Neuropsychological functioning in children with ADHD: Symptom persistence is linked to poorer performance on measures of executive and nonexecutive function†
- †This research was funded by the New Zealand Neurological Foundation. Thomas Robinson was in receipt of a University of Otago Postgraduate Scholarship.
Correspondence concerning this article should be sent to: Gail Tripp, Human Developmental Neurobiology Unit, Okinawa Institute of Science and Technology Graduate University, Onna, Onna-son, Kunigami-gun, Okinawa 904-0411, Japan. (E-mail: firstname.lastname@example.org)
The present study compared the current intellectual and neuropsychological functioning of 55 children diagnosed with attention deficit hyperactivity disorder (ADHD group) 4 years earlier with that of an age- and sex-matched control sample. The children in the ADHD group performed less well than the control group on measures of intellectual function, design fluency, spatial organization, and visual memory. Those children who continued to meet DSM-IV criteria for ADHD (persistent ADHD, n = 32) evidenced greater impairment than those showing some symptom remission (ADHD in partial remission, n = 23). These data confirm the presence of neuropsychological deficits in late childhood/early adolescence among those previously diagnosed with ADHD. The data also suggest that greater cognitive impairment is a feature of persistent ADHD.
Attention deficit hyperactivity disorder (ADHD) is a persistent and debilitating neuropsychiatric disorder with an estimated worldwide prevalence of 5% (Polanczyk, de Lima, Horta, Biederman, & Rohde, 2007). The disorder is characterized by developmentally inappropriate levels of inattention, hyperactivity, and impulsivity, which cause impairment in functioning (American Psychiatric Association, 2000).
The underlying cause of ADHD is not known. Studies of the neuropsychological functioning of children with ADHD have featured prominently in efforts to understand the pathology of the disorder. Recognition of similarities between the behavior of monkeys and human adults with frontal-lobe lesions and symptoms of ADHD led researchers to question a possible role for frontal lobe functioning in the etiology of ADHD (Mattes, 1980; Zametkin & Rapoport, 1986). Subsequently, many studies have focused on assessing executive functions (e.g., set shifting, response inhibition, planning, working memory) believed to be mediated by neural circuits in the prefrontal cortex and striatum (Halperin & Schulz, 2006). The results of these studies demonstrate that, as a group, children with ADHD perform less well on tests of executive functions (Castellanos, Sonuga-Barke, Milham, & Tannock, 2006; Halperin & Schulz, 2006; Nigg, Willcutt, Doyle, & Sonuga-Barke, 2005; Seidman, 2006; Sonuga-Barke, Sergeant, Nigg, & Willcutt, 2008; Willcutt, Doyle, Nigg, Faraone, & Pennington, 2005). However, deficits are not observed in all studies, for all measures, or seen in all children with ADHD, nor are they unique to the disorder. Currently, executive function deficits (EFD) are regarded as an important aspect of ADHD, but not sufficient to explain all cases of the disorder. Nonexecutive cognitive functions (e.g., spatial span, spatial recognition memory, delayed matching to sample, pattern recognition, vocabulary) have also been found to differentiate children with and without ADHD (Kempton, Vance, Maruff, Luk, Costin, & Pantelis, 1999; Rhodes, Coghill, & Matthews, 2004; Rucklidge & Tannock, 2001, also see the meta-analysis by Frazier, Demaree, & Youngstrom, 2004).
The behavioral deficits associated with ADHD improve over time, with many children showing a reduction in symptoms during late childhood and adolescence (Biederman, Mick, & Faraone, 2000; Lahey, Pelham, Loney, Lee, & Willcutt, 2005). Lahey et al. (2005) reported that almost one-quarter (22%) of children in their longitudinal study no longer met the criteria for ADHD 7–8 years after the initial diagnosis. Biederman et al. (2000) described a syndromatic remission rate of 60% by age 20 years, but cautioned that the majority of their participants continued to struggle with a substantial number of symptoms, together with high levels of dysfunction.
Changes in neuropsychological functioning over time and their relation to symptom persistence and remission have received limited attention. Most studies are cross-sectional, only including participants with a current diagnosis of ADHD. The majority of these studies have been conducted with elementary-school-age children. In his review of neuropsychological functioning in ADHD across the lifespan Seidman (2006) concluded that while group data indicate executive dysfunctions are correlated with ADHD irrespective of age, longitudinal research is needed to assess the extent of neuropsychological continuity.
A small number of longitudinal studies have evaluated neuropsychological deficits in adolescents and young adults 5–14 years after they were diagnosed with ADHD (Biederman, Petty, Fried, Doyle, Seidman, Gross, Poetzl, & Faraone, 2007; Fischer, Barkley, Edelbrock, & Smallish, 1990; Fischer, Barkley, Smallish, & Fletcher, 2005; Halperin, Trampush, Miller, Marks, & Newcorn, 2008; Hinshaw, Carte, Fan, Jassy, & Owens, 2007; Hopkins, Perlman, Hechtman, & Weiss, 1979; Miller, Ho, & Hinshaw, 2012). In the earliest of these, Hopkins et al. (1979) reported that young adults in their study, diagnosed as hyperactive in childhood, showed deficits on measures of impulsivity and set shifting. Eight years after diagnosis, the adolescents in the study by Fisher et al. (1990) demonstrated deficits in academic achievement, sustained attention, and impulse control, but not on measures of executive function. Five years later, deficits in attention and inhibition were identified in the Hyperactive group, arising largely from the performance of participants who met the criteria for ADHD in adulthood (Fischer et al. 2005). Biederman et al. (2007) tested boys with ADHD on several common measures of executive function. Those with at least two abnormal test scores (i.e., 1.5 standard deviations from controls) were classified as having an EFD. Reassessed 7 years later, 18 of 26 participants (69%) continued to meet the criteria for an EFD. Hinshaw et al. (2007) report the findings from a prospective study of girls with ADHD. Five years after diagnosis the ADHD group showed moderate to large deficits in executive/attentional performance. Although these effects did not survive statistical control for IQ, differences between those continuing to meet the criteria for ADHD and the controls remained significant. Assessed again, 10 years after diagnosis, the girls with both persistent and remitted ADHD evidenced deficits in executive functions (Miller et al., 2012). Halperin et al. (2008) explicitly examined neuropsychological performance as a function of the persistence of ADHD. The participants with childhood ADHD differed from the controls on the Working Memory Index of the third edition of the Wechsler Adult Intelligence Scale (WAIS-III) and several continuous performance test (CPT) parameters. Separate analyses of remitters and persisters indicated there were more differences between the persisters and their control group than for those in remission.
The results of these longitudinal studies indicate impairment in cognitive/neuropsychological functioning of those diagnosed with ADHD several years earlier. There is some evidence that the degree, and possibly the nature, of the impairment is linked to the current level of symptomatology. Interpretation is complicated by the lack of overlap in the neuropsychological measures included in each study, the wide age range of the participants, and the length of the follow-up period. Additional follow-up studies are needed.
Here we compare the neuropsychological test performance of children diagnosed with DSM-IV ADHD 4 years earlier with that of an age-and sex-matched control group. Measures of general intellectual function, together with tests of executive and nonexecutive functions were included, and the data analyzed with regard to previous and current diagnostic status. Based on the extant literature, we predicted group differences, irrespective of the current diagnostic status of the ADHD group participants, on tests of intellectual and executive functioning. The results from recent longitudinal studies suggest higher levels of impairment will be observed in children who continue to meet the diagnostic criteria for ADHD.
Fifty-five children/adolescents (46 boys) diagnosed with DSM-IV ADHD 4 years earlier (ADHD group) and 55 typically developing children (control group) individually matched to members of the ADHD group for age (± 3 months), sex, and socioeconomic status participated in the current study.
The parents of all children diagnosed with ADHD at a University Research clinic between 1997 and 2001, who agreed to the researchers maintaining contact, were invited to participate with their child in the current study. Seventy-nine (76.7%) families responded to the invitation, 55 (53.4%) of whom met the inclusion criteria and agreed to participate. The baseline demographic and diagnostic characteristics of those who did and did not participate were not significantly different.
Typically developing children were recruited through letters of invitation distributed by local schools. Ninety-seven children were recruited, 55 of whom were individually matched to the ADHD participants. The diagnostic and demographic characteristics of these two groups are presented in Table 1. The mean socioeconomic status of the control group was higher than that of the ADHD group, t(54) = 2.642, p = .011.
Table 1. Demographic and diagnostic characteristics of the children in the ADHD and matched control groups
|Family socioeconomic statusa ||4.0||1.5|| ||3.4||1.2|| |
|Ethnicity|| || || || || || |
|NZ European||52||94.5|| ||53||96.4|| |
|NZ Maori||3||5.5|| ||0||0|| |
|Other||0||0|| ||2||3.6|| |
|Sex|| || || || || || |
|Male||46||83.6|| ||46||83.6|| |
|Female||9||16.4|| ||9||16.4|| |
|Initial diagnosis|| || || || || || |
|Combined||40||72.7|| || || || |
|Inattentive||10||18.2|| || || || |
|Hyperactive/impulsive||5||9.1|| || || || |
|Oppositional defiant disorder||19||34.5|| || || || |
|Conduct disorder||4||7.3|| || || || |
|Current diagnosis|| || || || || || |
|ADHD||32||58.2|| || || || |
|ADHD in partial remission||23||41.8|| || || || |
|Currently medicated||37||67.0|| || || || |
|ADHD||22||68.8|| || || || |
|ADHD in partial remission||15||65.2|| || || || |
All children included in the study had WISC-III full-scale IQ scores of at least 70 and showed no evidence of neurological disorder or psychosis. 1 At the time of testing, 37 (67%) of the children in the ADHD group were prescribed methylphenidate for the management of ADHD. For these children medication use was suspended for at least 24 h prior to neuropsychological testing.
Ethical approval for the study was obtained from the Otago Ethics Committee. All participants (parents, teachers, and children) gave written consent to take part.
The children in the ADHD group were initially diagnosed with ADHD following comprehensive multimethod multi-informant assessments that included parent, teacher, and child interviews, parent- and teacher-completed behavioral questionnaires, and an assessment of the child's cognitive functioning. All sources of information were used to determine if the children met the DSM-IV criteria for ADHD. In applying these criteria, the children were required to exhibit at least six symptoms of inattention and/or six symptoms of hyperactivity/impulsivity in at least one setting (home or school), evidence of symptoms in a second setting, and impairment from symptoms.
For the current study an abbreviated diagnostic assessment was conducted with all participating children. Information from this was used to determine the current diagnostic status of the children in the ADHD group and to assess for the presence of behavioral and/or emotional problems in the control group.
Parents and teachers were asked to complete the Disruptive Behavior Disorders (DBD) scale (Molina, Pelham, Blumenthal, & Galiszewki, 1992), to assess for the presence and severity of the DSM-IV symptoms of ADHD, oppositional defiant disorder (ODD), conduct disorder (CD), and the Child Behavior Checklist (CBCL; Achenbach, 1991a) or Teacher Report Form (TRF; Achenbach, 1991b) to screen for behavioral and emotional problems. Parents were interviewed face-to-face and teachers via the telephone about the children's current social, academic, and behavioral functioning. The children were interviewed about their friends, hobbies, and school performance.
Data from the interviews and the DBD were used to determine if the children in the ADHD group continued to meet the DSM-IV criteria for ADHD as described above (persistent ADHD). Those displaying fewer than six symptoms in at least one setting, or whose symptoms were present in only one setting were recorded as having ADHD in partial remission. Data from the interviews, the DBD, and the CBCL/TRF were used to rule out significant behavioral or emotional disturbance among the control group participants.
All sessions were scheduled in the morning and were approximately 3 h in duration. Testing was carried out by the first author, a doctoral student in clinical psychology at the time of the study.
Measures, administration, and scoring2
- Wechsler Intelligence Scale for Children Third Edition (WISC-III; Wechsler, 1991). Intellectual functioning was assessed using nine subtests from the Australian adaptation of the WISC-III. Verbal IQ (VIQ), Performance IQ (PIQ), and Full-scale IQ (FSIQ) scores were prorated according to the procedures outlined by Wechsler (1991). Freedom From Distractability (FFD) and Processing Speed (PS) Index scores were also calculated.
- Verbal Fluency Test (British Abilities Scales, Elliot, 1983). The participants were given 1 min each to name as many “things you can eat” and then “animals.” The total correct responses in the first and second 30-s interval and overall were recorded together with the number of incorrect and preservative responses (repetitions, including singular-plural) for each condition.
- Design Fluency Test (Jones-Gotman & Milner, 1977). For this test the participants were instructed to draw as many unnamable objects as they could in a 3-min period (free condition), followed by unnamable objects made up of four lines (fixed condition). The designs were scored following Jones-Gotman and Milner (1977). Five scores were derived: correct responses, perseverative responses, unacceptable designs, total errors, and total responses.
- Wisconsin Card Sorting Task (WCST; Heaton, Chelune, Talley, Kay, & Curtis, 1993). Following standard instructions, the participants were required to sort cards into different categories with limited examiner feedback. Raw scores for the number of errors, nonperseverative errors, perseverative responses, and perseverative errors were converted to T-scores, with higher T-scores indicating better performance.
- Children's Trail Making Test (TMT; Reitan, 1969). For part A, the participants were required to draw a line through small circles labeled 1–15 distributed in a random pattern on the page. Part B required them to connect alternating numbers and letters (i.e., 1, a, 2, b 3 … through to 8). The time taken to complete each part was recorded, including the time to correct mistakes. The test was discontinued after three consecutive errors, or more than 5 min to complete either part.
- Stroop Test (Golden, 1978). The participants completed three trials. In the first trial they read aloud the words “red,” “green,” or “blue” printed in black. Next they named the color of “XXX”s printed in red, green, or blue ink. Finally, they named the color in which the words “red,” “green,” and “blue” were written (with no word written in its own color). Interference T-scores were calculated using the discrepancy between the predicted and actual color-word scores.
- Rey Osterreith Complex Figure Test (ROCF; Meyers & Meyers, 1995). The participants were asked to copy the figure and draw it from memory after 3 min (immediate recall) and 30 min (delayed recall). All drawings were scored by the same doctoral-level clinical psychology student and 40% by a second student. The Spearman rank correlations for copy, immediate, and delayed recall were.98,.99, and.97, respectively. The total scores for each administration were recorded.
Data analytic plan
The performance of the ADHD group, based on the diagnostic status 4 years earlier, and their matched controls were compared on all the measures. We also compared the performance of the children who continued to meet the criteria for ADHD (persistent ADHD) with their matched controls and those who no longer met the full criteria (ADHD in partial remission) with their controls. The means and standard deviations for all groups on all measures are reported in Tables 2-5.
Table 2. Pro-rated WISC-III scores for the ADHD, persistent ADHD, and ADHD in partial remission groups and their matched controls
|WISC-III||M (SD)||M (SD)||M (SD)||M (SD)||M (SD)||M (SD)|
|Full Scale IQ||88.1 (13.8)||99.1 (13.2)||84.6 (13.3)||98.3 (11.9)||92.8 (13.2)||100.7 (15.1)|
|Verbal IQ||86.7 (13.7)||98.1 (14.2)||84.4 (13.5)||95.6 (12.6)||89.8 (13.4)||100.6 (16.5)|
|Performance IQ||91.7 (14.7)||100.1 (13.5)||87.1 (13.0)||100.9 (13.8)||98.1 (14.7)||98.3 (13.2)|
|Freedom from Distractability||87.5 (13.1)||100.3 (16.1)||85.6 (13.4)||99.2 (13.7)||90.1 (12.6)||101.8 (19.2)|
|Processing Speed||90.9 (15.0)||100.8 (12.6)||88.4 (14.4)||101.7 (13.1)||94.5 (15.4)||99.6 (12.0)|
Table 3. Verbal Fluency and Design Fluency scores for the ADHD, persistent ADHD, and ADHD in partial remission groups and their matched controls
|Verbal Fluency||N pairs = 55||N pairs = 32||N pairs = 23|
|Eat|| || || || || || |
|0–30 s||10.00 (2.99)||10.93 (3.35)||9.97 (3.10)||11.19 (3.30)||11.22 (2.50)||12.13 (3.20)|
|31–60 s||6.35 (3.77)||5.76 (2.71)||6.22 (3.62)||5.72 (2.11)||5.61 (2.73)||5.83 (3.14)|
|Total correct||16.16 (5.46)||16.69 (5.16)||16.19 (5.69)||16.91 (4.12)||16.83 (3.97)||17.96 (5.09)|
|Perseverativea||0.15 (0.56)||0.11 (0.37)||0.19 (0.64)||0.09 (0.39)||1.00 (3.74)||0.09 (0.29)|
|Incorrecta||0.07 (0.26)||0.70 (0.26)||0.06 (0.25)||0.06 (0.26)||0.13 (0.34)||0.13 (0.34)|
|Animal|| || || || || || |
|0–30 s||11.15 (2.73)||11.42 (3.14)||11.09 (2.92)||10.91 (3.05)||10.04 (2.90)||10.57 (3.48)|
|31–60 s||5.27 (2.98)||5.47 (3.35)||5.03 (3.17)||5.22 (3.52)||6.52 (4.06)||5.83 (3.42)|
|Total correct||16.09 (5.13)||16.6 (5.83)||16.13 (5.09)||15.63 (6.21)||16.13 (5.25)||16.39 (6.42)|
|Perseverativea||0.60 (2.46)||0.09 (0.29)||0.31 (0.69)||0.09 (0.30)||0.09 (0.42)||0.13 (0.34)|
|Incorrecta||0.09 (0.29)||0 (–)||0.60 (0.25)||0.06 (–)||0.09 (0.29)||0.17 (0.39)|
|Design Fluency||N pairs = 51||N pairs = 29||N pairs = 22|
|Free condition|| || || || || || |
|Novel||8.55 (5.16)||10.37 (6.18)||6.93 (3.47)||10.24 (5.17)||10.68 (6.24)||10.55 (7.43)|
|Incorrecta||0.71 (1.55)||0.22 (0.61)||0.62 (0.94)||0.24 (0.69)||0.82 (2.13)||0.18 (0.50)|
|Perseverative errors||0.59 (124)||0.22 (0.58)||0.52 (1.06)||0.28 (0.65)||0.68 (1.46)||0.14 (0.47)|
|Errors||1.29 (2.13)||0.43 (0.86)||1.14 (1.64)||0.52 (0.99)||1.50 (2.66)||0.32 (0.65)|
|Total||9.84 (5.84)||10.80 (6.34)||8.07 (4.28)||10.76 (5.48)||12.05 (6.89)||11.05 (7.79)|
|Fixed condition|| || || || || || |
|Novel||8.67 (4.33)||12.14 (5.29)||7.34 (3.92)||11.93 (5.78)||10.14 (4.23)||12.43 (4.50)|
|Incorrect||2.22 (1.76)||1.04 (1.57)||2.14 (1.58)||1.24 (1.83)||2.32 (2.01)||0.77 (1.15)|
|Perseverative errors||0.98 (1.63)||1.04 (1.90)||0.83 (1.20)||1.21 (2.19)||1.18 (2.09)||0.82 (1.44)|
|Errors||3.20 (2.47)||2.08 (2.35)||0.19 (0.23)||2.45 (2.68)||3.50 (3.10)||1.59 (1.76)|
|Total||11.86 (4.43)||14.22 (6.10)||10.31 (3.97)||14.38 (6.61)||13.91 (4.24)||14.00 (5.47)|
Table 4. WCST T-scores and Stroop interference T-scores for the ADHD, persistent ADHD, and ADHD in partial remission groups and their matched controls
|WCST||N pairs = 53||N pairs = 32||N pairs = 21|
|Errors||45.36 (13.00)||48.15 (12.40)||42.34 (12.15)||47.88 (10.74)||49.95 (13.18)||48.57 (14.79)|
|Perseverative responses||46.72 (11.76)||51.47 (12.84)||43.00 (10.78)||50.34 (11.14)||52.38 (11.12)||53.19 (10.38)|
|Perseverative errors||47.81 (11.69)||50.92 (11.02)||44.00 (10.58)||50.81 (11.26)||53.62 (11.09)||51.10 (10.92)|
|Nonperseverative errors||45.43 (13.49)||45.58 (11.54)||44.00 (13.74)||46.22 (10.02)||47.62 (13.14)||44.62 (13.76)|
|Stroop||N pairs = 45||N pairs = 28||N pairs = 17|
|Interference||31.61 (7.84)||30.00 (7.14)||32.29 (7.70)||30.10 (7.66)||31.18 (7.96)||30.41 (6.68)|
Table 5. TMT and ROCF test scores for the ADHD, persistent ADHD, and ADHD in partial remission groups and their matched controls
|TMT||N pairs = 51||N pairs = 29||N pairs = 22|
|Part A||16.69 (5.67)||18.69 (7.05)|| || || || |
|Z-score||−0.23 (0.82)||0.14 (1.09)||−0.11 (0.78)||0.32 (−0.34)||−0.40 (0.85)||−0.10 (0.87)|
| ||N pairs = 50||N pairs = 29||N pairs = 21|
|Part B||38.87 (21.90)||35.53 (15.05)|| || || || |
|Z-score||−0.11 (1.25)||−0.35 (0.96)||−0.04 (1.35)||−0.34 (1.07)||−0.20 (1.15)||−0.37 (0.84)|
|ROCF||N pairs = 53||N pairs = 31||N pairs = 22|
|Copy||18.92 (8.82)||24.65 (6.71)||19.16 (8.68)||24.95 (6.82)||18.57 (9.21)||24.23 (6.68)|
|Immediate||10.58 (7.10)||14.71 (5.28)||10.30 (6.40)||15.30 (5.40)||11.00 (8.10)||13.90 (5.20)|
|Delay||10.25 (6.67)||13.33 (5.09)||9.66 (5.61)||12.85 (4.71)||11.10 (8.00)||14.00 (5.60)|
As the children with ADHD and their controls were individually matched, paired analyses were undertaken. Prior to carrying out the analyses the data were checked to ensure that the mean difference between the variables was normally distributed. If necessary the data were transformed. When data transformation was not successful the data were analyzed using nonparametric statistics. Paired-sample t-tests and the nonparametric two-tailed Wilcoxon signed-rank test were used for these group comparisons. Where missing data occurred, list-wise deletion of the participant's data and that of their matched pair was undertaken. The Hochberg method was applied to control for multiple comparisons (Manly, Nettleton, & Hwang, 2004). 3
IQ was not treated as a covariate in these analyses. As IQ and ADHD are known to correlate, it is argued that controlling for IQ may remove variance due to the disorder (Frazier et al., 2004; Miller & Chapman, 2001; Willcutt et al., 2005). However, in view of the IQ discrepancy between the ADHD and control groups and the absence of a clear consensus in the literature on whether IQ should be controlled, we conducted secondary analyses to assess the possible contribution of IQ to any observed group differences. For these analyses matching was broken, the analyses re-run with age-corrected scores to confirm group differences, and an ANCOVA performed with pro-rated Full Scale IQ as the covariate. 4
Where differences in the pattern of results for the persistent ADHD and ADHD in partial remission groups were found, the age-corrected scores for the two groups were compared directly using independent t-tests. Alpha was maintained at p = .05 for all secondary analyses.
Summary scores for the ADHD groups and their matched controls on the WISC-III are presented in Table 2. The children in the ADHD group obtained significantly lower scores than the controls on the FSIQ, t(54) = −5.541, p < .001, the VIQ, t(54) = −5.142, p < .001, the PIQ, t(54) = −3.424, p = .001, the FFD, t(54) = −5.410, p < .001, and the PS, t(54) = −4.448, p < .001. Comparison of the performance of the persistent ADHD group with their matched controls produced similar results, with p < .001 for all comparisons. After controlling for multiple comparisons, the ADHD in partial remission group did not differ from their controls, although p < .05 for the FSIQ, the VIQ, and the FFD index.
The scores of the ADHD and control groups on the verbal and nonverbal fluency tasks are presented in Table 3. For the Verbal Fluency test, no significant differences were found for any of the group comparisons, and the effect sizes were small for all parametric analyses. On the Free condition of the Design Fluency task, no group differences remained significant after controlling for multiple comparisons. In the Fixed condition the ADHD group produced fewer novel responses, t(50) = −3.827, p < .001, and more incorrect responses, t(50) = 4.328, p < .001. The persistent ADHD group generated significantly fewer novel and total responses than their matched controls, while the ADHD in partial remission group made significantly more errors than their matched controls.
There were no significant group differences on any of the WCST variables (see Table 4) for the full ADHD sample or for the ADHD in partial remission group after taking account of multiple comparisons. In contrast, the children who continued to meet the criteria for ADHD obtained significantly lower T-scores than their matched controls for Errors, Perseverative Responses, and Perseverative Errors, indicating their poorer performance on these measures. The interference T-scores from the Stroop task (Table 4) were not significantly different for any of the ADHD group comparisons.
The findings from the TMT test and the ROCF are presented in Table 5. The children in the ADHD group completed part A of the TMT more quickly than those in the Control group; however, this effect was no longer significant after controlling for multiple comparisons. The ADHD and control groups did not differ on the time taken to complete part B of the test. Neither of the ADHD subgroups differed from their respective control groups for the time to complete either part of the TMT.
On the ROCF test, the children in the control group copied the initial design with higher fidelity than those in the ADHD group, t(52) = −4.514, p < .001, and recalled more details in both the immediate, t(52) = −4.104, p < .001, and delayed recall conditions, t(52) = −2.797, p = .007. A similar pattern of significance was seen for the children with persistent ADHD and their controls. The children who no longer met the criteria for ADHD differed from their matched controls on the copy condition only.
The ADHD and control group scores on the fixed version of the Design Fluency test (novel responses and errors) and the ROCF were submitted to an ANCOVA, with prorated FSIQ as the covariate. Group comparisons remained significant for the number of errors during the fixed condition of the Design Fluency task, F(1, 101) = 6.935, p = .016, and for the Immediate, F(1, 105) = 4.936, p = .028, and Delayed, F(1, 105) = 7.231, p = .008, recall conditions of the ROCF.
Differences in the pattern of results for the two ADHD subgroups (persistent ADHD and ADHD in partial remission) were found for the WISC-III, Design Fluency (fixed condition), the WCST, and the ROCF. The performance of the two ADHD groups was directly compared on these measures using age-corrected scores. Significant group differences were found for the WISC-III FSIQ, t(53) = −2.249, p = .029, and PIQ, t(53) = −2.949, p = .005, Design Fluency novel responses, t(50) = −2.076, p = .043, WCST perseverative responses, t(51) = −3.061, p = .004, errors, t(51) = −2.156, p = .036, and perseverative errors, t(51) = −3.176, p = .003. For all of these comparisons the partial remission group performed better than the persistent ADHD group. Appendix B presents a summary of all significant group comparisons.
The current study extends previous research on the neuropsychological functioning of children with ADHD, assessing their performance 4 years after diagnosis against a carefully matched control sample. The study design allowed the performance of children with subthreshold and persistent ADHD to be compared with that of their typically developing peers and with one another, providing valuable information on the developmental course of the disorder.
Previous studies detailing deficits in cognitive functioning in children, adolescents, and adults with ADHD suggest that such deficits are a persistent feature of the disorder (Seidman, 2006). The results of the current study confirm the presence of neuropsychological deficits in older children previously diagnosed with ADHD. When tested 4 years after diagnosis, the ADHD group obtained significantly lower scores than their matched controls on measures of intellectual functioning, verbal working memory, and attention (WISC-III FFD), nonverbal fluency (Design Fluency), and visuo-constructional abilities and visual memory (ROCF). The differences on these latter tests remained significant after controlling for IQ. No significant differences emerged on tests assessing verbal fluency, response inhibition (Stroop), inhibitory control (TMT Part B), or set shifting and planning (WCST). The presence of group differences on only some measures of neuropsychological functioning seems to be a consistent finding in the literature.
A review of the pattern of findings for the children with persistent ADHD and those who no longer met the full criteria is informative. On the WISC-III, the Design Fluency test, the WCST, and the ROCF, the performance of children meeting criteria for ADHD was significantly lower than that of their matched controls. The partial remission group differed from their controls on fewer measures. Direct comparisons of these two ADHD groups confirmed the superior performance of the partial remission group on aspects of the WISC-III, the Design Fluency task and the WCST.
These findings suggest greater cognitive impairment in those displaying more symptoms of ADHD or pervasiveness of symptoms across settings. Similar outcomes were reported in two recent longitudinal studies (Halperin et al., 2008; Hinshaw et al., 2007). Reports from both studies described greater impairment in participants in their longitudinal studies who continued to meet the criteria for ADHD. Together with these two studies, our results suggest greater cognitive impairment is a feature of persistent ADHD. The causal direction in this association is not entirely clear.
Halperin and Schulz (2006) have proposed that ADHD results from noncortical neural dysfunction that is static across the lifespan. They attribute symptom reduction over the course of development to compensatory prefrontally mediated executive functions (Halperin et al., 2008; Halperin and Schulz 2006). Based on their hypothesis, the children in the current study who showed remission of symptoms should have displayed fewer deficits on the measures of executive function. Our findings were consistent with this prediction. Compared to the persistent ADHD group, those in partial remission produced more novel designs on the Design Fluency test and made fewer errors, perseverative responses, and perseverative errors on the WCST. A similar pattern was observed when these two groups were compared to their matched controls.
In previous studies, the WCST has not demonstrated as strong a relationship to ADHD as some other measures of executive function (Willcutt et al., 2005). The current findings suggest that set-shifting and planning may be particularly impaired in those displaying persistent ADHD. Relatively few studies of ADHD report using the Design Fluency task. In this study, participants with ADHD, particularly those for whom the disorder is persistent, experienced difficulty in the production of novel designs rather than with perseverative responding, possibly reflecting difficulties with creativity as opposed to self-regulation. In contrast the ROCF has been more widely used in studies of ADHD, albeit with different scoring criteria. The children who had been previously diagnosed with ADHD performed less well than the typically developing children on both the copy and recall aspects of the task, with little separating those who did and did not currently meet the criteria for the disorder.
The number and range of neuropsychological tests administered is an important strength of the current study. The participating children/adolescents completed a full morning of testing that allowed assessment of general intellectual functioning, academic achievement (not reported here), and executive and nonexecutive functioning. Many of the tests included have been used extensively with ADHD populations, allowing comparison with other studies. This breadth of testing further confirmed the variability in findings across measures and demonstrated group differences on tests of both executive and nonexecutive function.
The age of the children in the current study addresses an important gap in the literature. Most studies assessing neuropsychological functioning in ADHD have been conducted with elementary-age children (6–12 years). In his review Seidman (2006) noted at least 100 such studies since the early 1970s. At the time they were followed up, the children in our study ranged in age from 10–14 years.
The ADHD sample in the current study is relatively small. This is offset somewhat by the inclusion of a well-matched control group and the use of paired analyses. A review of effect sizes suggests nonsignificant group comparisons were not due to low power. The participants in the ADHD group were representative of those seen in clinical practice with respect to sex, ADHD subtype, the presence of comorbid externalizing disorders, as well as the proportion no longer meeting the full criteria for ADHD. These characteristics increase the generalizability of the findings to the wider population of children with ADHD.
In splitting the ADHD group into those who continued to meet diagnostic criteria and those who did not, our partial remission group is heterogeneous. As a consequence, the two ADHD groups may be more similar than those in studies setting out to compare persisters and remitters (Halperin et al., 2008; Miller et al., 2012). Given the breath of our partial remission group, the extent of the observed differences lends further weight to the conclusion that persistent ADHD is associated with impaired neuropsychological functioning. The manner in which we formed our ADHD subgroups and the smaller sample size do not allow us to comment on the neuropsychological status of those who no longer display symptoms of ADHD.
To control for the large number of analyses undertaken, we applied conservative alpha levels. While we can be confident in the results presented, group differences may be underestimated. This seems most likely with the Design Fluency task and the WISC-III. It is likely that the participants with ADHD in partial remission do have poorer verbal comprehension scores than their matched controls.
The participants in the ADHD group completed several of the neuropsychological measures when first assessed 4 years earlier, whereas the control group had no previous exposure. The long interval between the assessments decreases the likelihood of practice effects in the ADHD group; however, such effects cannot be ruled out. As a result, the differences in the performance of the ADHD and control groups on the measures previously administered to the ADHD group may be reduced. Coupled with the correction for multiple comparisons, there is a risk of type II errors.
The children who were prescribed methylphenidate for symptom management were withdrawn from medication at least 24 h prior to testing. Although this practice is common and ensures no active drug effects during testing, it does not address receptor changes resulting from medication use. In the current study, the proportion of participants in the two ADHD subgroups who were prescribed methylphenidate was very similar.
The cross-sectional nature of the current study does not allow us to comment on the neuropsychological trajectories of the participants who were diagnosed with ADHD. We can conservatively say, as a group, that these children demonstrate poorer performance than controls on a measure of general intellectual functioning and a limited number of measures assessing both executive and nonexecutive functions. Four years after diagnosis, the group who no longer met the full criteria for ADHD were less impaired than those who still met the criteria. These findings contribute to the limited research on the developmental course of neuropsychological deficits in ADHD.
Three children in the ADHD group had full scale IQs greater than 70 at the first, but not the second, assessment.
Participants in the ADHD group completed tests 1–5 at the time of the initial assessment and again at follow up. The constructs measured by these tests are summarized in Appendix A.
Contrasts were grouped into families on the basis of their properties of interest. The contrasts within each family were ranked in descending order by their p-values and tested against a progressively increasing alpha level (p/n, where n = the contrast number).
For the Design Fluency and ROCF tests age-corrected Z-scores were calculated using the data from the 97 typically developing children recruited to the study.
Appendix: Appendix A
Functional domains assessed by the neuropsychological tests.
|Intellectual Functioning|| |
|Wechsler Intelligence Scale for Children-III (Wechsler, 1991)||General intellectual functioning|
|Executive Functioning|| |
|Verbal Fluency (semantic) (British Abilities Scales; Elliot, 1983)||Factor analytic studies suggest attentional control/working memory play an important role in performance (Strauss, Sherman, & Spreen, 2006).|
|Design Fluency (Jones-Gotman & Milner, 1977)||Thought to involve creativity, constructional skills, and working memory abilities (Strauss et al., 2006). Also described as involving the executive function subdomains of initiation, shift, self-regulation and self-monitoring (Baron, 2004).|
|Wisconsin Card Sorting Task (Heaton et al., 1993)||Designed to assess the ability to form abstract concepts, shift and maintain set and utilize feedback. Test requires planning, organized searching and the ability to control impulsive responding (Strauss et al., 2006).|
|Stroop (Golden, 1978)||Measure of cognitive control: goal maintenance, attention and response inhibition (Strauss et al., 2006).|
|Attention/Executive Functioning|| |
|Trail Making Test: Children's Version (Reitan, 1969)||Measure of attention, speed and mental flexibility. Part B described by some researchers as a measure of executive function. Thought to be sensitive to the executive function domains of shift and sustain and inhibitory control (Baron, 2004).|
|Rey Osterrieth Complex Figure (Meyers & Meyers, 1995)||Measure of visual-spatial constructional ability, spatial organization, and visual memory (Mitrushina, Boone, Razani, & D'Elia, 2005; Strauss et al., 2006). The Copy condition is considered as an executive function measure when scored according to the Waber and Holmes (1985) Developmental Scoring System (Miller & Hinshaw, 2010). This scoring system was not applied in the current study.|
Appendix: Appendix B
Summary of significant group comparisons
|Wechsler Intelligence Scale for Children-III|| || || || |
|Freedom from distractibility||✓||✓||✗||✗|
|Design Fluency|| || || || |
|Fixed|| || || || |
|Wisconsin Card Sorting Task|| || || || |
|Non perseverative errors||✗||✗||✗||–|
|Rey Osterrieth Complex Figure|| || || || |