Department of Physical Medicine & Rehabilitation
Examining Preseason ImPACT® Scores in Adolescent Girls with Attention-Deficit Hyperactivity Disorder
Grant L. Iverson, Ph.D.
Department of Physical Medicine and Rehabilitation, Harvard Medical School; &
Red Sox Foundation and Massachusetts General Hospital Home Base Program
Joseph E. Atkins, Ph.D.
Department of Psychology, Colby College; & Maine Concussion Management Initiative
Ross Zafonte, D.O.
Department of Physical Medicine and Rehabilitation, Harvard Medical School &
Red Sox Foundation and Massachusetts General Hospital Home Base Program
Paul D. Berkner, D.O.
Health Services, Colby College; & Maine Concussion Management Initiative
Presented at the 42nd Annual Meeting of the International Neuropsychological Society, February 12-15, 2014, Seattle, WA, USA
Objective: Baseline neurocognitive testing is used in amateur sports as a component of a comprehensive concussion management
program. Baseline testing might be more important for athletes who have a developmental condition, such as attention-deficit
hyperactivity disorder (ADHD). The purpose of this study was to determine whether adolescent girls with ADHD perform more
poorly on ImPACT® than girls who do not have ADHD.
Participants and Methods: A large de-identified database of student athletes from Maine was used for this study. The initial sample
consisted of 2,618 girls between the ages of 13 and 18, who (a) spoke English as their primary language, (b) denied a history of
neurological problems (i.e., epilepsy, meningitis, and brain tumors), and (c) obtained valid ImPACT® test scores during preseason
testing in 2010. Of this sample, 100 (3.8%) girls self-reported a diagnosis of ADHD. Through sorting and visual inspection, a sample
of 100 control subjects was precisely matched on age, sport, and number of self-reported past concussions.
Results: The two groups were compared on the four primary composite scores using mulivariate analysis of variance (MANOVA)
followed by univariate ANOVAs. Box’s M Test was not significant (p = .208), indicating that the covariance matrices did not
significantly differ across the dependent variables, and Levene’s tests for equality of error variances were nonsignificant for all four
variables. There was a significant multivariate effect [Wilks’ Lambda=.91; F(4, 195)=5.1, p<.001, partial eta squared=.095]. The
univariate ANOVA results revealed significantly worse neuropsychological test scores for girls with ADHD on the Visual Memory
(p<.031; Cohen’s d=.31) and Processing Speed (p<.001, d=.57) composites. The groups did not differ on the Verbal Memory or
Reaction Time composites. Mann Whitney U tests were used for the Impulse Control composite and the Post-Concussion scale due to
non-normal distributions and heterogeneity of the variances. The girls with ADHD had worse scores on the Impulse Control
composite (p<.033, d=.44) and they endorsed more symptoms (p<.001, d=.44) than the girls who did not have ADHD.
Table 1. Descriptive statistics and effect sizes for the ImPACT composite scores.
Composite Score
Group
M
Verbal Memory
Control
84.7
ADHD
83.4
Visual Memory
Control
71.4
ADHD
67.6
Processing Speed
Control
37.7
ADHD
33.9
Reaction Time
Control
.60
ADHD
.62
Impulse Control
Control
6.5
ADHD
8.7
Total Symptoms
Control
7.4
ADHD
13.3
SD
9.6
9.3
11.8
12.7
6.9
6.4
.09
.08
4.2
5.7
11.6
15.1
p
.344
Cohen’s d
.14
.031
.31
.001
.57
.16
.02
.033
.44
.001
.44
Note: P-values are from univariate ANOVAs for the first four scores and Mann Whitney U Tests for the Impulse Control composite and the Post-Concussion Scale
(total symptoms). By convention, effect sizes are interpreted as follows: .2 = small, .5 = medium, and .8 = large. A modified Bonferonni procedure would set the
experimentwise alpha at .008 (.05/6).
Conclusions: Girls with ADHD performed more poorly on computerized cognitive testing with ImPACT®, and they reported more
subjective symptoms, than girls who do not have ADHD. These results support the recommendation of baseline testing for athletes
with ADHD.