Neurocognitive profiles in children with epilepsy

Authors


Address correspondence to Claudia L. Kernan, Semel Institute for Neuroscience and Human Behavior, 760 Westwood Plaza, Los Angeles, CA 90024, U.S.A. E-mail: ckernan@mednet.ucla.edu

Summary

Purpose:  The presence of specific neurocognitive deficits may help explain why school achievement and psychosocial functioning are often worse in children with epilepsy than would be predicted by their global intellectual functioning. This study compared children with two forms of epilepsy: localization-related epilepsy with complex partial seizures (CPS) and childhood absence epilepsy (CAE), to determine whether they display distinct neurocognitive profiles.

Methods:  Fifty-one children with CPS, 31 children with CAE, and 51 controls underwent neuropsychological testing assessing verbal memory, visual memory, and executive functioning. Groups were compared in these cognitive domains. Within-group analyses were also conducted to examine seizure-related factors that may be related to neuropsychological test performance.

Key Findings:  When compared to controls, children with CPS showed a mild generalized cognitive deficit, whereas children with CAE did not. When we controlled for intelligent quotient (IQ), both epilepsy groups showed poorer performance relative to controls in the domain of verbal memory. When the epilepsy groups were compared to one another, the CPS group performed significantly poorer than the CAE group on a test of generalized cognitive functioning. However, in the specific domains of executive functioning, verbal memory, and visual memory the epilepsy groups did not differ when compared to one another.

Significance:  Neurocognitive deficits present in the context of grossly intact global intellectual functioning highlight the importance of neuropsychological screening in both children with CPS and children with CAE.

Although a significant proportion of children with severe forms of epilepsy score below average on intelligence tests, the majority of children with epilepsy have intellectual functioning within the average range (Berg et al., 2008). Despite their mostly intact global intellectual functioning, however, children with epilepsy as a group are at high risk for poor academic functioning (Sillanpaa, 2004; Aldenkamp et al., 2005; Fastenau et al., 2008) and negative psychosocial outcomes (Wirrell et al., 1997; Sillanpaa et al., 1998). Researchers have therefore highlighted the importance of neuropsychological testing in children with epilepsy, positing that specific cognitive deficits in domains other than general intelligence may help explain why school achievement and psychosocial functioning are often worse than would be predicted by their intelligence scores alone (Nolan et al., 2004; Jocic-Jakubi & Jovic, 2006; Fastenau et al., 2008).

Past research has indicated that children with epilepsy display specific cognitive weaknesses, with a number of studies demonstrating deficits in the domains of memory and executive functioning (Jambaque et al., 1993; Bailet & Turk, 2000; Lassonde et al., 2000; Pavone et al., 2001; Culhane-Shelburne et al., 2002; Nolan et al., 2004; Henkin et al., 2005; Hoie et al., 2005; Borden et al., 2006; Hermann et al., 2006; Hommet et al., 2006; Jocic-Jakubi & Jovic, 2006; MacAllister & Schaffer, 2007; Parrish et al., 2007; Seidenberg et al., 2007; Hermann et al., 2008; Pulsipher et al., 2009). However, the question of whether children with different types of seizures or epilepsy syndromes show distinct cognitive profiles has yet to be resolved in the literature. Most studies to date looking at cognitive functioning in children with epilepsy have compared children with epilepsy to a healthy control group or standard scores. The epilepsy groups in these studies have been heterogeneous, limiting the conclusions that can be drawn in terms of delineating distinct cognitive profiles. Although several recent studies have directly compared children with different types of seizures and epilepsy syndromes, findings have been mixed with regard to whether there are differences among epilepsy groups. In addition, children have been grouped together using different criteria in terms of type of seizure and epilepsy syndrome, making comparisons across studies difficult.

For example, some studies have compared children with generalized seizures to children with focal epilepsies and found that children with focal epilepsies perform poorer on tests of memory (Jambaque et al., 1993; Nolan et al., 2004). Other studies comparing children with localization-related epilepsies to children with generalized seizures have not found significant differences in memory functioning (Williams et al., 1998, 2001; Hermann et al., 2006). In the domain of executive functioning, some studies comparing children with localization-related epilepsy to children with idiopathic generalized epilepsy have found no or small differences between the groups (Hermann et al., 2006; Parrish et al., 2007). Similarly, a study comparing children with absence seizures to children with complex partial seizures found no differences among the groups in executive functioning (Williams et al., 1998). In contrast, a study comparing children with juvenile myoclonic epilepsy (JME) to children with benign childhood epilepsy with centrotemporal spikes (BCETS) found children with JME to have significant executive dysfunction, whereas children with BCECTS did not show the same pattern (Pulsipher et al., 2009).

In addition to seizure type and epilepsy syndrome, a number of other seizure-related variables have been identified as contributing to variability in cognitive outcomes for children with epilepsy. Age at onset of epilepsy has been shown to predict later cognitive ability, with younger age of onset associated with poorer cognitive outcomes (Jambaque et al., 1993; Schoenfeld et al., 1999; Pavone et al., 2001; Nolan et al., 2004; Hoie et al., 2006; Jocic-Jakubi & Jovic, 2006). Other disease-related factors such as longer duration of illness (Nolan et al., 2004), greater numbers of seizures (Hoie et al., 2006; Jocic-Jakubi & Jovic, 2006; Fastenau et al., 2009), and treatment with one or more antiepileptic drugs (AEDs) (Schoenfeld et al., 1999; Nolan et al., 2004; Fastenau et al., 2009) have also been associated with increased cognitive difficulties.

The purpose of the present study was to compare neurocognitive functioning in children with localization-related epilepsy with complex partial seizures (CPS), usually thought to involve cognitive impairment (particularly memory), and children with childhood absence epilepsy (CAE), often regarded as a benign disorder without cognitive or other comorbidities. Children with general cognitive functioning in the average range were compared on neuropsychological measures of verbal memory, visual memory, and executive functioning. We also investigated the effects of clinical variables on neurocognitive functioning in this population, including age at seizure onset, duration of illness, seizure frequency, number of AEDs, as well as lateralization and localization of epileptic activity on EEG. The current study included a larger sample size than in most previous studies of neurocognitive functioning in children with epilepsy, a healthy control group, and groups with comparable demographic variables. If patterns specific to epilepsy group and clinical seizure variables can be identified, this could aid in the early identification of children at risk of developing cognitive problems and in tailoring interventions that address their specific needs.

Methods

Subjects

The current study included 51 children with CPS, 31 children with CAE, and 51 controls, aged 6–16 years with IQ scores between 70 and 130. Table 1 presents demographic features of the sample and the recruitment sites of the subjects. We determined socioeconomic status (SES) using the Hollingshead 2 factor index (Hollingshead, 1973), based on parental occupational and educational status. As summarized in Table 1, there were significantly more Caucasian subjects in the CPS group than in the CAE and control groups (χ2 = 6.5, p < 0.05). There were no significant differences in the demographic variables of the CPS, CAE, and control groups.

Table 1.   Demographics
VariablesDiagnosis
ControlCPSCAE
  1. CPS, complex partial seizures; CAE, childhood absence epilepsy.

  2. aSignificantly more Caucasian subjects in CPS group than in CAE and Control group, p < 0.05.

n515331
Male (%)514945
Female (%)495155
Chronological age (year) (SD)10 (3)10 (3)9 (2)
Ethnicity (%)   
 Caucasiana396239
 Non-Caucasian613861
Socioeconomic status (%)   
 High293326
 Low716774

The primary study inclusion criterion for each epilepsy subject was that he/she had a diagnosis of epilepsy (CPS or CAE) and at least one seizure in the year prior to participation in the study. At each site a pediatric neurologist made the diagnosis of CPS or CAE based on clinical history and electroencephalography (EEG) findings, according to the International Classification of Epilepsy (Commission, 1989). Children with a clinical history of CPS but no EEG evidence of epileptic activity were also included in this study. All CAE patients had EEG evidence of 3 Hz spike and wave in addition to absence seizures induced by hyperventilation. We excluded patients with a mixed seizure disorder, previous epilepsy surgery, atypical spike and wave complexes, juvenile myoclonic epilepsy, generalized tonic–clonic seizures, a neurologic illness other than epilepsy, chronic medical illness, imaging evidence for structural brain abnormalities, a metabolic disorder, a hearing disorder, mental retardation based on school/classroom placement, and bilingual speakers of American English who did not attend English speaking schools or speak English at home.

We recruited 42% of the CPS and 33% of the CAE subjects from tertiary centers (i.e., UCLA- and USC-based clinics) and 58% of the CPS and 67% of the CAE from community services (i.e., Los Angeles and Anaheim Kaiser Permanente, the Los Angeles and San Diego Chapters of the Epilepsy Foundation of America, and private practices). There were no differences between the CPS group and the CAE group in terms of recruitment site (Table 2). UCLA institutional review board (IRB)–approved recruitment flyers were available for parents of children with CPS and CAE at each recruitment site. Parents who decided to enter their children into the study contacted the study coordinator who provided information about the study and used a UCLA IRB–approved telephone script to determine if the children met the study’s inclusionary but none of the exclusionary criteria. The study coordinator also contacted the child’s pediatric neurologist to confirm the child’s diagnosis and to rule out exclusionary criteria. One UCLA pediatric neurology investigator (W.D.S.) reviewed the history, EEG records, and diagnosis of each CAE subject from the different recruitment sites. If he did not concur with the diagnosis or EEG findings, the child was not included in the study.

Table 2.   Recruitment location and seizure characteristics
VariablesDiagnosis
CPSCAE
  1. N/A, not applicable.

  2. aCAE significantly different than CPS, p < 0.0001.

  3. bLog-transformed seizure frequency was used in the analyses because the seizure frequency ranged from 1–50,000 seizures per year in the CAE subjects.

n5331
Recruitment location (%)  
 Tertiary4233
 Community5867
Seizure frequency (log) (SD)a,b3 (2)6 (3)
Antiepileptic drugs (%)  
 None26
 Monotherapy7568
 Polytherapy2326
Age of onset (year) (SD)6 (3)6 (2)
Duration of epilepsy (year) (SD)4 (2)3 (2)
Seizure lateralization (%)  
 Left28N/A
 Right20N/A
 Bilateral28N/A
 No lateralization24N/A
Seizure localization (%)  
 Temporal36N/A
 Frontotemporal30N/A
 Not localized34N/A
Secondary generalization (%)250

The parents and children’s medical records provided information on seizure-related variables, including seizure frequency, current AEDs, age of onset, and illness duration (Table 2). EEG recordings done around the time of diagnosis indicated left, right, bilateral, and no lateralization in 28%, 20%, 28%, and 24% of the CPS subjects, respectively. The localization of epileptic activity was temporal in 36%, frontotemporal in 30%, and not localized in 34% of the CPS subjects. Thirteen of the CPS children had secondary generalization.

To include children from a wide range of ethnic and socioeconomic status backgrounds similar to that of the CPS and CAE group, we recruited the control subjects from four public and two private schools in the Los Angeles community. The study coordinator screened potential participants for neurological, psychiatric, language, and hearing disorders through a telephone conversation with a parent. We excluded children with diagnoses of these disorders in the past from the study.

Procedures

This study was conducted in accordance with the policies of the University of California, Los Angeles Human Subjects Protection Committees. Informed assents and consents were obtained from all subjects and their parents, respectively.

Neuropsychological assessment

All subjects were administered neuropsychological tests by individuals trained in standardized testing procedures. For safety reasons, testers were not blind to whether the subjects had epilepsy, but testers were blind to the epilepsy group (CPS or CAE) and all other seizure variables. The Wechsler Intelligence Scale for Children-Third Edition (WISC-III) was administered to children to obtain a measure of global intellectual functioning (Wechsler, 1991). The Full Scale IQ score was generated from each test. Executive functioning was assessed with subtests from the Stroop Color-Word Test (Stroop Color Naming, Stroop Interference) (Golden, 1975), the Wisconsin Card Sorting Test (Total Errors, Perseverative Errors)(Heaton et al., 1993), and the Test of Memory and Learning (TOMAL; Digits Forward, Digits Backwards) (Reynolds & Bigler, 1994). Verbal memory was assessed with subtests from the California Verbal Learning Test-Children’s Version (CVLT-C; CVLT-C Total, Long-Delay Free Recall, Retention, and Discriminability) (Delis et al., 1994), the TOMAL (Memory for Stories, Memory for Stories-Delayed), and the Doors and People Test (Auditory Name Recall, Long-Delayed Auditory Name Recall, Visual Name Recall, and Auditory Retention) (Baddeley et al., 1994). Visual memory was assessed with subtests from the Doors People Test (Door Recall, Shapes Recall, Long Delayed Shapes Recall, and Visual Retention).

Data analysis

A multivariate analysis of variance (MANOVA) was used to compare the CPS, CAE, and control groups in the domain of verbal memory. Standardized scores were used for this analysis. Because standardized test scores were not uniformly available for all neurocognitive tests within the executive functioning and visual memory domains (i.e., Stroop, and Doors and People tests), multivariate analyses of covariance (MANCOVAs) using raw scores, controlling for age, were used to compare the CPS, CAE, and control groups on neurocognitive tests in these domains. For all MANOVAs and MANCOVAs, subjects missing data were excluded. Significant MANOVA and MANCOVA effects were subsequently analyzed using univariate analysis of variance (ANOVA) and post hoc comparisons when appropriate. To determine whether demographic variables confounded results, we analyzed the relationship between gender, SES, and ethnicity on neurocognitive tests scores using separate ANOVAs. Gender, SES, and ethnicity did not significantly influence test scores for any of the three groups, and were therefore not controlled for in the analyses described above.

Within the domain of executive functioning, Perseverative Errors on the WSCT were not included in the data analyses, because an error in data recording during test administration resulted in loss of data for this variable. In the domain of verbal memory, a separate MANCOVA using raw scores controlling for age was conducted comparing the groups on verbal memory subtests from the Doors and People test, because significantly fewer subjects (N = 48) completed this test. This was because the Doors and People test was added to our battery later in the study. Doors and People completers did not differ from the rest of the sample in terms of demographics.

In addition, to determine whether neuropsychological testing discriminated between the epilepsy and control groups above and beyond what would be predicted by their general intellectual functioning, the between-group analyses described above were also conducted controlling for IQ.

For only those cognitive domains on which the groups differed significantly, within-group analyses were performed to explore the association of seizure-related variables (age of onset, log seizure frequency, lateralization and localization, and number of AEDs) and neurocognitive domains. Specifically, for the executive functioning and verbal memory domains, within-group MANCOVAs were conducted with the neurocognitive test scores as the dependent variables and the seizure variables as predictors, controlling for age. Significant MANCOVA effects were subsequently analyzed using univariate analysis of variance (ANOVA) and post hoc comparisons when appropriate.

Results

Comparisons between the control and epilepsy groups

Table 3 summarizes the means and standard deviations for each of the three groups on the global intellectual functioning, executive functioning, verbal memory, and visual memory variables. There were significant differences among the groups in terms of global intellectual functioning. Specifically, the ANCOVA comparing the CPS, CAE, and control groups, with age as a covariate revealed significant group differences: F2,135 = 10.29, p ≤ 0.0001, with the control group and the CAE group scoring significantly higher than the CPS group (mean difference = −13, 95% CI [−17 to −8] and mean difference = −7, 95% CI [−11 to −0.2]), respectively).

Table 3.   Neurocognitive performances in control, CPS, and CAE groupsa
  Control
Mean (SD)
CPSCAE
Mean (SD)95% CIMean (SD)95% CI
  1. *p < 0.05.

  2. **p < 0.01.

  3. ***p < 0.0001.

  4. aRaw scores are reported unless otherwise indicated. Lower scores indicate better performance for Stroop Color Naming, Stroop Interference, and WCST Total Errors. For all other tests, higher scores indicate better performance.

  5. bSignificant difference between CPS and controls.

  6. cSignificant difference between CPS and CAE.

  7. dSignificant difference between CAE and controls.

Full Scale IQb***c*107 (12)94 (15)−17 to −8101 (16)−11 to 0.2
Attention and executive functioning     
 Stroop Color Namingb*53 (15)69 (30)7 to 2465 (23)4 to 20
 Stroop Interference43 (30)37 (29)−14 to 250 (35)−6 to 50
 WCST Total Errorsb*d*30 (21)43 (24)5 to 1938 (24)−1 to 17
 Digits Forward39 (17)34 (21)−11 to 128 (12)−16 to −7
 Digits Backward23 (13)16 (15)−10 to −218 (13)−9 to 0.4
Verbal memory     
 CVLT Total (T score)b**d*56 (9)48 (11)−11 to −551 (11)−9 to −1
 CVLT Long-Delay Free (z-score)b*1 (1)−0.4 (1)−1 to 00.03 (1)−1 to −0.1
 Discriminability (z-score)b*1 (1)−0.1 (1)−1 to 00.2 (1)−1 to −0.01
 TOMAL Memory for Stories (standard score)b**d**12 (3)10 (3)−3 to −110 (3)−3 to −1
 TOMAL Memory for Stories- Delayed (standard score)b**d**11 (3)8 (3)−4 to −29 (3)−4 to −1
Verbal Memory (Doors and People)     
 Auditory Name Recall23 (7)19 (8)−7 to 0.120 (7)−7 to 1
 Long-Delay Auditory Name Recall9 (3)6 (4)−4 to −18 (3)−2 to 1
 Visual Name Recall16 (6)13 (7)−7 to 0.217 (5)−2 to 3
 E–F Difference (Auditory Retention)1 (2)3 (2)0.1 to 21 (2)−1 to 1
Visual memory     
 Door Recall17 (4)15 (5)−5 to −0.318 (3)−1 to 2
 Shapes Recall32 (5)26 (9)−10 to −131 (4)−4 to 2
 Long Delayed Shapes11 (2)10 (2)−2 to 0.111 (1)−0.4 to 1
 G–H Difference (Visual Retention)0.2 (2)1 (2)−1 to 20.4 (1)−1 to 1

For the executive functioning tests, the MANCOVA comparing the CAE, CPS, and control group scores, with age as a covariate, yielded a significant effect of group: F10,218 = 2.24, p < 0.05. Univariate analyses yielded significant group differences for Stroop Color Naming Time: F3,117 = 5.72, p < 0.01; and WCST Total Errors: F3,117 = 3.98, p < 0.05. Post hoc tests revealed that the CPS group performed more poorly than the controls on Stroop Color Naming Time (mean difference = 16, 95% CI [7–24]) and WCST Total Errors (mean difference = 13, 95% CI [5–19]). The CAE group performed more poorly than the controls on WCST Total Errors only (mean difference = 8, 95% CI [−1 to 17]).

In the domain of verbal memory, the MANOVA comparing the groups on the CVLT and TOMAL measures yielded a significant effect for group: F10,252 = 3.86, p < 0.0001. Univariate analyses yielded significant group differences for CVLT Total: F2,133 = 6.63, p < 0.01; CVLT Long Delay Free Recall: F2,133 = 8.7, p < 0.01; CVLT Discriminability: F2,133 = 4, p < 0.05; TOMAL Memory for Stories: F2,133 = 8.93, p < 0.01; and TOMAL Memory for Stories Delayed: F2,133 = 15.43, p < 0.0001. Post hoc tests revealed that the CPS group performed more poorly than the controls on CVLT Total: mean difference = −8, 95% CI (−11 to −5); CVLT Long-Delay Free Recall: mean difference = −0.9, 95% CI (−1 to −0.5); CVLT Discriminability: mean difference = −0.7, 95% CI (−1 to −0.3); TOMAL Memory for Stories: mean difference = −2, 95% CI (−3 to −1); and TOMAL Memory for Stories Delayed: mean difference = −3, 95% CI (−4 to −2). The CAE group performed more poorly than the controls on CVLT Total: mean difference = −5, 95% CI (−9 to −1); TOMAL Memory for Stories: mean difference = −2, 95% CI (−3 to −1); and TOMAL Memory for Stories Delayed: mean difference = −2, 95% CI (−4 to −1). For the Doors and People verbal memory tests, the MANCOVA did not yield a significant effect.

In the domain of visual memory, MANCOVA comparing the groups on the Doors and People measures yielded a significant effect for group: F8,82 = 1.83, p < 0.05; however, univariate analyses did not yield significant effects.

Comparisons between the control and epilepsy groups, controlling for IQ

For the executive functioning tests, the MANCOVA comparing the CAE, CPS, and control group scores, with age and IQ as a covariates, did not yield a significant effect: F10,236 = 1.01, p = 0.44. For the Doors and People tests, neither visual memory nor verbal memory MANCOVA yielded a significant effect: F8,88 = 1.83, p = 0.08, and F8,104 = 1.07, p = 0.39, respectively.

In the domain of verbal memory, the MANCOVA comparing the groups on the CVLT and TOMAL measures, controlling for IQ, yielded a significant effect of group (F10,272 = 2.34, p = 0.05). Univariate analyses yielded significant group differences for CVLT Long Delay Free Recall: F3,144 = 4.39, p < 0.05; TOMAL Memory for Stories: F3,144 = 5.40, p < 0.01; and TOMAL Memory for Stories Delayed: F3,144 = 7.10, p < 0.01. Post hoc tests revealed that the CPS group performed more poorly than the controls on CVLT Long Delay Free Recall (adjusted means = −0.26 and 0.39) and TOMAL Memory for Stories Delayed (adjusted means = 9.37 and 10.95). The CAE group performed more poorly than the controls on TOMAL Memory for Stories (adjusted means = 10.12 and 11.80) and TOMAL Memory for Stories Delayed (adjusted means = 9.07 and 10.95).

Relationship between neurocognitive performance and seizure-related variables

In the executive functioning domain, for the CPS group, the within-group MANCOVA exploring the association between the seizure-related variable and the tests of neurocognitive functioning did not yield a significant relationship between seizure-related variables and tests of neurocognitive functioning. For the CAE group, there was a significant main effect for age of seizure onset (F5,20 = 7.19, p < 0.001). Post hoc t-tests revealed that younger age of onset was associated with poorer performance on Digits Forward (t28 = −3.34, p < 0.05).

In the domain of verbal memory, for the CPS group, the within-group MANCOVA exploring the association between the seizure-related variables and the tests of neurocognitive functioning in the verbal domain where there were group differences (CVLT and TOMAL) yielded a significant main effect for seizure frequency (F5,40 = 2.84, p < 0.03). Post hoc t-tests were performed and revealed that greater seizure frequency was associated with poorer performance on CVLT Total (t48 = −2.06, p < 0.05) and TOMAL Memory for Stories Delayed Recall (t48 = −2.30, p < 0.05). For the CAE group, the MANCOVA did not yield a significant relationship between the seizure related variables and neurocognitive performance on the CVLT and TOMAL. AEDs, lateralization, and localization were not related to neurocognitive functioning.

Discussion

The present study is one of few that compare neurocognitive functioning in relatively large samples of children with different types of epilepsy with comparable demographic variables. When compared to each other, the CPS and CAE group differed significantly in generalized cognitive functioning, but not in the specific cognitive domains of executive functioning, verbal memory, and visual memory. Compared to controls, children with CPS showed a mild generalized cognitive deficit reflected by a mean Full Scale IQ of 94 (SD = 15), whereas children with CAE did not (mean Full Scale IQ = 102, SD = 16). Relative to controls, both groups showed poorer performances in the domains of executive functioning and verbal memory, with the CPS group showing poorer across a greater range of tests. This is consistent with a study of children with CPS, which found a pattern of relatively diffuse and generalized cognitive dysfunction in children with CPS (Schoenfeld et al., 1999).

To determine whether neuropsychological testing could help identify specific neurocognitive weaknesses in children with CPS and CAE above and beyond what would be predicted by their general intellectual functioning, we also compared our study groups controlling for IQ. After controlling for IQ, no differences were found between the epilepsy groups and the controls in the domain of executive functioning. However, both the CPS and CAE groups demonstrated mild verbal memory deficits that were not associated with deficits in generalized cognitive functioning. Specifically, controlling for IQ, the CPS group performed poorer than the controls on delayed verbal memory tasks (memory for word lists and memory for stories), and the CAE performing poorer than the control group on story memory tasks (both immediate and delayed). The present study is among the few to document cognitive impairments in children with CAE. CAE has historically been considered a “benign” disorder, with little effect on neurocognitive functioning. However, increasing evidence suggests that this may not be the case. In a prospective study with 39 children with generalized absence seizures, children with absence seizures performed more poorly on tests of attention/executive functioning/construction than controls (Fastenau et al., 2009). Although we found intact nonverbal memory skills with mild verbal memory deficits in children with CAE, two other studies found the reverse pattern of mild nonverbal memory problems, with intact verbal memory skills (Pavone et al., 2001; Nolan et al., 2004). These inconsistencies in findings may be partially due to small sample sizes and subsequent lack of power to detect differences. Varying degrees of seizure control in the study populations and differences across studies in instruments used to assess verbal and visual memory may also partially account for these discrepant findings. More research is needed to better characterize the factors that account for within-group variability in cognitive functioning among children with CAE. Nonetheless, these findings, along with studies showing language impairments (Caplan et al., 2009) in children with CAE suggest that CAE is a syndrome that affects multiple cognitive functions.

The literature is somewhat inconclusive regarding the relationship between seizure variables and cognitive functioning in children with epilepsy, with some studies finding relationships and others not (Schoenfeld et al., 1999; Williams et al., 2001; Nolan et al., 2004; Hermann et al., 2006; Jocic-Jakubi & Jovic, 2006; Fastenau et al., 2009; Austin & Fastenau, 2010; see Jambaque et al., 1993). In our sample, seizure frequency was associated with cognitive test performance in the CPS group, but not in the CAE group. Younger age of onset was associated with poorer performance on cognitive tests for the CAE group, but not the CPS group. It may be that the relationship between age of seizure onset and frequency and neurocognitive performance differs depending on type of epilepsy and involved brain regions. However, the literature does not adequately address this question, and further exploration of the mechanisms underlying the association between these factors and neurocognitive functioning or lack thereof is needed. In particular, large prospective studies initiated close after the onset of epilepsy will be useful in characterizing the course of neuropsychological functioning over time and in determining the relative contributions of seizures variables on neuropsychological impairment (Hermann et al., 2006; Fastenau et al., 2009; Austin & Fastenau, 2010).

Along with the cross-sectional design of this study and the lack of power to detect differences in the visual memory domain, other limitations should be noted. The study was not population-based, which limits the generalizability of the findings. Although the control group was well-matched to the epilepsy group on demographic variables, the control group performed better than average on many of the neuropsychological tests. In addition, information on lateralization and localization of the EEG findings of the CPS group was limited because it was based on reports of EEG recordings done at the time of the diagnosis of epilepsy.

Despite its limitations, the present study demonstrates that although children with CPS and CAE may look similar in terms of their academic and behavioral problems (Caplan et al., 2004, 2008; Fastenau et al., 2008; Caplan et al., 2009), the cognitive underpinnings of their functional deficits may be distinct, with the CPS group showing more diffuse cognitive impairment. In addition, the present study demonstrated verbal memory deficits in children with CAE and CPS in the context of intact global intellectual functioning. Although the deficits were mild in magnitude for many of the children studied, it is important to note that, overall, a disproportionate number of children with CPS and CAE fell more than one standard deviation below controls, even when controlling for IQ. There is clearly a subset of children who are at risk for academic and functional difficulties due to memory problems. More research is needed to determine whether these memory difficulties may be present in the nonverbal domain, as the present study was underpowered to detect visual memory differences. Because neuropsychological testing is costly in terms of both time and financial resources, findings may be useful for designing more targeted memory screenings of children with epilepsy.

Acknowledgments

This study was supported by grants NS32070 (RC) and MH 67187 (RC).

Disclosure

None of the authors has any conflicts of interest to disclose. 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.

Ancillary