Attention deficit hyperactivity disorder and neurocognitive correlates after childhood stroke

Journal of the International Neuropsychological Society (2003), 9, 815–829.
Copyright 2003 INS. Published by Cambridge University Press. Printed in the USA.
DOI: 10.10170S1355617703960012
Attention deficit hyperactivity disorder andneurocognitive correlates after childhood stroke JEFFREY E. MAX,1 KATHERINE MATHEWS,2 FACUNDO F. MANES,3BRIGITTE A.M. ROBERTSON,4 PETER T. FOX,5 JACK L. LANCASTER,5AMY E. LANSING,6 AMY SCHATZ,6 and NICOLE COLLINGS6 1University of California, San Diego and Children’s Hospital and Health Center, San Diego, California2University of Iowa, Department of Pediatrics and Neurology, Iowa City, Iowa3Raul Carrea Institute for Neurological Research—FLENI, Buenos Aires, Argentina4GlaxoSmithKline, Research Triangle Park, North Carolina5Research Imaging Center, University of Texas Health Science Center, San Antonio, Texas6Children’s Hospital and Health Center, Department of Psychiatry, San Diego, California (Received March 28, 2002; Revised August 22, 2002; Accepted October 3, 2002) Abstract
We investigated the frequency and neurocognitive correlates of attention deficit hyperactivity disorder and traits ofthis disorder (ADHD0Traits) after childhood stroke and orthopedic diagnosis in medical controls. Twenty-ninechildren with focal stroke lesions and individually matched children with clubfoot or scoliosis were studied withstandardized psychiatric, intellectual, academic, adaptive, executive, and motivation function assessments. LifetimeADHD0Traits were significantly more common in stroke participants with no prestroke ADHD than in orthopediccontrols (16028 vs. 7029; Fisher’s Exact p , .02). Lifetime ADHD0Traits in the orthopedic controls occurredexclusively in males with clubfoot (7013; 54%). Participants with current ADHD0Traits functioned significantlyworse ( p , .005) than participants without current ADHD0Traits on all outcome measures. Within the stroke group,current ADHD0Traits was associated with significantly lower verbal IQ and arithmetic achievement ( p , .04),more nonperseverative errors ( p , .005), and lower motivation ( p , .004). A principal components analysis ofselected outcome variables significantly associated with current ADHD0Traits revealed “impaired neurocognition”and “inattention-apathy” factors. The latter factor was a more consistent predictor of current ADHD0Traits inregression analyses. These findings suggest that inattention and apathy are core features of ADHD0Traits afterchildhood stroke. This association may provide clues towards the understanding of mechanisms underlying thesyndrome. (JINS, 2003, 9, 815–829.) Keywords: ADHD, Childhood stroke, Cognition, Apathy, Clubfoot, Scoliosis
INTRODUCTION
1955), childhood hemiplegia (Goodman & Graham, 1996),and encephalitis (Ebaugh, 1923). Such diverse etiologies This is the first study of neurocognitive correlates of atten- would tend to suggest that the syndrome is a final common tion deficit hyperactivity disorder (ADHD) after childhood pathway of varied pathophysiological processes. The syn- stroke. ADHD is one of the most common psychiatric syn- drome may be associated with a varied pattern of symptom dromes which manifest after a variety of brain injuries in- clusters and neurocognitive correlates depending on the na- cluding traumatic brain injury (TBI) (Bloom et al., 2001; ture and extent of the brain injury. Pathophysiological and Gerring et al., 1998; Max et al., 1998a; Max et al., in press), neurocognitive research in idiopathic ADHD is more ad- very low birth weight0premature infants (Botting et al., vanced than corresponding research in ADHD that follows 1997; Lou, 1996; Whitaker et al., 1997), cerebral palsy (Bre- brain injury, yet it is far from conclusive. Therefore the slau & Chilcoat, 2000; Ingram, 1956), epilepsy (Ounsted, investigation of poststroke ADHD and idiopathic ADHDare likely to have mutual relevance.
Children with focal stroke lesions provide a potentially Reprint requests to: Jeffrey E. Max, M.B.B.Ch, Children’s Hospital useful model for the investigation of ADHD after brain and Health Center, San Diego, 3020 Children’s Way, MC 5033, SanDiego, CA 92123. E-mail: jmax@ucsd.edu injury. We recently reported a trend towards an association of ADHD and stroke lesions of the putamen in participants construct of executive function is nonunitary. We report on with lesions Յ10 cc (Max et al., 2002). In the current study, two measures of executive function, the Wisconsin Card we shall report neurocognitive correlates of poststroke Sorting Test (WCST) (Grant & Berg, 1948) and the Con- ADHD regardless of lesion volume. Our focus will be on trolled Oral Word Association Test (Borkowski et al., 1967).
intellectual, academic, adaptive, and executive functions, The WCST is a test that demands flexibility of sorting strategies for multidimensional figures, and requires work-ing memory and behavioral inhibition (Barkley, 1997). Chil-dren with ADHD make more perseverative errors and Intellectual Function
nonperseverative errors on the WCST than controls, but Children with idiopathic ADHD have a lower IQ, particu- this is not a uniform finding (Barkley, 1997; Klorman et al., larly verbal IQ, than controls. This difference is small but 1999; Pennington & Ozonoff, 1996). Performance on the reaches statistical significance (Barkley, 1997). In con- WCST has not been reported in children with brain-injury- trolled studies, this can affect comparisons of neurocogni- tive functions that correlate significantly with IQ. Barkley The Controlled Oral Word Association Test (Borkowski (1997) has challenged the wisdom of statistically control- et al., 1967) is a test of verbal fluency. The literature on ling for IQ in studies of idiopathic ADHD because findings children with ADHD is mixed with respect to documented related to the independent variable of interest, ADHD, might differences in this domain of executive function (Barkley, be eliminated. Intellectual function in brain-injury-related 1997). However, tests which use letters, for example, the ADHD has been studied in children with TBI. Brain-injury- Controlled Oral Word Association Test, rather than seman- related ADHD is associated with significantly lower IQ when tic categories may be more challenging and more likely to analyses include mild to severe TBI children (Max et al., in discriminate children with ADHD from controls (Barkley, press) and IQ scores were lower but reached statistical sig- 1997). Verbal fluency associated with brain-injury-related nificance in only 1 of 2 studies when analyses were limited to more severely injured TBI participants (Gerring et al.,1998; Max et al., in press).
Motivation
Academic Function
There is extensive evidence for difficulties in the self- Academic deficits have been associated with idiopathic regulation of motivation, particularly persistence of effort, ADHD. Specifically, reading disorders are often comorbid in children with idiopathic ADHD (Barkley, 1997). It may with ADHD (Shaywitz & Shaywitz, 1994; Willcutt & Pen- be that children with ADHD have a deficit in the executive- nington, 2000). ADHD is also associated with poorer long- function capacity to bridge delays in reinforcement and per- mit the persistence of goal-directed acts (Barkley, 1997).
educational achievement measured in years of formal edu- There are no published reports of motivation or its con- cation (Wood & Felton, 1994) and poorer math skills (Nuss- verse, apathy, associated with brain-injury-related ADHD.
baum et al., 1990). In contrast, academic function associated Against this background, we set out to study ADHD in with brain-injury-related ADHD has not been studied.
children with neuroimaging evidence of focal stroke le-sions, controlling for age, gender, socioeconomic status(SES), and the presence of and timing of onset of a chronic Adaptive Function
non-central nervous system (non-CNS) medical condition.
Over 70% of children with idiopathic ADHD in an epide- (See Table 1 for categorization of ADHD with onset after miological study had deficits in adaptive function (Costello the diagnosis of the respective medical condition: stroke et al., 1996). These deficits are often considered markers of and chronic non-CNS disorders.) We hypothesized, first, more severe and pervasive impairments in this population that significantly more children with stroke than controls (Shelton et al., 1998). However, adaptive function (Max would have a lifetime history of the full ADHD syndrome.
et al., in press) or overall disability (Gerring et al., 1998) Second, we hypothesized that when children with a lifetime associated with brain-injury-related ADHD has been stud- history of the full ADHD syndrome and children with only ied only in childhood TBI. Taken together, these studies ADHD traits (defined below) were considered as one group indicate significantly worse functioning in brain-injury- (lifetime ADHD0Traits), significantly more children with related ADHD participants regardless of whether the sam- stroke than medical controls would exhibit lifetime ADHD0 ples included a full range of injury severity or only more Traits. Third, we hypothesized that children with current ADHD0Traits (i.e., excluding resolved ADHD cases) woulddemonstrate significantly more impairments in intellectual,academic, adaptive, and executive functions as well as mo- Executive Function
tivation than participants without current ADHD0Traits Executive-function deficits have been implicated in the either when stroke and control participants were analyzed expression of idiopathic ADHD (e.g., Barkley, 1997). The or only when participants with stroke were considered.
Table 1. Postmedical diagnosis ADHD0Traits
since stroke; and (4) English as first language. The follow- ing exclusions were applied: (1) neonatal bleeds (e.g., in-traventricular hemorrhages, germinal matrix hemorrhages) potentially associated with prematurity; (2) neonatal water- shed infarcts associated with hypoxia; (3) hemoglobinopa- thies; (4) progressive neurometabolic disorders; (5) Down’s syndrome and other chromosomal abnormalities; (6) malig- nancy; (7) congenital hydrocephalus; (8) shunts; (9) con- genital and acquired CNS infections; (10) clotting factor deficiency; (11) stroke in a pregnant minor; (12) previous organ or bone marrow transplant; (13) cerebral cysts; (14) trauma; (15) transient ischemic attack; (16) Moya Moya; (17) severe and profound mental retardation; (18) quadri- plegia, triplegia, or diplegia diagnoses; (19) syndromic vas- Legend. One additional stroke participant had ADHD before the medical cular malformations (excluding arterio-venous malformation diagnosis. Hypothesis 1 concerned lifetime postmedical diagnosis ADHD (AVM) aneurysm ruptures); (20) systemic lupus erythema- (n 5 13 in stroke cohort; n 5 5 in orthopedic cohort) which consists of tosis; and (21) multiple lesions (unless in close proximity).
participants with a history of the full postmedical diagnosis ADHD syn-drome at some point in their life (postmedical diagnosis ADHD current, Inclusion criteria for controls were as follows: Children postmedical diagnosis ADHD resolved, postmedical diagnosis ADHD par- with congenital clubfoot and children with scoliosis were tial resolution); Hypothesis 2 concerned lifetime postmedical diagnosis individually matched to participants with stroke according ADHD0 Traits (n 5 16 in stroke cohort; n 5 7 in orthopedic cohort) whichconsists of participants with a history of the full postmedical diagnosis to age of onset of stroke (i.e., early vs. late). Matching was ADHD syndrome or postmedical diagnosis ADHD traits at some point in based on gender, ethnicity, SES, and age within 1 year. Age their life (postmedical diagnosis ADHD current, postmedical diagnosis matching had to be extended to 16 months in three cases.
ADHD resolved, postmedical diagnosis ADHD partial resolution, post-medical diagnosis ADHD traits); Hypothesis 3 concerns current postmed- The following exclusion criteria were applied for controls: ical diagnosis ADHD0 Traits (n 5 15 in stroke cohort; n 5 6 in orthopedic Orthopedic controls were excluded when they had evi- cohort) which consists of participants with current postmedical diagnosis dence in the chart of acquired or congenital CNS injury that ADHD, postmedical diagnosis ADHD traits, and postmedical diagnosisADHD in partial resolution.
may be part of broader (e.g., neuromuscular) syndromesunrelated to the common idiopathic syndromes. Matchingwas possible for all but two children with late stroke le-sions. These two late-onset stroke participants were matched The design and concept of this study was strongly influ- Stroke participants evaluated included 17 with early le- enced by British research on psychiatric aspects of neuro- sions and 12 with late lesions. The strokes were ischemic in logical disorders over the past three decades (e.g., Goodman 21 cases and hemorrhagic in eight cases. Etiology included & Graham, 1996; Rutter et al., 1970; Seidel et al., 1975).
15 idiopathic occlusive cases, two idiopathic hemorrhagic The research design, previously reported in detail (Max cases, four cases occurred in participants with congenital et al., 2002), is a cross-sectional study of children with a heart disease (three after cardiac surgery or catheterization history of a single stroke and a medical control group. The and one after varicella zoster infection), five cases of arterio- study focus was on psychiatric outcome in children with venous malformation rupture, one case of ruptured angi- strokes in addition to neuropsychological, academic, adap- oma, one case possibly linked to comorbid ulcerative colitis, tive, executive, and family function outcomes. In accor- and one case followed a varicella infection. Both cases as- dance with previous studies (Riva & Cazzaniga, 1986; sociated with varicella infections were presumed to be due Woods, 1980), stroke participants were considered to have to vasculitis and there was no evidence for encephalitis “early” lesions if their brain lesion occurred prenatally or (Roach & Riela, 1995). The distribution of the brain lesions up to 12 months of postnatal life. The “late” lesion group included seven cases of predominantly putamen lesions, consisted of children who acquired their stroke at age 12 nine cases of large middle cerebral artery (MCA) distribu- months or later. We matched “early” stroke participants with tion infarcts including deep gray structures, ten cases of children with clubfoot, with the rationale that physical de- smaller MCA distribution fronto-temporal or temporo- formity in both groups was an early, and frequently congen- parietal lesions sparing the deep gray, and three cases of ital, insult. We matched “late” stroke participants with parietal or parieto-occipital strokes. Forty-eight partici- children who had scoliosis because these children were with- pants (including all stroke participants) were recruited from out physical deformity prior to their acquired disorders.
one university hospital while ten participants were re- Inclusion criteria for stroke cases were as follows: cruited from a second university hospital due to the reloca- (1) Neuroimaging documentation of a focal, nonrecurrent and nonprogressive supratentorial brain parenchymal le- The stroke and orthopedic groups were not significantly sion caused by a stroke before age 14 years; (2) Participants different on matching variables of age and SES. Respective aged 5–19 years at the time of the assessment; (3) Ն1 year age means (SD) of stroke and orthopedic participants were 12.1 (3.9) and 11.9 (3.9), df 5 56, t 5 2.135, p . .8.
tal behavior-problem score, the CBCL provides two “broad Respective SES means (SD) of stroke and orthopedic par- band” subscales (internalizing and externalizing symp- ticipants were 2.45 (.95) and 2.45 (1.06), df 5 56, t 5 0, toms) and eight “narrow band” scales (withdrawn, somatic p 5 1.0. There were 18 males, 27 Caucasians, and two complaints, anxious0depressed, social problems, thought biracial children in each of the stroke and orthopedic groups.
problems, attention problems, delinquent behavior and ag-gressive behavior).
In addition, the Children’s Motivation Scale (Gerring et al., Measures
1996) was administered. This is a 16-item rating scale that Psychiatric and behavioral measures measures the behavioral, cognitive, and emotional concom-itants of motivation, the conceptual converse of apathy. The Diagnostic and Statistical Manual–Fourth Edition (DSM– scale has fair-to-good test–retest reliability, internal con- IV) psychiatric diagnoses (American Psychiatric Associa- sistency, and interrater reliability. The measure correlated tion, 1994) were derived by utilizing a semistructured significantly with an independent measure of withdrawal interview, the Schedule for affective disorders and schizo- but not with depression. Normative samples have a mean phrenia for school-aged children, present and lifetime ver- of 50 with a standard deviation of approximately 10. Psy- sion (K-SADS-PL) (Kaufman et al., 1997). The K-SADS-PL chiatric inpatient or outpatient samples have a mean of 31 is an integrated parent–child interview which generates di- with a standard deviation of 10. Parents completed this agnoses based on a clinician synthesizing data collected from parent and child separately, querying present and life-time symptoms as well as providing data regarding the tim- Family psychiatric and behavioral assessments ing of symptom onset in relation to the stroke and orthopedicdiagnosis. If participants have significant symptoms on ques- The Family History Research Diagnostic Criteria (Andre- tions for a particular syndrome in a K-SADS-PL screen asen et al., 1977) interview was conducted in most cases by interview, a corresponding K-SADS-PL supplementary in- a trained research assistant and in other cases by JEM. Cri- terview module is completed to clarify the diagnosis.
teria were modified to conform with Diagnostic and Statis- The outcome measures were the diagnoses of DSM–IV tical Manual–Third Edition–Revised (DSM–III–R) criteria.
ADHD and “ADHD traits.” This approach recognized the At least one parent acted as the informant. Family ratings dimensional nature of ADHD symptomatology (Levy et al., were summarized for first-degree relatives on a four-point 1997). The diagnosis of ADHD was made when the symp- scale (0 to 3) (Max et al., 1998a) of increasing severity.
tom complex resulted in clinically significant impairment, Among specific disorders recorded in both first- and second- even after considering overall developmental level of the degree relatives was ADHD. We also recorded another vari- child, and was not based simply on symptom counts. The able, a merged category of ADHD or ADHD symptoms in ADHD subtypes (combined, predominantly inattentive, pre- first- and second-degree relatives. This category was pur- dominantly hyperactive0impulsive, and not otherwise spec- posefully broader than the first because it is difficult for ified) were applied only to participants with a clinically family members to endorse sufficient ADHD symptoms to significant DSM–IV ADHD syndrome. The designation of meet diagnostic criteria for their relatives who may in fact ADHD traits was given to participants with a subsyndro- mal condition. ADHD traits were defined a priori as at least Global family functioning was assessed by using the Mc- three of four symptoms in the screening interview for ADHD Master Structured Interview of Family Functioning. The rated positive but “subthreshold” or at least one screener interview is used to derive scores on the Clinical Rating question rated “threshold” and at least five additional symp- Scale (CRS) (Miller et al., 1994). The CRS is comprised of toms on the supplementary ADHD interview rated “sub- seven domains, including global family functioning, which threshold” or “threshold”. The age-of-onset (7 years) are rated 1–7 on a Likert Scale. Scores of 5–7 indicate criterion for ADHD was waived so that we could document healthy family function and lower scores indicate un- the development of this behavioral syndrome in partici- healthy family functioning. Two trained research assistants, pants whose stroke or scoliosis was diagnosed later.
who remained blind to the psychiatric findings, conducted Fifty-seven of 58 interviews were administered by JEM, who is a board-certified child and adolescent psychiatrist,and all were videotaped. AEL, a trained Ph.D. level re- searcher, administered one interview. Eleven interviews wereselected randomly to be rated by a second child psychia- SES assessment was accomplished through the Four Fac- trist, BAMR, to ascertain interrater reliability. Agreement tor Index (Hollingshead, 1975). Classification into five lev- regarding the presence of ADHD, ADHD subtype, ADHD els (1 to 5) (Max et al., 2002) depends on scores derived from a formula involving both mother’s and father’s edu- The Child Behavior Checklist (CBCL) (Achenbach, 1991) cational levels and occupational levels. Lower scores re- was completed by a parent. The CBCL is a well-standardized flect higher SES. Controls were matched within two levels assessment of child-behavior problems. In addition to a to- 5 mm skip 1 mm). Twenty-six of 29 stroke participantsunderwent research scans that were the basis of their lesion The Wechsler Intelligence Scale for Children, Third Edition location analyses. The other three participants who could (WISC–III) (Wechsler, 1991) was used. Prorated Full Scale not have a research MRI (due to refusal, concern about IQ (FSIQ) was derived from a prorated Performance IQ intracerebral metallic clips, and equipment failure, respec- (PIQ: Picture Arrangement, Block Design, and Coding sub- tively) had lesion location determined from previous clini- tests) and a prorated Verbal IQ (VIQ: Information and Sim- cal computed tomography (CT) scans (2) or MRI scan (1).
ilarities subtests). We applied the upper age limit norms for A neurologist, FFM, marked the lesions on hard-copy this test to the few participants who were above the age films. Guided by these lesion markings, an experienced neuroanatomist supervised by PTF and JLL “painted” eachlesion using a three-dimensional (3-D) brain-morphometrics package (Display, Montreal Neurological Institute). Lesion The Wide Range Achievement Test, Revised (WRAT–R, volume was computed in absolute units (cm3 ) before and Jastak & Wilkinson, 1984) was administered to assess after normalization for intersubject differences in brain size achievement in reading, spelling, and mathematics. The (Lancaster et al., 1995). Size normalization was performed WRAT–R consists of two alternate forms with two levels using the spatial normalization (SN) software which has (level 1 for children ages 5.0 to 11.11; level 2 for persons the user mark the front, back, left, right, top, and bottom of the brain following anterior commissure–posterior commis-sure (AC–PC) alignment. SN then sized the brain along each axis to the template size, thus correcting for brain size.
Adaptive functioning assessment was completed by trained research assistants using the Vineland Adaptive BehaviorScale interview (Sparrow et al., 1984) through a nondirec- A standard history and examination was administered by tive interview with the primary caretaker. The Vineland scales KM or JEM. Scores on a neurological severity index were survey activities that the child habitually demonstrates in rated (Max et al., 2002). Scale items consisted of ratings of the environment, yielding an overall composite score and head circumference, degree of hemiparesis, function of the separate standard scores for Socialization, Daily Living, “good” side, and history of seizures. Higher scores re- Microcephaly (, 3rd percentile) was present in 5025 (20%) cases where head circumference was measured. Therewas no hemiparesis in 12029 (41%) cases, mild hemipare- A computerized version of the Wisconsin Card Sorting Test sis in 4029 (14%) cases, typical hemiparesis in 11029 (38%) (WCST) (Grant & Berg, 1948) was administered to partici- cases, and worst hemiparesis in 2029 (7%) cases. The side pants to assess aspects of executive function. Participants of the body ipsilateral to the brain lesion was normal in were asked to match each stimulus card (appearing at the 27029 (93%) cases, had slightly decreased coordination in bottom of the screen) with one of the four key cards at the 1029 (3%) cases, and was poorly coordinated in 1029 (3%) top. Correct responses were signified by a “high beep” and cases. The possibility of bilateral physical signs with uni- a “dull buzz” denoted errors. Participants’ responses were lateral lesions is a well-known phenomenon (Goodman & typed into a keyboard and there were no time limits. “Cor- Yude, 1997). Eleven of 29 participants (38%) had a history rect” sorting strategies were changed without announce- of seizures, but only five were receiving anticonvulsant med- ment or explanation after a participant had completed ten ication (carbamezepine monotherapy in 3; carbamezepine correct sorts under a specific principle (e.g., color). A max- plus mysoline in 1; phenytoin in 1) at the time of assess- imum of six categories across 128 cards was possible.
Verbal fluency was tested by means of the Controlled Oral Word Association Test (Borkowski et al., 1967). The task required participants to generate as many different words Group differences were tested with independent sample t beginning with a particular letter, “F ”, “A”, then “S” within tests and x2(or Fisher’s Exact Test) analyses when the vari- discrete 60-s periods for each respective letter. Proper names ables of interest were continuous or categorical, respectively.
For hypotheses 1 and 2, we compared the rates of life- time ADHD and lifetime ADHD orthopedic participants with Fisher’s Exact Test. For hy- Protocol magnetic resonance imaging (MRI) scans were pothesis 3, we correlated current intellectual, academic, adap- obtained ( T1-weighted volumetric mode, SPGR 0400, tive, and executive functioning, and motivation in TR 5 26, TE 5 7, Matrix 256 3 192, NEX 5 2, 1.5-mm participants currently affected with ADHD0Traits versus thickness with no skip; T2-weighted multiecho, FSE0V, participants not currently affected with ADHD0Traits (re- TR 5 2350, TE 5 170102, Matrix 256 3 192, NEX 5 1, solved ADHD0Traits and never any ADHD0Traits). We rea- soned that resolved ADHD0 Traits status would be depressive disorder not otherwise specified, 2; simple pho- accompanied by improvements in function in these various bia, 2; overanxious disorder, 1; chronic motor tic disorder, 1; chronic vocal tic disorder, 1; and stereotypic movement Other analyses of general interest in the description and characterization of the sample but not a part of our three Three orthopedic controls had ADHD (two inattentive hypotheses were conducted. Analyses on family history of and one combined type). Two orthopedic participants had psychiatric disorder and family history of ADHD con- ADHD traits, and one had a completely resolved ADHD cerned participants with any lifetime history of ADHD0 (not otherwise specified subtype). One other orthopedic par- Traits versus those with no lifetime history of ADHD0 ticipant had a partially resolved ADHD (not otherwise spec- Traits. We assumed that a family history of these conditions ified subtype) with only two inattentive symptoms current might predispose participants to persisting and more tran- at the time of the assessment. The latter participant was sient ADHD0Traits. Similarly, we compared stroke chil- assigned to the ADHD0Traits grouping. Table 2 shows de- dren with a lifetime history of ADHD0Traits versus those mographic, lesion, and ADHD characteristics of each with no lifetime history of ADHD0Traits with respect to onset of stroke (early vs. late), lesion laterality, and lesion These data confirmed our first hypothesis: 13028 (46%) volume. This allowed us to document both persistent and of stroke participants had a lifetime history of the full ADHD more transient correlates of these lesion characteristics.
syndrome compared with 5029 (17%) of orthopedic con- Certain independent variables that were found to be sig- trols (Fisher’s Exact Test, p , .03). Our second hypothesis nificantly correlated ( p , .05) with current ADHD0Traits was likewise confirmed: 16028 (57%) eligible stroke par- from each of the domains of interest (intellectual function, ticipants compared with 7029 (24%) orthopedic partici- academic function, adaptive function, executive function, pants had lifetime ADHD0Traits (ADHD, ADHD traits, and motivation) were chosen for a principal components partially remitted ADHD, or resolved ADHD) (Fisher’s Ex- analysis with a varimax rotation. This was done to extract act Test, p , .02). The surprisingly high rate of ADHD0 the largest amount of meaningful variation among the inde- Traits among the orthopedic participants was accounted for pendent variables of which several were conceptually re- exclusively by males with clubfoot (7013; 54%).
lated. Factors with an eigenvalue .1 were identified and Table 3 shows CBCL scores of participants with and with- named according to their dominant characteristics.
out ADHD0Traits. The ADHD0Traits group was signifi-cantly more impaired regarding attention problems andshowed a statistical trend regarding increased total behav- ior scores in analyses of the entire cohort and of strokeparticipants only. The ADHD Incidence of ADHD
tistical trend for increased aggressive behavior in analyses Table 1 shows the distribution of ADHD in the groups. One of the entire cohort and for delinquent behavior in analyses stroke participant was diagnosed with prestroke ADHD and was dropped from further analyses regarding the develop-ment of postmedical disorder ADHD0Traits. There were no Outcome in stroke and control participants other participants who had premorbid symptoms reachingour defined level for traits of ADHD. This yielded 28 stroke To provide context for the analyses concerning ADHD0 Traits, Table 4 shows comparisons of intellectual, aca- participants were diagnosed with ongoing ADHD (six in- demic, adaptive, and executive function, as well as attentive, four not otherwise specified, one hyperactive motivation between stroke and orthopedic participants. Chil- impulsive, and one combined subtype); and three had dren with stroke scored significantly worse than orthopedic ongoing ADHD traits. One additional stroke participant had controls in all domains except motivation.
a resolved ADHD, not otherwise specified diagnosis. Par-ticipants with the “not otherwise specified” subtype had Characteristics of children with ADHD0 Traits predominantly inattentive symptoms. Notably, all but two The child with partially resolved ADHD was considered in participants with ADHD0Traits or resolved ADHD0Traits the ADHD0Traits group for all analyses. However, there had at least one hyperactivity0impulsivity symptom at thresh- were no meaningful changes in the analyses when this child old or subthreshold intensity on the K-SADS-PL. Only four was excluded from the ADHD0Traits group.
of the 16 participants with a lifetime history of ADHD0Traits developed this problem after age 7 because their Nine of 16 stroke children with a lifetime history of Children with ADHD0Traits were significantly more im- ADHD0Traits had other DSM–IV psychiatric disorders paired than children without ADHD0Traits regarding intel- present at the time of assessment: oppositional defiant dis- lectual, academic, and adaptive functioning, WCST order, 4; personality change disorder, 3; separation anxiety measures, motivation, and family psychiatric history disorder, 2; agoraphobia without panic, 2; social phobia, 2; (Tables 5–7). The ADHD0Traits group was not signifi- Table 2. Demographic, lesion, and ADHD characteristics of participants
Legend. Participants are listed according to ADHD status of stroke participants in the hierarchical order of Table 1 and then according to age at evaluation. Eachstroke participant is followed by the individually matched orthopedic control designated with a “c” at the end of the subject ID. * No CBCL score is available forthis participant. Ant-lat 5 anterior lateral; AVM 5 arterio-venous malformation; CBCL 5 Child behavior checklist; F 5 female; Fr-T0Temp-P 5 Fronto-temporal0temporo-parietal lesions sparing the deep gray structures; H0I 5 hyperactive0impulsive; Hem. 5 hemorrhage; L 5 left; M 5 male; MCA 5 large middle cerebralartery distribution infarcts including deep gray structures; mos 5 months; N0A 5 not applicable; NOS 5 not otherwise specified; Occ 5 occlusive; P0P-Occip. 5Parietal0Parieto-Occipital; part resol 5 partial resolution; Pre 5 prenatal; R 5 right; yrs 5 years.
Table 3. Child behavior checklist scores and ADHD
cantly different regarding verbal fluency, global family func- tion, family history of ADHD, family history of ADHD0ADHD symptoms, age, and gender (16036 males vs. 5021 Independent variables found to be significantly correlated ( p , .05) with ADHD0Traits from each of the domains ofinterest (intellectual function, academic function, adaptive function, executive function, and motivation) were chosenfor a principal components analysis for the following rea- Analyses limited to stroke participants that compared sons: VIQ was selected because it provides a more accurate stroke participants with ADHD0Traits to stroke partici- reflection of overall intellectual function in stroke partici- pants without ADHD0 Traits revealed the following pants than PIQ (and therefore FSIQ) due to motor impair- (Tables 5–7): Children with ADHD0Traits were signifi- ments such as hemiplegia (Goodman & Yude, 1996). Reading cantly more impaired regarding VIQ, arithmetic scores, and standard score was selected from the significant academic certain executive function measures (WCST total errors and function tests because of the well-known comorbidity of nonperseverative errors). The groups were not significantly reading disability and ADHD (Shaywitz & Shaywitz, 1994; different regarding age, gender (10018 males vs. 5010 fe- Willcutt & Pennington, 2000). The Vineland adaptive be- males), neurological severity summary score, seizure activ- havior composite was chosen because it captures overall ity history (6015 participants with ADHD0Traits had a adaptive function. Both perseverative errors (standard score) seizure history compared with 4013 participants without and nonperseverative errors (standard score) were selected ADHD0Traits), FSIQ, PIQ, reading, spelling, adaptive func- from the WCST because these measures reflect distinct do- tion, family function, family history of ADHD, and family mains of information processing. Finally, the total score on history of ADHD0ADHD symptoms. Neither were the the Children’s Motivation Scale was chosen.
groups significantly different in terms of perseverative as- These six variables were then entered into a principal pects of executive function (responses and errors) on the components analysis with a varimax rotation. The two- factor final solution is shown in Table 8. We termed the first Table 4. Intellectual, academic, adaptive, and executive function, and motivation in stroke and control participants
Executive function (Wisconsin Card Sorting Test) Legend. Means (SD) of standard scores (SS).
aWechsler Intelligence Scales for Children–Third Edition.
bWide Range Achievement Test–Revised.
cVineland Adaptive Behavior Scales.
dChildren’s Motivation Scale.
COWA 5 Controlled Oral Word Association; PIQ 5 Performance IQ; VIQ 5 Verbal IQ; FSIQ 5 Full Scale IQ; ns 5 not significant.
factor the “impaired neurocognition” because it was corre- df 5 1, p , .002) and “impaired neurocognition” factor lated highly with general intellectual function, specific read- (Wald x2 5 4.12, df 5 1, p , .05). The second regression ing ability, overall adaptive function, and perseveration which based on only stroke participants was also significant is typically considered to reflect a measure of neurological (22 log likelihood x2 5 19.52, df 5 2, p , .003), and integrity. We termed the second factor the “inattention– correctly predicted 82.6% of the ADHD0Traits cases. How- apathy” because it was correlated highly with nonpersever- ever, only the “inattention–apathy” factor (Wald x2 5 5.71, ative errors (which reflects inattention) and low motivation 0 df 5 1, p , .02) significantly contributed to the presence apathy. The derived “impaired neurocognition” and the of ADHD0Traits when simultaneously entered in the re- “inattention–apathy” factors, respectively, captured 56.5% gression with the “impaired neurocognition” factor (Wald and 18.3% (total 74.7% with rounding) of the variance within x2 5 .76, df 5 1, p . .38).
the set of independent variables entered. The results weresimilar when the analyses were repeated utilizing only the Lesion correlates of lifetime ADHD0 Traits stroke participants: the derived “impaired neurocognition”and the “inattention–apathy” factors, respectively, captured Lifetime ADHD0Traits was not significantly related to 51.6% and 25.5% (total 77.2% with rounding) of the vari- whether the lesion onset was early or late (9017 children ance within the set of independent variables entered.
with early lesions vs. 7011 children with late lesions had Logistic regression analyses were conducted for the pres- this behavior disturbance). Lifetime ADHD0Traits was not ence of ADHD0Traits using both the “impaired neurocog- significantly related to lesion laterality (7015 children nition” and the “inattention–apathy” factors. The first with right-sided lesions vs. 9013 children with left-sided regression which included stroke and orthopedic participants lesions had this behavior disturbance). Finally, lesion vol- was significant (22 log likelihood x2 5 37.79, df 5 2, ume, which was highly skewed, was not significantly re- p , .00005), and correctly predicted 77.8% of the ADHD0 lated to lifetime ADHD0Traits (Mann–Whitney U Test 5 Traits cases. Furthermore, each factor independent of the 70.0): the mean rank of lesion volume for the participants other significantly contributed to the presence of ADHD0 with and without lifetime ADHD0Traits was 12.7 (n 5 15) Traits: “inattention–apathy” factor (Wald x2 5 10.23, and 13.5 (n 5 10), respectively.
Table 5. Characteristics of ADHD
Legend. Means (SD) of standard scores.
aWechsler Intelligence Scales for Children–Third Edition.
bWide Range Achievement Test–Revised.
cVineland Adaptive Behavior Scales.
dChildren’s Motivation Scale.
PIQ 5 Performance IQ; VIQ 5 Verbal IQ; FSIQ 5 Full Scale IQ; ns 5 not significant.
DISCUSSION
in partial resolution, and resolved ADHD were combined(lifetime ADHD0Traits), the rate in the stroke group (57%; The main finding from this study was that ADHD develop- 16028) was significantly higher than the rate in the ortho- ment after stroke in children occurred at a rate (46%; 13028) pedic control group (24%; 7029). These increased rates could which was significantly higher than ADHD occurring after not be explained by differences in age, gender, SES, race, an orthopedic diagnosis in controls (17%; 5029). We found family function, family history of ADHD, or the presence also that when children with ADHD, ADHD traits, ADHD of a chronic medical condition requiring medical attention.
Table 6. Tests of executive function and ADHD
Controlled Oral Word Association Test (percentile) Controlled Oral Word Association Test (percentile) Note. SS 5 standard score. ns 5 not significant.
The differences were therefore probably related to the brain tively diagnosed with clubfoot, this may have implications in the search for genetic markers for both conditions.
The differences in incidence of lifetime ADHD0Traits Differences on neurocognitive measures between chil- between the two groups may have actually been greater dren (stroke plus orthopedic controls) with current ADHD0 were it not for an unexpectedly high rate of lifetime ADHD0 Traits and those with no current ADHD0Traits reflected the Traits in males with clubfoot (7013; 57%) which accounted dominant influence of the stroke condition on function. This entirely for the occurrence of lifetime ADHD0Traits in or- was true for intellectual, academic, adaptive, and executive thopedic controls. A specific association between ADHD function. However, this was not the case for motivation and clubfoot has not been noted before; however, there is which was not significantly different between stroke and some evidence that minor physical anomalies are overrep- orthopedic groups but was significantly lower in children resented in children with attention deficit disorder and their (stroke plus orthopedic controls) with ADHD. Therefore, first-degree relatives (Deutsch et al., 1990). If the associa- this is a clue that the neural substrate of low motivation or tion between ADHD and males with clubfoot is replicated apathy may play a central role in the pathophysiology of in a larger orthopedic clinic sample of children consecu- Table 7. ADHD and family psychiatric history
Table 8. Rotated factor matrix of independent variables of interest
aWechsler Intelligence Scales for Children–Third Edition.
bWide Range Achievement Test–Revised.
cVineland Adaptive Behavior Scales.
dWisconsin Card Sort Test.
eChildren’s Motivation Scale.
The differences in neurocognitive measures between ticipants and between those with and without ADHD0 stroke participants with or without current ADHD0Traits Traits suggests that additional measures would have to be were more limited but informative. These differences in- employed to document possible specific executive function cluded significantly lower VIQ, lower arithmetic scores, deficits in children with poststroke ADHD0Traits.
more nonperseverative errors on the WCST, and lower mo- The factor analysis of independent variables which were tivation. Particularly striking was the fact that VIQ of stroke significantly associated with ADHD0Traits yielded “im- children with no current ADHD0Traits was in the average paired neurocognition” and “inattention–apathy” factors.
range compared with below average scores for current This supported earlier research that idiopathic ADHD is ADHD0Traits children. As noted before, VIQ is considered associated with cognitive differences including small but a more accurate measure of overall intelligence than PIQ or significant IQ decreases, and academic function and adap- FSIQ in the stroke population (Goodman & Yude, 1996).
tive function deficits. The “neuro” component of the “im- In contrast to VIQ, academic scores were depressed in paired neurocognition” factor reflects the finding of increased the non-ADHD0Traits stroke participants and the ADHD0 perseverative errors typically associated with brain dam- Traits stroke participants. Only scores for arithmetic were age. Logistic regression demonstrated that the “inattention– significantly different between the ADHD0Traits and non- apathy” factor was a more consistent predictor than the ADHD0Traits groups. The specificity of this finding is un- “impaired neurocognition” factor in accounting for current clear because reading and spelling scores were also lower ADHD0Traits. This suggests that inattention and apathy but not significantly so. It is possible that in stroke partici- are core impairments with respect to ADHD0Traits and that pants, children with ADHD0Traits have a pattern of injury the syndrome is not merely a reflection of more general and0or a pattern of neuronal repair that disproportionately affects working memory, visual memory, and visual-spatial Clinically, it is important to differentiate (1) inattention skills, whose recruitment is central to the completion of related to ADHD or depressive disorder, and (2) apathy limited to “personality change due to stroke, apathetic sub- The increase in nonperseverative errors but not persev- type” (American Psychiatric Association, 1994; Max et al., erative errors on the WCST in stroke participants with cur- 1998b) or apathy which may be part of a depressive disor- rent ADHD0Traits compared with stroke participants with der. Inattention and apathy loaded on the same factor and no current ADHD0Traits suggested that this group had dif- they also tend to respond to similar treatment, for example, ficulty with the marshalling of nonspecific attention to task.
stimulants (Marin et al., 1995). This suggests that they may In fact, stroke participants with no current ADHD0Traits have related neural mechanisms. Not surprisingly, the CBCL had scores in the average range on this and other measures profile exhibited by the children with ADHD0Traits showed recorded from the WCST. This is consistent with the pre- significant increases on the attention problems scale which dominant difference found in a study of idiopathic ADHD includes hyperactivity0impulsivity symptoms. These chil- in which children with ADHD, combined type differed from dren tended to have increased total problems as well as controls only in nonperseverative errors on the WCST (Klor- aggressive and delinquent behaviors. At the level of the man et al., 1999). Furthermore, with respect to verbal flu- individual, we found that just over half the children with ency, the other executive function test assessed, the absence ADHD0Traits had comorbid externalizing and0or internal- of significant differences between stroke versus control par- izing psychiatric disorders. The psychiatric interview and CBCL findings support the position that we have not sim- referral biases. Sixth, the psychiatrist did not have the ben- ply identified a group of children with compromised atten- efit of a teacher’s report in reaching diagnostic decisions.
tional resources related to brain damage and measurable byneuropsychological tests, but rather DSM–IV ADHD with FUTURE DIRECTIONS
comorbid psychiatric problems probably related to braindamage.
ADHD after childhood stroke should be further clarified.
Striking by the absence of an apparent association was This will require a larger sample of stroke participants and the relationship between ADHD and family history of controls. It will be important to confirm the preponderance ADHD. This is a departure from the pattern seen in idio- of ADHD, inattentive and the “not otherwise specified” sub- pathic ADHD. There was however a significant association types and its association with apathy. Other neurocognitive of intensity of family psychiatric history and ADHD, which correlates should be investigated including dimensions of suggests the existence of a relatively less specific and less executive function such as inhibitory control. Direct mea- direct relationship between family psychopathology and sures of attention would clarify the attentional problem more ADHD. Most cases of ADHD were children with stroke. It specifically (Posner & Peterson, 1990). A larger study should is likely that in the apparent absence of a relationship with investigate structural lesion–ADHD correlates and func- familial ADHD, lesion location (Max et al., 2002) and pos- tional imaging studies in ADHD participants may reveal sibly dysfunctional neuronal circuits or connections result- characteristic patterns of abnormal activation on neurocog- ing from imperfect reparative processes (Goodman, 1989) nitive tasks including tests of attention. A comparison group may be implicated in the manifestation of poststroke ADHD.
of children with idiopathic ADHD would help clarify what We must acknowledge a number of limitations in this might be specific in terms of the neurocognitive profile of study. First, the sample is small and findings on larger sam- poststroke ADHD. Finally, a treatment study would provide ples of carefully screened children with stroke and their critically important clinical data about whether children with appropriate controls are needed. Nevertheless, this repre- poststroke ADHD benefit in a similar manner to children sents one of the largest reports of childhood stroke. Second, about one-third of the orthopedic control children were re-cruited from a different site than the children with stroke.
Unknown biases may be operative as a result of this. How- ACKNOWLEDGMENTS
ever, all controls were carefully selected to match the par- We thank Chris Cook for image processing, Stephan Arndt for ticipants in age, gender, SES, and the presence of a chronic statistical consultation, and Damien Ihrig and Jennifer Smith for medical condition. Furthermore, the stroke and control data collection and0or management. This study was supported groups did not differ on family function or family psychi- by NARSAD (Dr. Max); T32 MH18399 (Dr. Robertson); and atric history. Third, the psychiatric interviewer was not UTHSCSA portion of 5 PO2 MH52176-07 (NIMH, NIDA, NCI) blinded to the group affiliation of the participants. How- ever, excellent interrater reliability was recorded with an-other child psychiatrist who watched randomly selected REFERENCES
videotaped psychiatric interviews and who was blind togroup affiliation of the participants. In addition, a parent- Achenbach, T.M. (1991). Manual for the Child Behavior Checklist0 completed behavioral questionnaire found significant dif- 4-18 and 1991 Profile. Burlington, Vermont: University of Ver- ferences in attention problems in the group diagnosed with American Psychiatric Association (1994). Diagnostic and Statis- ADHD0 Traits by clinical interview. Fourth, premorbid tical Manual of Mental Disorders (4th ed.). Washington, DC: ADHD status was carefully assessed in the clinical inter- American Psychiatric Association Press.
view. Yet there remains the possibility that some proportion Andreasen, N.C., Endicott, J., Spitzer, R.L., & Winokur, G. (1977).
of participants with congenital conditions or even condi- The family history method using research diagnostic criteria: tions with onset in the first few years of life may have Reliability and validity. Archives of General Psychiatry, 34, developed ADHD0Traits regardless of their medical condi- tion. This problem was mitigated by using a control group Barkley, R.A. (1997). ADHD and the nature of self-control. New and also by the finding of a lack of association between family history of ADHD and children with ADHD Bloom, D.R., Levin, H.S., Ewing-Cobbs, L., Saunders, A.E., Song, Fifth, the stroke sample is not an epidemiological sample J., Fletcher, J.M., & Kowatch, R.A. (2001). Lifetime and novel but rather represents the results of a case-finding strategy of psychiatric disorders after pediatric traumatic brain injury. Jour-nal of the American Academy of Child and Adolescent Psychi- children diagnosed with stroke at a university teaching hos- pital. The stroke children were not referred for their psychi- Borkowski, J.G., Benton, A.L., & Spreen, O. (1967). Word flu- atric disorders including ADHD0 Traits but rather for ency and brain damage. Neuropsychologia, 5, 135–140.
neurological diagnosis, treatment for cardiac problems, or Botting, N., Powls, A., & Cooke, R.W.I. (1997). Attention deficit orthopedic procedures for residual neurologically based mus- hyperactivity disorders and other psychiatric outcomes in very culoskeletal problems. The controls were subject to similar low birthweight children at 12 years. Journal of Child Psychol-ogy and Psychiatry, 8, 931–941.
Breslau, N. & Chilcoat, H.D. (2000). Psychiatric sequelae of low Levy, F., Hay, D.A., McStephen, M., Wood, C., & Waldman, I.
birth weight at 11 years of age. Biological Psychiatry, 47, (1997). Attention-deficit hyperactivity disorder: A category or a continuum? Genetic analysis of a large-scale twin study. Jour- Costello, E.J., Angold, A., Burns, B.J., Erkanli, A., Stangl, D.K., & nal of the American Academy of Child and Adolescent Psychi- Tweed, D.L. (1996). The Great Smoky Mountains Study of Youth: Functional impairment and serious emotional distur- Lou, H.C. (1996). Etiology and pathogenesis of attention-deficit bance. Archives of General Psychiatry, 53, 1137–1143.
hyperactivity disorder (ADHD): Significance of prematurity Deutsch, C.K., Matthysse, S., Swanson, J.M., & Farkas, L.G. (1990).
and perinatal hypoxic-haemodynamic encephalopathy. Acta Genetic latent structure analysis of dysmorphology in attention Paediatrica, 85, 1266–1271.
deficit disorder. Journal of the American Academy of Child Marin, R.S., Fogel, B.S., Hawkins, J., Duffy, J., & Krupp, B. (1995).
and Adolescent Psychiatry, 29, 189–194.
Apathy: A treatable syndrome. Journal of Neuropsychiatry and Ebaugh, F.G. (1923). Neuropsychiatric sequelae of acute epi- Clinical Neuroscience, 7, 23–30.
demic encephalitis in children. American Journal of Diseases Max, J.E., Arndt, S., Castillo, C.S., Bokura, H., Robin, D.A., of Children, 25, 89–97.
Lindgren, S.A., Smith, W.L., Jr., Sato, Y., & Mattheis, P.J.
Gerring, J.P., Freund, L., Gerson, A.C., Joshi, P.T., Capozzoli, J., (1998a). Attention-deficit hyperactivity symptomatology after Frosch, E., Brady, K., Marin, R.S., & Denckla, M.B. (1996).
traumatic brain injury: A prospective study. Journal of the Amer- Psychometric characteristics of the Children’s Motivation Scale.
ican Academy of Child and Adolescent Psychiatry, 37, 841–847.
Psychiatry Research, 63, 205–217.
Max, J.E., Castillo, C.S., Lindgren, S.D., & Arndt, S. (1998b). The Gerring, J.P., Brady, K.D., Chen, A., Vasa, R., Grados, M., Bandeen- neuropsychiatric rating schedule: Reliability and validity. Jour- Roche, K.J., Bryan, R.N., & Denckla, M.B. (1998). Premorbid nal of the American Academy of Child and Adolescent Psychi- prevalence of ADHD and development of secondary ADHD after closed head injury. Journal of the American Academy of Max, J.E., Fox, P.T., Lancaster, J.L., Kochunov, P., Mathews, K., Child and Adolescent Psychiatry, 37, 647– 654.
Manes, F.F., Robertson, B.A.M., Arndt, S., Robin, D.A., & Goodman, R. & Graham, P. (1996). Psychiatric problems in chil- Lansing, A.E. (2002). Putamen lesions and the development of dren with hemiplegia: Cross sectional epidemiological survey.
attention-deficit 0hyperactivity symptomatology. Journal of the British Medical Journal, 312, 1065–1069.
American Academy of Child and Adolescent Psychiatry, 41, Goodman, R. & Yude, C. (1996). IQ and its predictors in child- hood hemiplegia. Developmental Medicine and Child Neurol- Max, J.E., Mathews, K., Lansing, A., Robertson, B.A.M., Fox, P., Lancaster, J., Manes, F.F., & Smith, J. (2002). Psychiatric dis- Goodman, R. & Yude, C. (1997). Do unilateral lesions of the orders after childhood stroke. Journal of American Academy of developing brain have side-specific psychiatric consequences Child and Adolescent Psychiatry, 41, 555–562.
in childhood? Laterality, 2, 103–115.
Max, J.E., Lansing, A.E., Koele, S.L., Castillo, C.S., Bokura, H., Goodman, R.N. (1989). Neuronal misconnections and psychiatric Schachar, R., Collings, N., & Williams, K.E. Attention deficit disorder: Is there a link? British Journal of Psychiatry, 154, hyperactivity disorder in children and adolescents following traumatic brain injury. Developmental Neuropsychology (in Grant, D.A. & Berg, E.A. (1948). A behavioral analysis of degree of reinforcement and ease of shifting to new responses in a Miller, I.W., Kabacoff, R.I., Epstein, N.B., & Bishop, D.S. (1994).
Weigl-type card sorting problem. Journal of General Psychol- The development of a clinical rating scale for the McMaster model of family functioning. Family Process, 33, 53– 69.
Hollingshead, A.B. (1975). Four factor index of social status. New Nussbaum, N.L., Grant, M.L., Roman, M.J., Poole, J.H., & Bigler, Haven, Connecticut: Yale University.
E.D. (1990). Attention deficit disorder and the mediating effect Ingram, T.T.S. (1956). A characteristic form of overactive behav- of age on academic and behavioral variables. Journal of De- iour in brain damaged children. Journal of Mental Science, velopmental and Behavioral Pediatrics, 11, 22–26.
Ounsted, C. (1955). The hyperkinetic syndrome in epileptic chil- Jastak, S. & Wilkerson, G.S. (1984). Wide Range Achievement dren. Lancet, 2, 303–311.
Test: Administration manual. Wilmington, Delaware: Jastak Pennington, B.F. & Ozonoff, S. (1996). Executive function and developmental psychopathology. Journal of Child Psychology Kaufman, J., Birmaher, B., Brent, D., Rao, U., Flynn, C., Moreci, and Psychiatry, 37, 51–87.
P., Williamson, D., & Ryan, N. (1997). Schedule for affective Posner, M.I. & Petersen, S.E. (1990). The attention system of the disorders and schizophrenia for school-age children—present human brain. Annual Reviews of Neuroscience, 13, 25– 42.
and lifetime version (K-SADS-PL): Initial reliability and va- Riva, D. & Cazzaniga, L. (1986). Late effects of unilateral brain lidity data. Journal of the American Academy of Child and lesions sustained before and after age one. Neuropsychologia, Adolescent Psychiatry, 36, 980–988.
Klorman, R., Hazel-Fernandez, L.A., Shaywitz, S.E., Fletcher, J.M., Roach, E.S. & Riela, A.R. (1995). Inflammatory vascular disor- Marchione, K.E., Holahan, J.M., Stuebing, K.K., & Shaywitz, ders. In E.S. Roach and A.R. Riela (Eds.), Pediatric Cerebro- B.A. (1999). Executive functioning deficits in attention-deficit0 vascular Disorders (2nd ed., pp. 121–139). Armonk, New York: hyperactivity disorder are independent of oppositional defiant or reading disorder. Journal of the American Academy of Child Rutter, M., Graham, P., & Yule, W. (1970). A neuropsychiatric and Adolescent Psychiatry, 38, 1148–1155.
study in childhood. Clinics in Developmental Medicine, Nos.
Lancaster, J.L., Glass, T.G., Lankipalli, B.R., Downs, H., May- 35036. London, UK: Heinemann0Spastics International Med- berg, H., & Fox, P.T. (1995). A modality-independent approach to spatial normalization. Human Brain Mapping, 3, 209–223.
Seidel, U.P., Chadwick, O.F., & Rutter, M. (1975). Psychological disorders in crippled children. A comparative study of children Whitaker, A.H., Van Rossem, R., Feldman, J.F., Schonfeld, I.S., with and without brain damage. Developmental Medicine and Pinto-Martin, J.A., Torre, C., Shaffer, D., & Paneth, N. (1997).
Child Neurology, 17, 563–573.
Psychiatric outcomes in low-birth-weight children at age 6 years: Shaywitz, B.A. & Shaywitz, S.E. (1994). Learning disabilities and Relation to neonatal cranial ultrasound abnormalities. Archives attention disorders. In K. Swaiman (Ed.), Principles of pediat- of General Psychiatry, 54, 847–856.
ric neurology (pp. 1119–1151). St. Louis, Missouri: Mosby.
Willcutt, E.G. & Pennington, B.F. (2000). Comorbidity of reading Shelton, T.L., Barkley, R.A., Crosswait, C., Moorehouse, M., disability and attention-deficit0hyperactivity disorder: Differ- Fletcher, K., Barrett, S., Jenkins, L., & Metevia, L. (1998).
ences by gender and subtype. Journal of Learning Disabilities, Psychiatric and psychological morbidity as a function of adap- tive disability in preschool children with aggressive and Wood, F.B. & Felton, R.H. (1994). Separate linguistic and atten- hyperactive-impulsive-inattentive behavior. Journal of Abnor- tional factors in the development of reading. Topics in Lan- mal Child Psychology, 26, 475– 494.
guage Disorders, 14, 42–57.
Sparrow, S.S., Balla, D., & Cicchetti, D. (1984). The Vineland Woods, B.T. (1980). The restricted effects of right-hemisphere Adaptive Behavior Scales. Circle Pines, Minnesota: American lesions after age one: Wechsler test data. Neuropsychologia, Wechsler, D. (1991). Wechsler Intelligence Scale for Children– Third Edition. New York: Psychological Corporation.

Source: http://www.iapediatricstroke.org/StrokeADHD%20Neurocognitive%20Correlates.pdf