Detecting epilepsy-related cognitive problems in clinically referred children with epilepsy: Is the WISC-IV a useful tool?

Authors

  • Elisabeth M. S. Sherman,

    1. Alberta Health Services and the Alberta Children’s Hospital Research Institute, Calgary, Alberta, Canada
    2. Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada
    3. Copeman Healthcare Centre, Calary, Alberta, Canada
    Search for more papers by this author
  • Brian L. Brooks,

    1. Alberta Health Services and the Alberta Children’s Hospital Research Institute, Calgary, Alberta, Canada
    2. Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada
    Search for more papers by this author
  • Taryn B. Fay-McClymont,

    1. Alberta Health Services and the Alberta Children’s Hospital Research Institute, Calgary, Alberta, Canada
    2. Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada
    Search for more papers by this author
  • William S. MacAllister

    1. Comprehensive Epilepsy Center, New York University, New York, New York, U.S.A.
    Search for more papers by this author

Address correspondence to Elisabeth M. S. Sherman, Ph.D., Neurosciences Program, Alberta Children’s Hospital, 2888 Shaganappi Trail NW, Calgary, T3B 6A8 AB Canada. E-mail: elisabeth.sherman@albertahealthservices.ca

Summary

Purpose:  The Wechsler Intelligence Scale for Children – Fourth Edition is the most widely used intelligence quotient (IQ) test in use today. However, despite numerous studies on IQ in childhood epilepsy, data exist almost exclusively from prior editions of the test, and no studies to date provide information on the sensitivity of specific WISC-IV scores (full-scale IQ [FSIQ], index, and subtest scores) to epilepsy-related cognitive impairments. The goal of this study was to determine the relative sensitivity of WISC-IV index and subscale scores in detecting cognitive problems in a group of clinically referred children with epilepsy compared to matched controls, and to define the relationship among WISC-IV scales, demographic factors, and epilepsy-related variables.

Methods:  WISC-IV data for children with epilepsy and high seizure burden were obtained from the Alberta Children’s Hospital (ACH) and the New York University Comprehensive Epilepsy Center (NYU), two tertiary care medical centers for pediatric epilepsy. All children were clinically referred and received a standard assessment including WISC-IV. Matched controls were obtained from the WISC-IV Canadian and American standardization samples.

Key Findings:  WISC-IV scores from 212 children were included: 106 children with epilepsy (46 girls, 60 boys; mean age 11.0 years, standard deviation [SD] 3.1; parental education 14.5 years, SD 2.8), and 106 controls matched for age, gender, ethnicity, and parental education. Of the children with epilepsy, 44 had a clearly lateralized focus on electroencephalography (EEG) involving either the right or left hemisphere (26 left, 18 right). FSIQ for the epilepsy group was significantly lower than for controls, and 36.8% of children had IQs compatible with intellectual disability (FSIQ < 70), versus <1% of controls. In children with epilepsy, Working Memory and Processing Speed Index scores were lower than those for Verbal Comprehension and Perceptual Reasoning (p < 0.01). At the subtest level, scores for children with epilepsy were highest on visual and verbal subtests measuring reasoning skills such as Matrix Reasoning, and lowest on Coding (mean 5.93, SD 3.6). In terms of percentage of children on each subtest with low scores (i.e., scores below 2 SDs from the expected normative mean of 10), the Coding subtest identified the most children (28.3%) with low scores, and the Similarities subtest identified the fewest (16%). Later age at onset and shorter epilepsy duration were both correlated with higher WISC-IV FSIQ and index scores (r correlation coefficient values ranging from 0.36 to 0.44, p < 0.0001), and number of current and previous antiepileptic drug trials were both inversely correlated with FSIQ and index scores (r −0.27 to −0.47, all p-values < 0.01). Neither the FSIQ nor the index scores were significantly related to seizure frequency. A similar pattern was found for subtest scores. No differences in FSIQ, index scores, or subtest scores were found between children with left- and right-hemisphere seizure foci, or between those with positive or negative magnetic resonance imaging (MRI) findings.

Significance:  The WISC-IV is sensitive to epilepsy-related cognitive problems in clinically referred children with high seizure burden, particularly problems relating to expressive verbal, working memory, and processing speed difficulties. Compared to healthy children, these children have a very high rate of cognitive difficulties as assessed by the WISC-IV. The usefulness of the WISC-IV in detecting cognitive deficits in children with milder forms of epilepsy remains to be determined.

Increasingly, cognitive deficits are considered a common feature of childhood epilepsy. Children with epilepsy are more likely to have cognitive deficits than healthy children, with approximately one in four children with epilepsy demonstrating subnormal cognitive function (Berg et al., 2008). Cognition may be affected adversely by several factors including antiepileptic drugs (AEDs), drug-resistance, younger age at onset, longer epilepsy duration, and seizure frequency (Cormack et al., 2007; Seidenberg et al., 2007; Berg et al., 2008). However, cognitive problems may also predate first seizure (Fastenau et al., 2009), which suggests that in some children, cognitive problems are a primary symptom of the abnormal brain circuitry giving rise to epilepsy, rather than a secondary effect attributable to the effects of seizures or drugs.

Intelligence tests are considered the first-line tool for assessing cognitive problems in children. Results from intelligence quotient (IQ) tests help guide diagnosis, treatment, and educational planning, and are used by clinicians and researchers alike. Consequently, validation studies are critical for supporting the use of specific IQ tests in clinical and research settings. Despite this, validation work in epilepsy on the most commonly used measure of intelligence in children, the Wechsler Intelligence Scale for Children – Fourth Edition (WISC-IV, Wechsler, 2003a) is lacking. The WISC-IV is arguably the gold standard for intelligence testing in clinical evaluations, and it is the recommended IQ test for children as part of the National Institutes of Health (NIH) National Institute of Neurological Disorders and Stroke (NINDS) Common Data Elements for Epilepsy (http://www.commondataelements.ninds.nih.gov). Although there are numerous studies on IQ in childhood epilepsy (Blackburn et al., 2007; Cormack et al., 2007; Guimaraes et al., 2007; Berg et al., 2008; van Mil et al., 2008), these studies used the previous editions of the test or combined data across several prior test editions. No studies have provided information on the sensitivity of specific WISC-IV scores (full-scale IQ [FSIQ], index, and subtest scores) to epilepsy-related cognitive impairments, or on the relationship between WISC-IV scores and epilepsy variables. In one study on the Wechsler Intelligence Scale for Children, Third Edition (WISC-III; Wechsler, 1991), a predecessor of the WISC-IV, neither epilepsy onset nor AED polytherapy were related to FSIQ. Although the statistical power was limited by a small sample size, children with epilepsy had lower scores across all index scores and subscales compared to controls (O’Leary et al., 2006). Notably, research on the WISC-III does not necessarily apply to the WISC-IV, since the latter differs from the WISC-III in both content and structure (Sattler & Dumont, 2004; Strauss et al., 2006).

Because of its widespread clinical use and its importance in epilepsy research, the goal of this study was to provide initial validation evidence for the WISC-IV in clinically referred children with epilepsy. The goals were to identify WISC-IV index and subscale strengths and weaknesses compared to a matched sample of children without epilepsy, and to define the relationship between WISC-IV scales and clinical variables, including demographic factors and epilepsy variables such as laterality of seizure focus.

Method

Retrospective data from consecutive referrals of children seen for neuropsychological assessment at the Alberta Children’s Hospital (ACH) and the New York University Comprehensive Epilepsy Center (NYU), two tertiary care medical centers for pediatric epilepsy, were included. All children were clinically referred, and all received a standardized assessment that included the WISC-IV. The majority of patients were seen for purposes of preoperative decision making or routine postoperative follow-up, whereas the remainder of the sample included referrals made to clarify learning or behavioral issues. Data were accessed after approval by hospital and university review boards at both sites. Inclusion criteria were the following: (1) epilepsy diagnosis by a pediatric neurologist or epileptologist, (2) current age between 6 and 16 years (consistent with WISC-IV normative data), and (3) availability of full WISC-IV scores, including scaled scores on all 10 core subtests and 5 index scores. Exclusion criteria included progressive neurologic or serious medical condition other than epilepsy. Control data were obtained from the WISC-IV Canadian and American standardization samples, matched for age, gender, ethnicity, parental education, and country of origin (United States or Canada).

Measures

The WISC-IV provides a FSIQ, as well as four composite Index scores that reflect functioning along four cognitive domains, namely, the Verbal Comprehension Index (VCI), the Perceptual Reasoning Index (PRI), the Working Memory Index (WMI), and the Processing Speed Index (PSI). The VCI reflects performance across three verbal subtests (Vocabulary, Similarities, and Comprehension), whereas the PRI is a composite index based on performance on three visual reasoning subtests (Block Design, Matrix Reasoning, and Picture Concepts). The WMI and PSI are each based on two subtests, respectively, measuring auditory working memory (WMI: Digit Span, Letter-Number Sequencing) and speed of thinking and motor speed (PSI: Coding, Symbol Search). All scaled scores and index scores were derived from raw scores based on the American and Canadian standardization sample data (Wechsler, 2003b,c).

Demographic variables included age at testing, gender, ethnicity, parental education, and handedness. Epilepsy severity variables included age at onset, duration of epilepsy, number of AEDs at time of assessment, number of previously failed AED trials (a proxy variable of epilepsy-treatment resistance), seizure frequency (number of seizures per month), laterality of seizure onset (left, right, or bilateral), and imaging findings (dichotomized as either positive or negative findings on magnetic resonance imaging (MRI).

Statistical approach

WISC-IV normative data from each respective country were used for the epilepsy and matched control sample analysis. The Canadian and American normed WISC-IV data were then combined to make a single epilepsy group (N = 106) and a single control group (N = 106) for analysis. Comparisons between the epilepsy and matched control samples were achieved using independent samples t-tests. Differences between index scores and between subtest scaled scores for the epilepsy sample were assessed with paired t-tests. As a measure of sensitivity to epilepsy-related cognitive impairments, the percentage of children with low scores on indexes and subtests was computed using chi-square analyses, with low scores defined as scores falling 2 SDs below the normative mean (i.e., a standard score of 70 for the FSIQ and index scores, and a scaled score of 4 on the subtests). Correlations were used to assess for associations between WISC-IV scores and demographic and epilepsy variables, with tests of normality (Kolmogorov-Smirnov) conducted to determine whether Pearson’s or Spearman’s r would be most appropriate. Two-tailed analyses were used where applicable. To address familywise error, an adjusted p-value of <0.01 was considered statistically significant. Analyses were conducted with SPSS version 15.0 (IBM, Armonk, NY, U.S.A.) for Windows.

Results

WISC-IV scores were available for 106 children with epilepsy. Demographic and neurologic characteristics are presented in Table 1. Overall, the sample was characterized by early onset epilepsy, multiple AEDs, and high seizure burden. There were no significant differences between sites (ACH site 57, NYU site 49) on demographic or epilepsy variables (age, parental education, age at onset, number of current AEDs, number of prior AED trials, and seizure frequency), WISC-IV FSIQ, or WISC-IV index scores.

Table 1.   Demographic and neurologic characteristics of the sample of children with epilepsy
Sample characteristics 
  1. aMRI reports available for 103 participants.

N106
Sex57% girls, 43% boys
Age11.0 (SD 3.1)
Ethnicity
 Caucasian68%
 Asian14%
 Hispanic 7%
 African American 5%
 Other 6%
Parental education
 At least a grade 12 education93%
 13 years of education or more62%
Handedness, right-handed67%
Age at epilepsy onset5.1 years (SD 4.2)
Duration of epilepsy5.9 years (SD 3.9)
Number of current AEDs1.3 (SD 0.9)
Number of prior AEDs2.2 (SD 2.3)
Seizure frequency per month18.4 (SD 47.9)
Presence of lateralized EEG focusn = 44 (26 left hemisphere, 18 right hemisphere)
MRI findingsn = 54 (50.9%) with positive MRIa
 Bilateral49
 Temporal15
 Extratemporal18
 Other21

From the WISC-IV Canadian and American standardization samples, 106 controls were selected, and were matched case-by-case with the epilepsy cases on age, gender, ethnicity, parental education, and country of origin (United States or Canada). After matching, as expected, there were no significant differences in any of these variables between epilepsy cases and controls. In addition, there were no significant differences between the American or Canadian control samples in terms of WISC-IV FSIQ or index scores. Of note, American and Canadian standardization samples differ slightly in terms of ethnic composition and parental education to reflect Census-based sociodemographic differences between the two countries, with the Canadian sample tending to be slightly more educated and ethnically homogenous than the U.S. standardization sample (see Wechsler, 2003a,b for details).

WISC-IV performance levels

The mean FSIQ for the epilepsy sample was in the borderline range (78.62, standard deviation [SD] 22.3), as was the mean for the Processing Speed Index. The remaining three index scores were in the low average range (see Table 2 for the index means). In all, 36.8% of the epilepsy group had FSIQ scores in the extremely low range (FSIQ < 70, or >2 SDs below the expected normative mean of 100) indicative of intellectual disability, compared to <1% of controls.

Table 2.   WISC-IV FSIQ and index scores
 Mean (SD)Percentage with extremely low scores (<2 SDs)
EpilepsyControlp-valueEpilepsy, %Control, %p-value
  1. SD, standard deviation; CI, confidence interval. Extremely Low scores defined as scores falling <2 SDs below the normative mean of 100 (i.e., scores of 69 or less). Independent samples t-tests were used to compare FSIQ and Index scores (d.f. 210). Chi-square analyses were used to compare percentage of participants with extremely low scores (<2 SDs).

FSIQ78.62 (22.3)101.36 (15.9)<0.00136.80.9<0.001
Verbal Comprehension Index (VCI)83.75 (20.8)101.26 (14.0)<0.00130.20.9<0.001
Perceptual Reasoning Index (PRI)84.99 (21.2)100.30 (16.6)<0.00128.32.8<0.001
Working Memory Index (WMI)80.30 (18.1)100.07 (14.6)<0.00132.11.9<0.001
Processing Speed Index (PSI)79.88 (19.6)102.22 (16.9)<0.00130.21.9<0.001

In terms of subtests, mean scaled scores for the children with epilepsy were highest on two of the three nonverbal subtests, Matrix Reasoning and Picture Concepts [means 7.97 (3.9), 7.96 (4.0), respectively; Table 3]. Mean scaled scores were lowest on two subtests measuring visual and auditory aspects of processing speed and working memory, specifically the Coding and Digit Span subtests [means 5.93 (3.6), 6.58 (3.7), respectively]. A subtest measuring processing speed and visual attention, with minimal motor component (Symbol Search) was also low [mean 6.70 (3.9)] as was a subtest measuring auditory working memory (Letter Number Sequencing) [mean 6.67 (3.2)]. Compared to controls, all WISC-IV scaled scores were significantly lower for the children with epilepsy (p < 0.0001).

Table 3.   WISC-IV subtest scaled scores
 Mean (SD)p-valuePercent with extremely low scores (<2 SDs)p-value
EpilepsyControlEpilepsy, %Control, %
  1. SD, standard deviation; CI, confidence interval. Extremely Low scores defined as scores falling <2 SDs below the normative mean of 10 (i.e., scores of 3 or less). Independent samples t-tests were used to compare subtest scores (d.f. 210). Chi-square analyses used to compare percentage of participants with extremely low scores (<2 SDs).

Matrix Reasoning7.97 (3.9)10.08 (3.1)<0.00117.00.9<0.001
Picture Concepts7.96 (4.0)10.04 (3.1)<0.00117.12.8<0.001
Similarities7.53 (3.6)10.25 (2.8)<0.00116.00.9<0.001
Vocabulary7.12 (3.5)10.41 (3.1)<0.00117.90.9<0.001
Block Design6.87 (3.6)10.06 (3.4)<0.00122.65.7<0.001
Symbol Search6.70 (3.9)10.32 (3.3)<0.00126.42.8<0.001
Letter Number6.67 (3.2)10.13 (2.9)<0.00118.91.9<0.001
Comprehension6.61 (4.0)10.07 (2.7)<0.00127.40.9<0.001
Digit Span6.58 (3.7)10.02 (3.0)<0.00125.50.9<0.001
Coding5.93 (3.6)10.39 (3.4)<0.00128.31.9<0.001

In terms of statistically significant differences between Index scores, VCI scores were significantly greater than WMI and PSI scores (p < 0.01, p = 0.01, respectively; Table 4). PRI scores were also significantly greater than WMI and PSI scores (p < 0.0001, p < 0.0001, respectively). No statistically significant differences were found between the VCI and PRI scores, or the WMI and PSI scores. Compared to the matched controls, all WISC-IV index scores were significantly lower for the children with epilepsy (p < 0.0001), and a much higher proportion of children had extremely low scores (Table 2).

Table 4.   Differences between WISC-IV index scores for the children with epilepsy
 Td.f.p-value
Verbal Comprehension Index – Perceptual Reasoning Index−1.081050.28
Verbal Comprehension Index – Working Memory Index2.931050.004
Verbal Comprehension Index – Processing Speed Index2.581050.011
Perceptual Reasoning Index – Working Memory Index3.861050.0001
Perceptual Reasoning Index – Processing Speed Index3.881050.0001
Working Memory Index – Processing Speed Index0.281050.78

The number of children with epilepsy who had low subtest scores (i.e., scores below 2 SDs from the expected normative mean of 10) was examined for each subtest (see Table 3). Three of the four subtests that had the largest percentage (25–28%) of children with extremely low scores measured aspects of processing speed and working memory (Coding, Symbol Search, and Digit Span subtests). Conversely, subtests measuring reasoning and abstraction in the verbal and visual modalities, the Similarities and Matrix Reasoning subtests, identified the fewest children (16% and 17%, respectively). Verbal subtests, considered individually, had the highest rate occurring for Comprehension (27%), and the lowest with Similarities (16%). Compared to matched controls, a much higher proportion of children with epilepsy had extremely low scores.

Correlation with epilepsy variables

In children with epilepsy, the FSIQ and all four index scores demonstrated moderate correlations with almost all epilepsy variables (Table 5). Later age at onset and shorter epilepsy duration were both correlated with higher WISC-IV FSIQ and index scores (r values ranging from 0.36 to 0.44, p < 0.0001), and number of current and previous AED trials were both inversely correlated with FSIQ and index scores (r values −0.27 to −0.47, all ps-values < 0.01). Neither the FSIQ nor the index scores were significantly related to seizure frequency.

Table 5.   Correlations between WISC-IV and epilepsy variables
 Age at onsetDurationNumber of AEDsNumber of failed AEDsSeizure frequency
  1. Correlations include Spearman’s rho. *p < 0.05; **p < 0.01.

Index scores     
 FSIQ0.43**−0.41**−0.34**−0.46**−0.21*
 VCI0.44**−0.41**−0.39**−0.42**−0.25*
 PRI0.44**−0.36**−0.27**−0.45**−0.18
 WMI0.36**−0.37**−0.30**−0.47**−0.18
 PSI0.36**−0.43**−0.35**−0.37**−0.21*
Subtest scores     
 Similarities0.32**−0.29*−0.37**−0.35**−0.22*
 Vocabulary0.42**−0.39**−0.40**−0.41**−0.21*
 Comprehension0.41**−0.41**−0.40**−0.40**−0.24*
 Block Design0.38**−0.35**−0.23*−0.40**−0.18
 Picture Concepts0.38**−0.30*−0.20−0.35**−0.18
 Matrix Reasoning0.43**−0.31**−0.26**−0.44**−0.15
 Digit Span0.33**−0.31**−0.23*−0.44**−0.16
 Letter Number0.32**−0.39**−0.36**−0.41**−0.17
 Coding0.36**−0.41**−0.32**−0.37**−0.20*
 Symbol Search0.31**−0.37**−0.37**−0.36**−0.25*

A similar pattern was found for subtest scores. All subtests were significantly correlated with age at onset, with correlations ranging from 0.31 (Symbol Search, p < 0.01) to 0.43 (Matrix Reasoning, p < 0.0001), and inversely related to epilepsy duration, with correlations ranging from −0.29 (Similarities, p < 0.01) to −0.41 (Comprehension and Coding, p < 0.0001). All the verbal (Similarities, Vocabulary, and Comprehension) and processing speed subtests (Coding, Symbol Search) were inversely related to the number of current AEDs, as was the working memory subtest of Letter Number Sequencing and the nonverbal subtest of Matrix Reasoning, with significant correlations ranging from −0.26 (Matrix Reasoning, p < 0.01) to −0.40 (Vocabulary, Comprehension, p < 0.0001). Nonverbal subtests (Block Design and Picture Concepts), as well as the working memory subtest of Digit Span, had modest correlations with number of AEDs that failed to reach statistical significance. All subtest scores were inversely correlated with number of prior AED trials, with correlations ranging from −0.35 (Similarities and Picture Concepts, p < 0.0001) to −0.44 (Matrix Reasoning and Digit Span, p < 0.0001). No index or subtest scores were significantly related to seizure frequency. No differences in index scores, subtest scores, or magnitude of the VCI-PRI discrepancy were found between children with left- and right-hemisphere seizure foci, or between children with positive or negative MRIs. In addition, index scores of children with a history of epilepsy surgery (N = 16) did not differ significantly from those of the larger group.

Discussion

Despite its widespread clinical use and its presence in the clinical and research field for several years, this is the first study to examine the utility of the WISC-IV in detecting cognitive problems in children with epilepsy. The results indicate that the WISC-IV is sensitive to epilepsy-related cognitive impairments in clinically referred children with high seizure burden, with >30% of children with FSIQs in the range associated with intellectual disability. In addition, difficulties with processing speed and working memory were most prominent in children with epilepsy. Approximately one third of children had clinically significant difficulties in these domains, with scores falling two or more standard deviations below the normative mean on these tasks. It is notable that the motor component of the tasks was apparently not a salient factor—the prevalence of children with low scores was similar whether a test with high motor component (Coding) or low motor component (Digit Span, Symbol Search) was employed, suggesting that the primary difficulty for children relates to complex attention and working memory rather than motor speed per se. Verbal subtests, considered individually, had the highest detection rate occurring with Comprehension (27%), and the lowest with Similarities (16%). Considered together, however, the composite VCI score identified one third of children with clinically significant deficits. In contrast, visual-spatial deficits as measured by the WISC-IV were less common, suggesting that in most clinically referred children, visual-spatial problems do not figure prominently. Similarly, few children had difficulties with subtests that measured reasoning and abstraction, whether verbal or visual, with a very low base rate of low scores on both the Similarities and Matrix Reasoning subtests. The reason for the low base rate of children with reasoning difficulties is unknown, but could be either epilepsy-specific (e.g., reasoning skills are a relative strength in children with epilepsy compared to other cognitive domains), or test-specific in terms of how skills are measured in these subtests versus other subtests. For example, Matrix Reasoning requires only pointing, and Similarities can be accomplished by responding with single-word answers, so it is possible that these subtests allow children to perform better than on other subtests requiring more complex demands on attention, motor skills, or verbal expression.

WISC-IV scores were related to age at onset, epilepsy duration, number of AEDs, and drug resistance as measured by number of prior failed AED trials, which is important for the test’s validity, given the association between all of these markers of epilepsy severity and cognitive dysfunction in children with epilepsy (MacAllister & Sherman, 2012). The only epilepsy severity variable not related to WISC-IV scores was seizure frequency, and this likely relates to the fact that seizure severity (rather than frequency) may be a more important determinant of epilepsy severity (Speechley et al., 2008). It is important to note that neither laterality, the presence of a positive MRI, nor a history of epilepsy surgery influenced WISC-IV scores. Similar findings on laterality were reported for the previous edition of the test (Blackburn et al., 2007). Therefore, the WISC-IV should not be used for inferring seizure laterality, and additional measures tapping more lateralized hemispheric functions should be added in the cognitive evaluation of children with epilepsy.

In terms of study limitations, it is important to note that the children in this study were a clinically referred group with early onset epilepsy, multiple AEDs, and high seizure burden, and consequently comprised a sample in which cognitive problems were likely prominent. Because of this, the results may not apply to other children with epilepsy, where cognitive deficits may be less common, or less severe. However, the results do provide needed evidence that the test has utility and validity in children with high seizure burden and high likelihood of cognitive difficulties. In the same vein, the epilepsy sample was not a population-based sample, and selection biases may have affected study inclusion. Characterization of WISC-IV profiles in different kinds of epilepsy syndromes, including milder forms of epilepsy from community samples, and subgroups with different and more well-characterized MRI findings await further research. It is important to note here that the data presented here are but one aspect of evaluating the validity of the WISC-IV in children with epilepsy. Future studies will need to evaluate other aspects of the test’s predictive and criterion validity (Strauss et al., 2006), including sensitivity to treatment-related changes such as surgery. In addition, the study does not inform as to whether the pattern of WISC performance in children with epilepsy is specific to epilepsy, but merely that it occurs in epilepsy. Comparisons to children with other kinds of neurologic conditions would be informative in this regard.

Lastly, the WISC-IV is an intelligence test, and is therefore designed expressly for measuring intelligence, not other cognitive domains. Therefore, despite its sensitivity to cognitive impairment in children with epilepsy, the WISC-IV is not sufficient for measuring cognition given the prevalence of other kinds of cognitive problems falling outside the test’s intended range, such as problems with memory, sustained attention, executive function, and aspects of language that are frequently reported in children with epilepsy (Williams et al., 2001; Hoie et al., 2006; Caplan et al., 2008; MacAllister & Sherman, 2012). The test is therefore best used as one component of a more comprehensive evaluation of cognition in children with epilepsy.

Acknowledgments

This study was supported by the Neurosciences program at Alberta Children’s Hospital and Alberta Health Services. We thank Dr. Helen Carlson, Linda Wang, and Michelle Huie for help with data collection.

Disclosure

We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. Dr. Sherman receives royalties from Oxford University Press and funds from Psychological Assessment Resources. Dr. Brooks receives funds from Psychological Assessment Resources. Dr. Fay-McClymont and Dr. MacAllister have no conflicts of interest to disclose.

Ancillary