neuropsychologiczne korelaty symptomów adhd.pdf

2005, Vol. 19, No. 4, 446–455

Neuropsychological Correlates of ADHD Symptoms in Preschoolers

David J. Marks

Mount Sinai Medical Center

Olga G. Berwid, Amita Santra, Elizabeth C. Kera,

and Shana E. Cyrulnik

Graduate Center of the City University of New York

Jeffrey M. Halperin

Mount Sinai Medical Center, Graduate Center of the City University of New York, and Queens College of the City

University of New York

The authors examined the neuropsychological status of 22 preschoolers at risk for attention-deﬁcit/

hyperactivity disorder (ADHD) and 50 matched control children, using measures of nonverbal working

memory, perceptual and motor inhibition, and memory for relative time. All tasks included paired control

conditions, which allowed for the isolation of discrete executive function constructs. Group differences

were evident on several measures of neuropsychological functioning; however, after accounting for

nonexecutive abilities, no deﬁcits could be attributed to speciﬁc functions targeted by the tasks.

Performance on executive measures was not related to objective indices of activity level or ratings of

ADHD symptoms. Yet, the fact that at-risk preschoolers were highly symptomatic casts doubt on whether

executive function deﬁcits and/or frontostriatal networks contribute etiologically to early behavioral

manifestations of ADHD.

Keywords: ADHD, preschool children, executive function

Attention-deﬁcit/hyperactivity disorder (ADHD) is a clinically

heterogeneous syndrome characterized by symptoms of inatten-

tion, hyperactivity, and impulsivity that are developmentally inap-

propriate and negatively impact an individual’s psychosocial func-

tioning. Research exploring the biological underpinnings of

ADHD has focused on a host of etiologic mechanisms, and al-

though the speciﬁc nature of the dysfunction remains unknown,

data from the ﬁelds of neuropsychology (Barkley, 1994, 1997;

Pennington & Ozonoff, 1996), neurochemistry (Solanto, 1998),

neuroimaging (Hynd et al., 1993), and molecular genetics (Rowe

et al., 1998) suggest that the core behavioral features of the

disorder and associated neurocognitive deﬁcits likely reﬂect a

multifactorial etiologic process involving circuits of the prefrontal

cortex.

Several investigators have suggested that the cognitive sequelae

of ADHD primarily encompass deﬁcits in executive functioning.

Although many deﬁnitions have been proposed (Eslinger, 1996),

executive functions are believed to incorporate planning, self-

regulation, organizational skills, goal setting, problem solving, and

judgment (Lezak, 1993) and are thought to be mediated by recip-

rocal connections between the prefrontal cortex and basal ganglia.

Consistent with such models, school-age children with the disorder

have been shown to perform signiﬁcantly worse than control

children on a host of neuropsychological indices believed to be

subserved by frontal–subcortical pathways, including measures of

working memory (Barkley, Grodzinsky, & DuPaul, 1992), plan-

ning (Pennington, Grossier, & Welsch, 1993), cognitive ﬂexibility

(Chelune, Fergusson, Koon, & Dickey, 1986; Tripp, Ryan, &

Peace, 2002), time perception (Rubia, Taylor, Taylor, & Sergeant,

1999), motor inhibition (Oosterlaan & Sergeant, 1985; Trommer,

Hoeppner, Lorber, & Armstrong, 1988), and phonemic ﬂuency

(Grodzinsky & Barkley, 1999).

While some investigators have gone so far as to suggest that the

symptoms of ADHD and related information-processing distur-

bances result from a primary deﬁcit in frontostriatal executive

functions, other researchers have been unable to provide empirical

support for such claims (Loge, Staton, & Beatty, 1990) or have

argued that executive theories of ADHD should be interpreted

cautiously because of a host of methodological limitations (e.g.,

small sample sizes, variability in selection and/or diagnostic cri-

teria), small effect sizes (Pennington & Ozonoff, 1996), and, in

some cases, disregard for psychiatric comorbidity (Sergeant,

Geurts, & Oosterlaan, 2002). Other researchers (e.g., Tripp et al.,

2002) have suggested that the lack of behavioral control observed

in children with ADHD might adversely affect their test-taking

skills, irrespective of the constructs under investigation, which in

turn could contribute to deﬁcits in nonexecutive domains (e.g.,

language, spatial abilities, etc.). Moreover, the fact that ADHD

probands, even after controlling for psychiatric comorbidity (Fara-

one, Biederman, Weber, & Russell, 1998), typically perform more

poorly than matched controls on measures of cognitive functioning

(Shallice et al., 2002) raises the possibility that intellectual com-

promise, not executive dysfunction per se, accounts for group

David J. Marks, Division of Child and Adolescent Psychiatry, Mount

Sinai Medical Center. Olga G. Berwid, Amita Santra, Elizabeth C. Kera,

and Shana E. Cyrulnik, Neuropsychology Doctoral Program, Graduate

Center of the City University of New York. Jeffrey M. Halperin, Division

of Child and Adolescent Psychiatry, Mount Sinai Medical Center; Neuro-

psychology Doctoral Program, Graduate Center of the City University of

New York; Department of Psychology, Queens College of the City Uni-

versity of New York.

This research was supported by Professional Staff Congress–City Uni-

versity of New York Grant 6269500-31 awarded to Jeffrey M. Halperin.

Correspondence concerning this article should be addressed to David J.

Marks, Division of Child and Adolescent Psychiatry, Box 1230, Mount

Sinai Medical Center, 1 Gustave L. Levy Place, New York, NY 10029.

E-mail: david.marks@mssm.edu

446

Neuropsychology

0894-4105/05/$12.00 DOI: 10.1037/0894-4105.19.4.446

NEUROPSYCHOLOGICAL CORRELATES OF PRESCHOOL ADHD

447

differences on neuropsychological indices. By extension, control-

ling for intellectual abilities has been shown to eradicate or atten-

uate group disparities in neuropsychological status (Seidman,

Biederman, Faraone, Weber, & Ouellette, 1997; Tripp et al.,

2002).

Although studies of school-age children are replete with explo-

rations into the neurobiological and neurocognitive underpinnings

of ADHD, such efforts have been far more limited with samples of

preschool children (Baving, Laucht, & Schmidt, 1999; Mariani &

Barkley, 1997). Several factors may account for this dichotomy.

First, neuroimaging techniques are difﬁcult to use with preschool

children because of the confounding effects of movement artifact,

and therefore, relative to other forms of data, may be less able to

tap into functional/neurocognitive impairments. Further, among

the measures used to evaluate information-processing models or

executive dysfunction theories, many are inappropriate for use

with preschoolers because they are often too long, insufﬁciently

engaging, or require reading skills that have not yet developed.

Within the past decade, a number of researchers (e.g., Corkum,

Byrne, & Ellsworth, 1995; Espy, Kaufmann, McDiarmid, &

Glisky, 1999) have responded to this perceived deﬁciency by

developing a host of tasks appropriate for use with preschool

children. In large part, such investigations have demonstrated that

the preschool years are characterized by rapid cognitive growth

typiﬁed by the emergence and maturation of elementary abstrac-

tion, aspects of cognitive ﬂexibility (Jacques & Zelazo, 2001),

response inhibition, basic set-switching skills (Espy, 1997), rule

use, working memory (Espy et al., 1999), visual search (Welsch,

Pennington, & Grossier, 1991), motor inhibition, impulse control,

planning, and attention regulation (Klenberg, Korkman, & Lahti-

Nuuttila, 2001). A subset of tasks designed to assess these con-

structs may be most appropriately conceptualized as measures

used to assess the core behavioral features of ADHD (e.g., atten-

tional deﬁcits and overactivity), whereas others have been adapted

from animal models and studies of human adults to speciﬁcally

assess domains of neuropsychological functioning (e.g., working

memory, inhibitory control, set shifting).

Byrne, DeWolfe, and Bawden (1998) reported that, relative to

matched controls, preschoolers with ADHD committed a signiﬁ-

cantly greater number of commission errors on a selective attention

task and were comparatively unable to restrain themselves from

touching a series of off limits toys (77% of probands vs. 0% of

controls). In contrast, they found no group difference in the num-

ber of cancellation task omission errors and did not identify dis-

crepancies in ratings of off-task behavior obtained during the

structured assessment procedure. Contrary to the nearly ubiquitous

reports of hyperactivity communicated by parents of ADHD chil-

dren during the interview process, both groups evinced similar

levels of out-of-seat behavior during structured and unstructured

periods and were indistinguishable with regard to their degree of

mobility while engaged in unstructured play. However, a subse-

quent study by the same investigators (DeWolfe, Byrne, & Baw-

den, 2000) using a larger cohort of preschool children did not

replicate many of their earlier results. Speciﬁcally, no signiﬁcant

group differences emerged in objectively deﬁned levels of impul-

sivity (rates of unsanctioned play: 60% of probands vs. 36% of

controls), whereas robust discrepancies were reported for measures

of off-task behavior and observed mobility during a play period

(ADHD controls).

As to the question of whether preschoolers with ADHD

display neuropsychological deﬁcits that mirror their school-age

counterparts, the answer has yet to be conclusively determined.

Relative to matched controls, preschool boys with ADHD have

been shown to exhibit deﬁcits in motor control and working

memory (Mariani & Barkley, 1997) and impairments in aspects

of academic achievement. Hughes, Dunn, and White (1998)

reported signiﬁcant group differences between hard-to-manage

preschoolers and control preschoolers on four of six executive

function indices, including measures of inhibitory control,

working memory, and cognitive ﬂexibility. However, virtually

all group differences were no longer signiﬁcant after controlling

for verbal ability and socioeconomic status. In addition, the fact

that performance on such tasks was also highly associated with

acts of aggression and antisocial behavior (Hughes, White,

Sharpen, & Dunn, 2000) calls into question whether patterns of

executive dysfunction were related to features of ADHD per se.

Sonuga-Barke, Dalen, Daley, and Remington (2002), after par-

tialing out the effects of age and IQ, did not ﬁnd an association

between working memory or planning skills and symptoms of

ADHD in a heterogeneous sample of preschool children.

Rather, the authors proposed that compromised behavioral in-

hibition may serve as a harbinger for the emergence of more

global executive function deﬁcits that appear later in

development.

Taken together, the ﬁndings from the above studies provide

limited support for the notion that preschool children with mani-

festations of ADHD display unique patterns of neurocognitive

functioning relative to normally developing preschool youths. Un-

fortunately, such ﬁndings have been inconsistent because of vari-

ations in selection criteria, assessment techniques, and statistical

methodology (e.g., covariation for age and/or IQ), which in turn,

render it difﬁcult to draw precise conclusions regarding both the

concurrent and predictive utility of neuropsychological parameters

in disruptive preschoolers. Furthermore, none of the tasks used in

previous studies incorporated paired reference conditions to isolate

the construct or constructs of interest. Consequently, it is unclear

whether reports of executive dysfunction reﬂect genuine deﬁcits in

speciﬁc neurocognitive domains (e.g., visual working memory) or

whether such phenomena occur secondary to problems with basic-

level functions (e.g., visual form discrimination) subsumed by

executive function indices.

The current study sought to remedy this gap by using four

novel executive function tasks designed to assess the following:

(a) memory for time and temporal sequencing, (b) nonverbal

working memory, (c) perceptual inhibition (i.e., susceptibility

to interference), and (d) motor inhibition (i.e., inhibition of

prepotent responding). All neuropsychological indices con-

sisted of both experimental and control conditions to isolate

process speciﬁcity and have been designed to be both brief and

engaging. Key issues addressed by this investigation included

(a) the extent to which at-risk and typically developing pre-

school children differ on the above executive function param-

eters, (b) the relationship of laboratory measures of activity

level to performance on the executive function tasks, and (c) the

association between early neuropsychological status and parent

and teacher ratings of ADHD symptoms.

448

MARKS ET AL.

Method

group (51 boys, 51 girls). The ratio of boys to girls among at-risk partic-

ipants (3:1) signiﬁcantly exceeded that of controls (1:1),

2 (1, N

Participants

.002. Among the 102 children from the Phase 1 cohort

who met criteria for the control group, 50 participants (23 boys, 27 girls)

agreed to come to the laboratory at Queens College for further on-site

assessments. Control participants who were and were not seen for Phase 2

evaluations did not differ signiﬁcantly with regard to parent and teacher

ratings of ADHD symptom domains (all p s .05). Among the 56 children

who met criteria for at-risk status, 22 (18 boys, 4 girls) were seen for

Phase 2 assessments. Once again, at-risk participants who did and did not

complete the Phase 2 evaluations did not differ signiﬁcantly with respect to

dimensional ratings of inattention and hyperactivity–impulsivity completed

by parents and teachers (all p s

9.33, p

Recruitment was conducted using a two-step procedure, previously

described by Campbell, Ewing, Breaux, and Szumowski (1986), to gather

parent and teacher ratings on a relatively large sample of children from

which an assessment cohort could then be selected. During this segment of

the project (Phase 1), principals of private and public preschools located in

Queens, New York, were contacted and permission was requested to screen

the school for children with attention and behavior problems. A subset of

Phase 1 participants was subsequently selected for inclusion in the on-

campus assessment (Phase 2) portion of the study. Selection procedures for

both project components are outlined below.

Phase 1 screening procedure. Thirteen preschools within close prox-

imity to Queens College agreed to participate, and parents of children in

these schools were sent Diagnostic and Statistical Manual of Mental

Disorders (4th ed.; DSM–IV ; American Psychiatric Association, 1994)

ADHD checklists along with consent forms that would allow us to collect

similar ratings from the child’s teacher. The checklists consisted of the 18

ADHD behaviors listed in DSM–IV , which were rated on a 4-point scale

.05).

As shown in Table 1, the two groups that came to the laboratory (i.e., at

risk and control) did not differ signiﬁcantly in age or estimated intellectual

functioning (according to the Wechsler Preschool and Primary Scale of

Intelligence—Revised Information subtest, Wechsler, 1989). However, the

at-risk group included disproportionately more boys relative to the control

group (81% vs. 46%),

.01. Signiﬁcant

differences were also observed with regard to dimensional ratings of

inattention and hyperactivity–impulsivity symptoms reported by parents

and teachers, which is not surprising given that DSM–IV ADHD checklist

ratings were used for group classiﬁcation. However, the groups also

differed signiﬁcantly in activity level as measured using solid-state acti-

graphs, providing further support for their validity (see Table 1). Dispar-

ities in objectively assessed activity level occurred in the absence of

differences in actigraph variability over time (all p s

2 (1, N

72)

7.35, p

very much ). As

has been suggested by other investigators (e.g., DuPaul et al., 1998;

Kadesj¨ , Kadesj¨, H¨ggl¨f, & Gillberg, 2001), a symptom was considered

to be present if it received a rating of 2 ( pretty much )or3( very much ). All

consent forms and behavioral rating scales were returned directly to the

principal investigator to preserve both the conﬁdentiality of parent reports

and the objectivity of teacher ratings.

As a result of these mailings, we received consent and both parent and

teacher ratings for 212 children (123 boys, 89 girls). Children for whom

DSM–IV ADHD checklists were completed ranged in age from 2.84 to 5.92

years with a mean age of 4.38 (S D

not at all ,1

s omewhat, 2

pretty much, and 3

.05; see Table 1). Of

note, separate consent forms were used for each portion (i.e., Phases 1 and

2) of the study, and parents were informed during the screening phase and

subsequent invitational telephone contact that participation in Phase 1 did

not obligate them to participate in the on-campus assessment.

Although racial/ethnic data were not collected on the screening form,

among the children evaluated in the laboratory, 35% were Caucasian, 3%

were African American, 21% were Latino, 24% were Asian American, and

18% were of other or mixed ethnicity according to demographic data

obtained from the parents of all Phase 2 participants. The families who

participated in the on-campus evaluation were of primarily middle-class

status, with 47% reporting a family income greater than $70,000 per year,

and 90% of the families reporting a total income above $25,000 per year.

Additionally, 44% of fathers and 48% of mothers reportedly had received

4-year undergraduate degrees. Among the children, 86% lived with both

parents, who were married. Parents were compensated $10.00 for trans-

0.63) years.

Selection of groups for Phase 2. Designation of at-risk and control

status was based on combined parent–teacher ratings modeled after

DSM–IV ADHD criteria. To be considered at risk, a child needed to receive

a pretty much or very much rating on at least six inattention or hyperactive–

impulsive items by either informant. To meet criteria for the control group,

a child needed to receive a pretty much or very much rating on fewer than

three inattention and three hyperactive–impulsive items according to both

parent and teacher ratings. This more liberal method of classiﬁcation of

at-risk status was used to cast a wide net for children with behavioral

difﬁculties.

On the basis of these criteria, 56 (25.3%) children (42 boys, 14 girls) met

criteria for the at-risk group, and 102 (46.2%) met criteria for the control

Table 1

Phase 2 Sample Characteristics

Control ( N 50)

At risk ( N 22)

Variable

t or F

Age

4.23

0.69

4.46

0.46

1.43

.10

WPPSI-R information

10.50

3.44

8.95

3.00

1.65

.10

Hyperactivity–impulsivity total: Parent

4.66

3.05

16.95

5.33

10.11

.001

Inattention total: Parent

4.74

3.05

10.68

6.05

4.37

.001

Hyperactivity–impulsivity total: Teacher

2.70

4.59

12.68

10.26

4.38

.001

Inattention total: Teacher

2.56

3.61

10.36

8.21

4.28

.001

Actigraph: Ankle a

655.03

117.24

1,175.77

196.79

5.12

.05

Actigraph: Waist a

220.87

37.90

381.41

63.61

4.66

.05

Actigraph variability: Ankle b

2,269.46

227.34

2,865.16

381.61

1.78

.10

Actigraph variability: Waist b

681.31

61.43

868.81

103.12

2.42

.10

Note. WPPSI-R Wechsler Preschool and Primary Scale of Intelligence—Revised.

a Data reﬂect age-controlled median and standard error values, respectively.

b Age-controlled data correspond to average intraindividual dispersion around

actigraph values and standard error values, respectively.

158)

NEUROPSYCHOLOGICAL CORRELATES OF PRESCHOOL ADHD

449

portation costs associated with the campus visit; children were provided

with snacks and stickers during the assessment.

There were no gender or ethnic restrictions, but both the children and

their parents were required to be English speaking. Children diagnosed

with mental retardation, a pervasive developmental disorder, or a neuro-

logic disorder (e.g., epilepsy) and those who were taking systemic medi-

cation for a chronic medical condition were excluded from participation.

3. Executive function battery. The Perceptual and Motor Conﬂict Test

(Nassauer & Halperin, 2003) consists of separate perceptual (stimulus) and

motor (response) conﬂict components, both of which required the child to

be seated in front of a computer monitor with an oversized two-button

computer mouse.

The SCT required fewer than 5 min to administer and consisted of three

conditions. For Condition 1, a series of left- or right-pointing arrows

appeared in the center of the screen. The child was required to press the

button corresponding to the side to which the arrow was pointing (direc-

tion). For Condition 2, a series of rectangular boxes appeared on either the

left or right side of the monitor, and the child pressed the button on the side

on which the rectangle appeared (location). Condition 3 consisted of trials

in which either left- or right-pointing arrows appeared on either the left or

right side of the screen. Children were required to inhibit the prepotent

response to stimulus location and to respond instead to arrow direction.

Errors and reaction times (RTs) for Condition 3 were shown to be signif-

icantly greater than for either of the other two conditions (Nassauer &

Halperin, 2003). For all three conditions, stimuli remained on the screen

until the child responded, with a subsequent interstimulus interval of 1,000

ms prior to the onset of the next stimulus. This task assessed the construct

of interference control similar to that of the Stroop Color and Word Test

(Golden, 1978). However, unlike the Stroop test, it can be used with young

children because it does not require reading.

The RCT is a brief ( 5 min) computerized measure of response orga-

nization and motor inhibition. The ﬁrst (compatible) condition was iden-

tical to that of the SCT; children were asked to press a button that

corresponded to the direction that a centrally located arrow was pointing.

Following 20 trials, the task changed such that the child was asked to press

the button that was on the opposite side from where the arrow was pointing

(incompatible condition). The child was therefore required to inhibit a

prepotent response and engage a new response set. Data indicated signif-

icant differences in error rates and RTs between compatible and incom-

patible task conditions (Nassauer & Halperin, 2003). As in the SCT, stimuli

remained on the screen until the child responded, with a subsequent

interstimulus interval of 1,000 ms prior to the onset of the next stimulus.

The DNMST is a nonverbal memory task that required the children to

hold a visual image online and is analogous to tasks used with monkeys to

demonstrate deﬁcits following prefrontal cortex lesions (Goldman, Ros-

vold, Vest, & Galkin, 1971). Nonverbalizable ﬁgural stimuli were pre-

sented to the children on a computer screen for 4 s, followed by a 1-s delay.

Then, a response screen containing the original ﬁgure and one new ﬁgure

was presented. The children were required to identify the new ﬁgure by

pointing to it. This task’s difﬁculty level increased every three trials, such

that the ﬁrst three trials contained stimuli with a single ﬁgure and a

response screen containing two ﬁgures, the second level contained two

ﬁgural stimuli and three response options, and so forth. The control task

was a nonmatching-to-sample task with no delay: The children viewed the

stimulus and response ﬁgures simultaneously. Both task conditions were

administered until a child got two or three trials incorrect within a given

level (up to four task levels; possible score range 0–12).

The RMT was designed to measure memory for relative time and

temporal sequencing and has been shown to effectively distinguish patients

with prefrontal cortex lesions from those with lesions in other parts of the

brain (Milner & Petrides, 1984). Verbalizable pictorial stimuli were pre-

sented on a computer screen at a rate of one per second. After the pictured

stimuli were displayed, a response screen containing two pictures was

presented. In the control (recognition) condition, only one response picture

had been previously seen, and the child was asked to indicate, by pointing,

which picture had been seen before. In the memory-for-time condition,

both pictures on the response screen had been previously seen, and the

children indicated which picture had been seen last, or most recently.

Similar to the DNMST, conditions increased in difﬁculty every three trials;

Procedures

The on-campus assessment took approximately 2 hr to complete and

involved the gathering of demographic information from the parent and

testing data from the child. Three forms of data were collected from the

children: (a) objective measures of activity level to validate the distinction

between at-risk and control groups, (b) an estimate of overall cognitive

function to characterize the sample, and (c) measures to assess four aspects

of executive functioning, including inhibition of response to distracting

stimuli (Stimulus Conﬂict Task [SCT]), response organization (Response

Conﬂict Task [RCT]), nonverbal working memory (Delayed Nonmatching

to Sample Test [DNMST]), and memory for time (Recency Memory Test

[RMT]). In light of reports highlighting the importance of process speci-

ﬁcity (Sergeant et al., 2002), all neuropsychological indices were devel-

oped in our laboratory and consisted of both experimental and control/

reference conditions that would allow for the isolation of the constructs of

interest. All participants were administered practice trials (on paper and

computer) prior to the start of each task to ensure comprehension of the

relevant instructions. Practice trials were repeated until participants could

demonstrate a sufﬁcient level of proﬁciency, operationalized as two or

three correct out of three practice trials for the DNMST and RMT. More

subjective criteria were used to gauge comprehension of SCT and RCT

instructions. A ﬁxed test-administration order was used to systematically

vary task types to help maintain interest and limit fatigue effects.

1. Objective measures of activity level. Motor activity was recorded

throughout the Phase 2 assessment using two solid-state actigraphs worn

around the waist and nondominant ankle. Both actigraphs (Model WAM-

7164) were obtained from Computer Sciences and Applications, Inc., of

Shalimar, Florida, and utilize internal accelerometers (acceleration magni-

tude, 0.05–2.00 G; frequency response, 0.25–2.50 Hz) to store data on the

number of movements per unit time. Prior to each Phase 2 assessment, both

actigraphs were initialized using a reader interface unit and programmed to

obtain movement counts in 60-s epochs. After each evaluation, the acti-

graphs were returned to the reader interface unit, and the data were

downloaded into Microsoft Excel spreadsheets. Descriptive statistics (i.e.,

mean, median, and standard deviation) were generated separately for each

participant’s ankle and waist actigraph data, using computational algo-

rithms included within Microsoft Excel. The same two actigraphs were

used during all Phase 2 assessments and were consistently placed on the

same location on each child (i.e., waist and nondominant ankle).

It is important to note that assessments of activity level taken during

structured test sessions are reliable (Reichenbach, Halperin, Sharma, &

Newcorn, 1992), yield measures that are correlated with parent and teacher

ratings of hyperactivity (Fairweather, Reilly, Grant, Whittaker, & Paton,

1999; Reichenbach et al., 1992), and have been shown to effectively

discriminate hard-to-manage preschool boys from comparison controls

(Campbell, Pierce, March, Ewing, & Szumowski, 1994). In addition,

Fairweather et al. (1999) have shown Computer Sciences and Applications

actigraphs to be a valid method for assessing physical activity in preschool

children and have identiﬁed signiﬁcant associations between actigraph

readings obtained from different locations on the torso (i.e., left vs. right

hip).

2. General cognitive functioning: Wechsler Preschool and Primary

Scale of Intelligence—Revised Information subtest (Wechsler, 1989).

This well-normed subtest has a strong association with general intelligence

and was administered to characterize the sample and to identify possible

group differences with regard to general cognitive ability.

450

MARKS ET AL.

administration was discontinued if a child provided two incorrect responses

out of three within a task level (up to eight task levels; possible score range

for each condition

RCT

0–24).

As predicted, a signiﬁcant main effect for condition was

identiﬁed, such that participants, regardless of group, commit-

ted signiﬁcantly more errors on the incompatible condition

relative to the compatible condition, F (1, 59) 14.41, p

.001, 2

Data Analysis

Objective 1: Proﬁles of neuropsychological status in at-risk versus

control preschoolers. Data from each of the executive function tasks

were submitted to a Group (control vs. at risk) Condition (control vs.

experimental) mixed factorial analysis of variance. Using this approach, a

main effect for condition, reﬂecting superior performance in the control

condition, would support the validity of the task manipulation. A main

effect for group, with those in the at-risk group performing more poorly,

would suggest greater cognitive impairment in that group. Given that

control and experimental task conditions were identical except for the

degree of executive function involvement, as posited by the additive factors

model (Sergeant et al., 2002), a deﬁcit speciﬁc to a given executive

function would be reﬂected by signiﬁcant a Group

.20 (see Figure 1A). In addition, a signiﬁcant main

effect for group status was also observed, with at-risk partici-

pants committing signiﬁcantly more errors across both RCT

conditions, F (1, 59) 29.53, p .001, 2

.33. However, as

evidenced by the parallel slopes, at-risk preschoolers did not

perform differentially worse on the incompatible condition rel-

ative to the compatible condition, F (1, 59) 0.40, p .10, 2

.007.

A similar pattern emerged when RT was used as the dependent

measure. Speciﬁcally, both groups evinced longer RTs for the

incompatible condition relative to the compatible condition: con-

dition main effect, F (1, 59) 12.94, p .001, 2

Condition interaction.

RTs and error rates served as dependent measures for the RCT and SCT.

There were two levels of condition for the RCT (compatible and incom-

patible) and three levels for the SCT (direction, location, and conﬂict). We

expected to ﬁnd a main effect of condition, such that children, irrespective

of group, would make more errors and exhibit longer RTs on the incom-

patible/conﬂict conditions relative to the compatible conditions; these

ﬁndings would support the validity of the task manipulations. It was also

predicted that at-risk preschoolers would display circumscribed deﬁcits in

motor inhibition (RCT) and interference control (SCT) relative to controls,

by performing disproportionately worse on the incompatible conditions

compared with the control conditions (Group

.18 (see

Figure 1B). At-risk participants did show a tendency to respond

more rapidly irrespective of condition; however, this ﬁnding did

not reach statistical signiﬁcance: group main effect, F (1,

59) 3.55, p .06, 2

.06. Finally, no signiﬁcant Group

Condition interaction was identiﬁed, F (1, 59) 1.01, p .10, 2

.02.

SCT

Condition interactions).

For the DNMST and RMT, the number of correct responses served as

dependent measures. As described above, group-speciﬁc deﬁcits in the

domains assessed by the two tasks (i.e., nonverbal working memory and

memory for relative time, respectively) would require a signiﬁcant

Group

.31 (see Figure 2A). Post hoc comparisons revealed that both

groups committed signiﬁcantly more errors on Condition 3 relative

to Conditions 1 and 2. There was no signiﬁcant main effect for

group, F (1, 64) 2.59, p .10, 2

Condition interaction, with at-risk preschoolers performing dif-

ferentially worse than controls on the experimental condition.

Objectives 2 and 3: Relationship of neuropsychological status to labo-

ratory measures of activity level and parent and teacher ratings of ADHD

symptoms. Linear regression procedures were used to remove the effects

of age on waist and ankle actigraph counts and to generate standardized

residual values. Performance discrepancies between the experimental and

control conditions were then calculated for each neuropsychological task.

For the RCT and SCT, the total number of errors on the control condition

was subtracted from the number of errors committed on the experimental

condition. For the DNMST and RMT, the number of correct responses on

the experimental condition was subtracted from the total number of correct

responses on the control condition. Finally, two-tailed Bonferroni-cor-

rected Pearson product–moment correlations were calculated to examine

the associations between each difference score and actigraph counts ob-

tained from the waist and ankle. Median movement counts were used in all

actigraph analyses to minimize the impact of outliers. Similarly, correla-

tions between the task difference scores and parent and teacher ratings of

hyperactivity–impulsivity (dimensional summary scores) were examined.

.04, nor was there a

signiﬁcant Group Condition interaction, F (2, 63) 1.02, p

.10, 2

.03.

Using RT as the dependent measure yielded a signiﬁcant main

effect for condition, F (2, 63) 18.47, p .001, 2

.37 (see

Figure 2B), with both groups exhibiting longer RTs on Condition 3

compared with Conditions 1 and 2. At-risk participants also re-

sponded more rapidly than controls across all three task condi-

tions: group main effect, F (1, 64) 4.42, p .04, 2

.07.

However, group differences in RT did not systematically vary as a

function of SCT condition: Group Condition interaction, F (2,

63) 2.50, p .09, 2

.07.

DNMST

As was the case for both the RCT and SCT, a main effect for

condition was found, with participants performing signiﬁcantly

worse on the experimental condition relative to the control condi-

tion, F (1, 52) 138.20, p .001, 2

.78 (see Figure 3).

However, there was no signiﬁcant main effect for group, F (1,

52) 0.11, p .10, 2

Results

.002, nor was there a signiﬁcant

Group Condition interaction, F (1, 52) 2.26, p .10,

Although the nature of the DNMST and the RMT precluded

evaluations of the tasks’ reliability, split-half reliability estimates

were calculated using error rates and RT variables from the dif-

ferent conditions of the RCT and SCT. Split-half reliabilities for

RCT error rates and RTs ranged from .71 to .77 and from .51 to

.61, respectively. Reliability estimates for error rates and RTs on

the SCT ranged from .78 to .84 and from .45 to .77, respectively.

.04.

RMT

As shown in Figure 4, a signiﬁcant main effect for condition

again supported the validity of the task manipulation, with children

In support of task validity, a signiﬁcant main effect for condition

was observed for total errors, F (2, 63) 14.41, p .001, 2

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: