|
Neuropsychological Evaluation
Based on both the literature and our previous experience with neuropsychological
assessment, we chose the following neuropsychological battery: 1) to assess
broad executive function, the computerized WCST (38); 2) to assess selective
attention, the abbreviated version of the ST (39); c) to assess attention,
2 additional specific measures (Digit Span [40] and Word Span [41]).
The WCST is a neuropsychological test that assesses the ability to form
abstract concepts, to sustain attention, and to shift cognitive set flexibly
in response to changing conceptual rules while inhibiting inappropriate
responses. It assesses organizational capacity, attention shifting, and
sustained attention. The WCST is generally considered to be sensitive to
frontal lobe dysfunction and is one of the most commonly used tests for
executive function in the school-aged population (42).
The ST measures the subject’s ability to shift perceptual set in response
to changing demands and to concentrate selectively or attend in situations
requiring inhibition of responses. The precise nature of the information-processing
mechanisms revealed by the ST remains controversial, but the potential
usefulness of this procedure for evaluating the efficiency of selective
attention in ADHD is compelling (5).
The Digit Span provides a simple measure of attention (40). The score reflects
the amount of material on which the subject can maintain focus during a
given time period. The Word Span requires subjects to repeat immediately
a list of words (nouns) just heard (41). This test assesses attention and
immediate memory of verbal content. Both tests have been used previously
in ADHD studies (8,17,43,44).
All tests were administered and scored by trained examiners who were unaware
of subjects’ ADHD status. The tests were always given in a fixed order,
and the battery required about 30 minutes to administer. Complete data
were obtained for all subjects. To estimate the adolescents’ overall IQ
(3,33), trained psychologists administered the vocabulary and block design
subtests of the Weschler Intelligence Scale-Third Edition (40).
We defined socioeconomic status (SES) according to a standard socioeconomic
measure frequently used in Brazil, the Socioeconomic Scale of the Brazilian
Association of Market Research Institutes (45). The ethical committee of
our university hospital approved the project.
Data Analysis
The comparisons among ADHD subtypes and control subjects on both demographic
and neuropsychological variables were performed using analysis of variance
(ANOVA) for those variables that showed a normal distribution. Differences
were located by least-squares means (LSMEANS). To check possible covariates,
we performed a partial correlation (residual analysis) among outcome (neuropsychological)
variables and both IQ and demographic variables (46). If correlations were
detected, we used Analysis of Covariance (ANCOVA). For variables that did
not show a normal distribution, we used the Kruskal-Wallis (KW) 1-way Anova,
and differences were located by the Dunn Test (47). We accepted a significance
level of 5% in all comparisons in this exploratory study. All statistical
tests were carried out using SPSS, version 8.0 for Windows (48) and SAS,
version 6.12 for Windows (49).
Results
The demographic data for the overall (n = 30) and specific ADHD groups
and the control group (n = 60) can be seen in Table 1. We found no significant
differences on demographic variables and IQ among groups, except for education
(KW = 9.4, df 3, P < 0.05). In post hoc analysis, the difference was located
only between adolescents with ADHD-C and control subjects (Q = 2.64, P
< 0.05) (Table 1).
Partial correlation analysis demonstrated the following significant correlations:
1) total of errors score on the WCST and education (r = –0.26, P < 0.05);
2) Stroop Color-Word (time to complete the test) score and education (r
= –0.43, P < 0.001), age (r = –0.26, P < 0.05), and SES (r = 0.33, P <
0.01). We therefore considered these demographic variables covariates in
statistical analysis (ANCOVA) that included these neuropsychological tests.
No effect of sex and IQ were detected in any measure.
Subjects with ADHD-I or ADHD-C had a worse neuropsychological performance
than did control subjects. Subjects with ADHD-HI did not show significant
differences on any test of the neuropsychological battery, compared with
control subjects.
Regarding the WCST, we detected a significant difference among groups in
the total errors score (ANCOVA, F = 3.02; df 3; P < 0.05). The difference
was localized between ADHD-C and the control group (LSMEANS, P < 0.05),
the ADHD-I group (LSMEANS, P < 0.05), and the ADHD-HI group (LSMEANS, P
< 0.01). We also detected a significant difference among groups in the
conceptual responses score (KW = 8.5, df 3, P < 0.05). The ADHD-C group
showed a lower score than did the control group (Q = 2.64, P < 0.05) (Table 2).
Regarding the ST, we detected a significant difference among groups in
the Color-Word score (KW = 13.8, df 3, P < 0.01). The difference was localized
between the group with ADHD-I and the control group (Q = 2.64, P < 0.01).
We also detected a significant difference among groups in the amount of
time needed to complete the test (ANCOVA, F = 3.8; df 3; P < 0.05). The
group with ADHD-I took more time to complete the test (ANCOVA, F = 3.8;
df 3; P < 0.05) than did the control group (LSMEANS, P < 0.01), the ADHD-C
group (LSMEANS, P < 0.01), and the ADHD-HI group (LSMEANS, P < 0.05) (Table 2).
We found a significant difference among groups in the Digit Span (ANOVA,
F = 6.9; df 3; P < 0.001). After post hoc analysis, the differences were
located between the control group and both the ADHD-C group (LSMEANS, P
< 0.001) and the ADHD-I group (LSMEANS, P < 0.05). In addition, the group
with ADHD-C showed a score that differed significantly from the ADHD-HI
group (LSMEANS, P < 0.01) (Table 2).
|
Table 1 Demographic characteristics and IQ of adolescents with ADHD and
control subjects
|
|
Demographics and IQ
|
All ADHD (n = 30)
|
ADHD-HI (n = 10)
|
ADHD-I (n = 10)
|
ADHD-C (n = 10)
|
Controls (n = 60)
|
|
Age
Mean (SD)
|
14.2 (1.2)
|
14.4 (1)
|
14.1 (1.3)
|
14.1 (1.4)
|
13.8 (1)
|
|
Grade level
Median
|
5.0
|
5.0
|
4.5
|
4.0a
|
5.0
|
|
Estimated IQ
Mean (SD)
|
88.3 (11.1)
|
91.3 (7.2)
|
87.8 (15.3)
|
85.8 (9.8)
|
92.9 (10.6)
|
|
Sex
Male, female
|
16, 14
|
3, 7
|
6, 4
|
7, 3
|
21, 39
|
|
Socioeconomic status (%)
A
|
6.6
|
0.0
|
10.0
|
10.0
|
5.0
|
|
B
|
20.0
|
20.0
|
10.0
|
30.0
|
21.7
|
|
C
|
60.0
|
60.0
|
70.0
|
50.0
|
60.0
|
|
D
|
13.4
|
20.0
|
10.0
|
10.0
|
13.3
|
|
E
|
0.0
|
0.0
|
0.0
|
0.0
|
0.0
|
|
Ethnicity (%)
European descent
|
73.3
|
70.0
|
80.0
|
70.0
|
75.0
|
|
Non-European descent
|
26.7
|
30.0
|
20.0
|
30.0
|
25.0
|
|
ADHD = Attention-deficit hyperactivity disorder; ADHD-HI = Predominantly
hyperactive-impulsive; ADHD-I = Predominantly inattentive; ADHD-C = Combined.
aP < 0.05 – versus control subjects
|
Discussion
In our sample, adolescents with ADHD-HI did not show significant differences
in any neuropsychological measure, compared with the control subjects.
Because significant differences were found when the 2 other groups (ADHD-I
and ADHD-C) and the control subjects were compared, neuropsychological
impairment seems to occur only in those ADHD subtypes wherein inattention
is clinically significant. Moreover, adolescents with ADHD-I presented
significant impairments on a neuropsicological test that assesses selective
attention, and adolescents with ADHD-C performed worst on a more broad
measure of executive function.
Our findings on neuropsychological impairment according to ADHD subtype
agree with other recent reports (7,18). Assessing cognitive performance
in children and adolescents with ADHD, Faraone and others found significant
differences for the combined and inattentive subtypes, compared with control
subjects (18). However, they were not able to find significant differences
among the 3 subtypes. In contrast, when Gadow and others compared a sample
of nonreferred adolescents, they were not able to find significant differences
among the neuropsychological performance of DSM-IV ADHD subtypes and control
subjects (32). In this study, however, subjects were allocated to ADHD
subtype groups according to DSM-IV ADHD criterion A only; that is, the
list of symptoms. Impairment and pervasiveness of symptoms criteria were
not assessed. As a result, significant differences in neuropsychological
measures might not have been detected among the groups because mild cases
might have been overrepresented in the ADHD groups.
The group with ADHD-HI did not show significant impairment in any neuropsychological
measure, compared with the 2 other subtype groups and the control subjects.
Thus, the findings from this exploratory study suggesting neuropsychological
deficits only in ADHD subtypes where inattention is significantly present
concur with studies that have demonstrated more academic impairments in
subjects with ADHD-I and ADHD-C, but not in subjects with ADHD-HI (18,25,26,50).
It therefore seems that the hyperactive dimension of symptoms is not linked
to significant cognitive problems.
In our study, adolescents with ADHD-C showed the worst performance on the
WCST, and those with ADHD-I presented the worst performance on the ST.
Thus, it is possible to speculate that these findings support the idea
that ADHD-C is associated with a more diffuse cognitive impairment, because
the WCST is a more comprehensive test of cognitive function (38). The performance
of the ADHD-I group on the ST supports Barkley’s model for ADHD, suggesting
that inattention in this ADHD type may be associated with more specific
deficits of selective attention and that inattention may be qualitatively
different in ADHD-C, involving more broad deficits of executive functions.
Limitations
Our findings should be understood in the context of some limitations. First,
we applied several neuropsychological tests, and this could have increased
the possibility of a type I error. However, the significant differences
in neuropsychological performance on several tests among ADHD subtypes
and control subjects were consistently in the same direction. Moreover,
even when we adjusted for multiple comparisons (Bonferroni’s correction),
significant differences among groups remained on both the ST Color-Word
score and the Digit Span (data not shown). Second, the sample size of the
ADHD subtype groups was small, limiting our statistical power. Even so,
we detected significant differences among the groups. This is an exploratory
study, however, and findings need to be replicated in other samples. Third,
the group with ADHD-HI presented a higher rate of adolescent girls than
all other groups (although the difference did not reach statistical significance,
probably owing to sample sizes). The lack of significant differences in
the neuropsychological profile between that group and the control subjects
could be attributed to a confounding effect of sex (for example, adolescent
girls might present lower impairment than adolescent boys on neuropsychological
tests). However, a previous investigation did not find significant differences
in neuropsychological functioning between boys and girls with ADHD (51).
As well, we detected no effect of sex in neuropsychological tests in our
study. Finally, we did not assess comorbidities in our sample, but several
previous studies have documented that other psychiatric disorders do not
significantly interfere with neuropsychological performance in subjects
with ADHD (7,15–17,52–54).
|
Table 2 Neuropsychological performance in adolescents with ADHD and control
subjects: means (SD) and comparison statistics
|
|
Tests
|
ADHD-HI
(n = 10)
|
ADHD-I
(n = 10)
|
ADHD-C
(n = 10)
|
Control subjects
(n = 60)
|
Post hoc analyses
|
|
Wisconsin Card-Sorting Test
Total of errors
|
47.5 (18)
|
43.4 (14.6)
|
69.1 (19.7)
|
50.7 (20)
|
C > Control subjectsa,HIa,Ib
|
|
Conceptual responses
|
51.2 (19.9)
|
54.9 (17.1)
|
29.8 (19.7)
|
49 (20.8)
|
Control subjects > Ca
|
|
Perseverative responses
|
23.5 (7.5)
|
21.5 (6.7)
|
25.8 (9)
|
25 (11.1)
|
|
|
Categories
|
4.5 (2.3)
|
3.9 (2.7)
|
2.2 (1.9)
|
4 (2.2)
|
|
|
Stroop
Word (error score)
|
0.2 (0.6)
|
0.8 (1.2)
|
0.0 (0)
|
0.4 (1.3)
|
|
|
Word (time taken to complete the test in seconds)
|
27.3 (5)
|
37.9 (32)
|
30.2 (10.7)
|
29.3 (8.3)
|
|
|
Color-Word (error score)
|
3.4 (4.1)
|
4.3 (2.9)
|
2 (2.5)
|
1.6 (2.3)
|
I > Control subjectsb
|
|
Color-Word (time taken to complete the test in seconds)
|
70.5 (14.7)
|
91.9 (36)
|
69.6 (14.2)
|
68.6 (13.3)
|
I > HIa,Cb, and
Control subjectsb
|
|
Digit Span
|
5.9 (2.5)
|
4.7 (1.9)
|
3.3 (0.7)
|
6.3 (2.2)
|
Control subjects > Ia,Cc; HI > Cb
|
|
Word Span
|
5.8 (1.3)
|
5.6 (1.3)
|
5.1 (1.1)
|
5.6 (1.1)
|
|
|
ADHD = Attention-deficit hyperactivity disorder; ADHD-HI = Predominantly
hyperactive-impulsive; ADHD-I = Predominantly inattentive; ADHD-C = Combined.
aP < 0 .05; bP < 0 .01; cP < 0 .001
|
The unmedicated subjects with ADHD-I and ADHD-C presented higher impairments
than the control subjects in a sample of adolescents aged 12 to 16 years,
drawn from a diverse culture. In addition, ADHD-HI seems not to have significant
cognitive deficits. Thus, our findings support the validity of the nosological
distinction among ADHD subtypes proposed by DSM-IV and also provide cross-cultural
support for this distinction. More studies evaluating the neuropsychological
performance in different ADHD subtypes from different age ranges and cultures
are needed, as are investigations to evaluate neuroimaging, genetics, and
treatment findings according to ADHD neuropsychological subtypes.
1 | 2 | 3 | 4
|
