF (1988), Assessing frontal lobe functioning in children: views from developmental psychology. Developmental Neuropsychology 4:199-230 Weschler 0 (1991), Weschler Intelligence Scale for Children-3rd Edition. San Antonio, TX: Harcourt Brace Jovanovich, Inc Zametkin AJ, Liebenauer LL, Fitzgerald GA er al. (1993), Brain metabolism in teenagers with attention-deficit hyperactiviry disorder. Arch Gen

Psychiatry 50:333-340

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 33:6, JULY/AUGUST 1994