The Neurodevelopmental Impact of Childhood-onset Temporal Lobe Epilepsy on Brain Structure and Function


Address correspondence and reprint requests to Dr. B. P. Hermann at C. G. Matthews Neuropsychology Laboratory, Department of Neurology, University of Wisconsin, 600 N. Highland, Madison, WI 53792, U.S.A. E-mail:


Summary:  Purpose: To characterize the neurodevelopmental correlates of childhood-onset temporal lobe epilepsy on brain structure and cognition compared with late-onset chronic temporal lobe epilepsy and healthy controls.

Methods: Healthy controls (n = 62) and patients with early (n = 37) versus late (n = 16) age at onset of temporal lobe epilepsy were compared with high-resolution quantitative magnetic resonance imaging (MRI) volumetrics and comprehensive neuropsychological assessment.

Results: Patients with childhood-onset temporal lobe epilepsy (mean onset age, 7.8 years) exhibited widespread compromise in neuropsychological performance and substantial reduction in brain tissue volumes extending to extratemporal regions compared with healthy controls and late-onset temporal lobe epilepsy patients (mean onset age, 23.3 years). Most evident was reduced total white-matter volume among the childhood-onset patients. Reduction in brain tissue volume, especially total white-matter volume, was associated with significantly poorer cognitive status, attesting to the clinical significance of the volumetric abnormalities.

Conclusions: Childhood-onset temporal lobe epilepsy appears to be associated with an adverse neurodevelopmental impact on brain structure and cognition that appears generalized in nature and especially evident in white-matter tissue volume.

Neuropsychological studies have demonstrated that an earlier age at onset of recurrent seizures is associated with poorer cognitive functioning. This relation, reported early in the last century (1), has been noted in studies of adult patients with diverse seizure types (2–6) and observed in neuropsychological studies of younger patients with complex partial and other types of seizures (7–9). In addition, generalized neuropsychological effects are evident in adults with the syndrome of mesial temporal lobe epilepsy (10), a syndrome defined by a focal neuropathologic substrate and early onset of recurrent seizures or initial precipitating injury (11). These cognitive effects associated with early-onset partial seizures appear more generalized than would be expected from a focal epileptogenic process.

Generalized cognitive compromise in association with early-onset localization-related syndromes of epilepsy, such as temporal lobe epilepsy, raises the question whether focal seizures and their treatment in childhood, or the etiologic insults that led to them, are associated with a more widespread influence on brain development and structure. To date, quantitative volumetric magnetic resonance imaging (MRI) studies in temporal lobe epilepsy have focused primarily on neural regions involved in the genesis and propagation of seizures, including the hippocampus (12–15), amygdala (16), entorhinal cortex (17), fornix (18), thalamus and basal ganglia (19), and temporal lobe (20–23). Fewer investigations have examined segmented volumes of gray and white matter and cerebrospinal fluid (CSF) in extratemporal lobar (21), regional (24), or total brain morphometrics (22,25).

The comparative impact of early versus late age at seizure onset on quantitative MRI volumetrics in temporal lobe epilepsy has not been systematically investigated, which is unexpected, given the neuropsychological literature reviewed earlier. In addition, some earlier animal studies suggested that seizures in the developing brain may adversely affect brain growth and development (26–28), and a human parallel has yet to be investigated. In this investigation we tested the hypothesis that childhood onset of focal temporal lobe epilepsy may be associated with a generalized neurodevelopmental effect on both brain structure and neuropsychological status. Patients with early and late onset of chronic temporal lobe epilepsy were compared with healthy controls by using high-resolution quantitative MRI volumetric measurements of total cerebral and lobar tissue volumes, including volumes of segmented gray and white matter, as well as CSF. Patients and controls also underwent comprehensive neuropsychological assessment to document the relation between age at onset of epilepsy and cognitive function and to examine directly the relation between neuropsychological status and quantitative MRI volumetric measurements.



Study participants (N = 115) included patients with temporal lobe epilepsy (n = 53) and healthy controls (n = 62). Initial selection criteria for the epilepsy patients included the following: (a) chronologic age from 14 to 60 years, (b) complex partial seizures of definite or probable temporal lobe origin, (c) absence of MRI abnormalities other than atrophy on clinical reading, and (d) no other neurologic disorder. A board-certified neurologist with special expertise in epileptology (P.R., R.S., K.R.) reviewed patients' medical records. This review, blinded to all quantitative imaging and cognitive data, included seizure semiology, previous EEGs, clinical neuroimaging reports, and all available medical records. Based on this review, each patient was classified as having complex partial seizures of definite, probable, or possible temporal lobe origin. Definite temporal lobe epilepsy was defined by continuous video-EEG monitoring of spontaneous seizures demonstrating temporal lobe onset; probable temporal lobe epilepsy was determined by review of clinical semiology with features reported to identify reliably complex partial seizures of temporal lobe origin versus onset in other regions (e.g., frontal) in conjunction with interictal EEGs, neuroimaging findings, and developmental and clinical history. Only those meeting criteria for definite and probable temporal lobe epilepsy proceeded to recruitment for study participation; patients with possible temporal lobe epilepsy were excluded.

There were no significant differences between patients with definite versus probable temporal lobe epilepsy in regard to demographic characteristics (age, gender, education), core clinical seizure features (age at onset, duration of epilepsy), quantitative MRI volumetric measurements (total tissue volume, total gray- and white-matter volume, total lobar volumes, total hippocampal volumes), or neuropsychological test performance across the administered battery. They were therefore combined for comparison with healthy controls. The results of MRI and cognitive comparisons between healthy controls and temporal lobe patients who underwent ictal monitoring are nonetheless reported in the Results section, in that the latter represent the gold standard for localization of seizure onset.

Selection criteria for healthy controls included the following: (a) chronologic age from 14 to 60 years, (b) either a friend or family member of the patient, (c) no current substance abuse or medical or acute psychiatric condition that could affect cognitive functioning, and (d) no psychotropic medications, loss of consciousness (LOC) >5 min, or history of developmental learning disorder. Chronologic age was closely comparable in the epilepsy and controls groups including means (33.4 vs. 34.1), medians (31.0 vs. 33.8), ranges (14–59 in both groups), and 95% confidence intervals (30.2–36.6 vs. 31.1–37.1). Other distributional characteristics (e.g., skewness, kurtosis) also were comparable, and we could find no evidence of asymmetric distributions or extreme outliers. The project was reviewed and approved by the University of Wisconsin Human Subjects Committee. All subjects were fully informed regarding the nature and purpose of the investigation, questions were answered, and signed consent was obtained.

Patients were interviewed, in the presence of a friend or family member whenever possible, regarding details of their epilepsy history and clinical course. Medical records were requested concerning all previous epilepsy-related hospitalizations, and records were requested from physicians who had treated the patients' epilepsy on an outpatient basis. These records were reviewed and abstracted by an individual blinded to the MRI and neuropsychological findings.

All participants underwent comprehensive neuropsychological assessment and high-resolution MRI with quantitative volumetric processing. Epilepsy patients were initially dichotomized into early (n = 37) and late (n = 16) age at onset groups based on a median split of epilepsy onset age (14 years) in the larger database of temporal lobe epilepsy patients from which this sample was selected. This resulted in groups with very disparate ages of seizure onset (mean early onset, 7.8 years; mean late onset, 23.3 years). The current consecutive sample was selected for study because quantitative MRI volumetric processing had been completed. Late age-at-onset patients with clear histories of early initial precipitating injuries (n = 7) were not included, as we were interested specifically in the effects of age at onset of recurrent seizures on brain structure and cognition. However, secondary analyses examined the potential relevance of early initial precipitating injuries.

Table 1 provides basic demographic and clinical seizure features of the subjects. Patients with childhood-onset temporal lobe epilepsy and healthy controls did not differ in chronologic age (p = 0.43), but both were significantly (p < 0.05) younger than late-onset patients. Early-onset temporal lobe epilepsy patients had significantly (p < 0.05) less education than both healthy controls and late-onset patients. Comparing the temporal lobe epilepsy groups, early-onset patients had a significantly earlier age at onset, as expected (7.8 vs. 23.3 years; p ≤ 0.001), and both temporal lobe epilepsy groups had chronic epilepsy, as evident from the long duration of seizures in each group (23.6 and 16.2 years), with significantly longer duration in the early-onset patients (p = 0.04). There was no significant difference in gender distribution across the groups (p = 0.39). The analyses to be described controlled for these demographic and clinical characteristics.

Table 1. Demographic and clinical characteristics of epilepsy patients and healthy controls
controls (C)
(n = 62)
Early onset
(n = 37)
Late onset
(n = 16)
  1. Standard deviation in parentheses.

  2. TLE, temporal lobe epilepsy.

Age (yr)33.4 (12.6)31.4 (11.6)39.6 (10.5)
Gender (M/F)25/3710/275/11
Education (yr)13.6 (2.4)12.4 (2.0)13.7 (2.3)
Age at onset (yr)7.8 (3.6)23.3 (7.3)
Duration of epilepsy (yr)23.6 (12.5)16.2 (10.0)

Images were obtained on a 1.5-Tesla GE Signa MRI scanner. Sequences acquired for each subject included (a) T1-weighted, three-dimensional SPGR acquired with the following parameters: TE = 5, TR = 24, flip angle = 40, NEX = 2, FOV = 26, slice thickness = 1.5 mm, slice plane = coronal, matrix = 256 × 192; (b) proton density (PD); and (c) T2-weighted images acquired with the following parameters: TE = 36 ms (for PD) or 96 ms (for T2), TR = 3,000 ms, NEX = 1, FOV = 26, slice thickness = 3.0 mm, slice plane = coronal, matrix = 256 × 192, and an echo train length = 8.

MRIs were acquired at the University of Wisconsin and transferred to the Image Processing Laboratory of the Mental Health Clinical Research Center at the University of Iowa, where they were processed by using a semiautomated software package [i.e., Brain Research: Analysis of Images, Networks, and Systems (BRAINS)](29–31). University of Iowa staff were blinded to the clinical and sociodemographic characteristics of the subjects. The T1-weighted images were spatially normalized so that the anterior–posterior axis of the brain was realigned parallel to the ACPC line, and the interhemispheric fissure was aligned on the other two axes. A 6-point linear transformation was used to warp the standard Talairach atlas space onto the resampled image. Images from the three pulse sequences were then co-registered by using a local adaptation of automated image registration software (32). After alignment of the image sets, the PD and T2 images were resampled into 1-mm cubic voxels, after which an automated algorithm classified each voxel into gray matter, white matter, CSF, blood, or “other”(33). The brains were then “removed” from the skull by using a neural network application that had been trained on a set of manual traces (31). Manual inspection and correction of the output of the neural network tracing was conducted. A stereotaxic method based on the Talairach atlas (34) yields measures of left and right frontal, temporal, parietal, and occipital lobes and cerebellum (35). The BRAINS software and procedures have been shown to be of high interrater reliability, intrarater reliability, and scan–rescan reproducibility, particularly for the MRI indices that are the focus of this study (29–31,33). MRI regions of interest for this investigation included total (supratentorial) cerebrum tissue volume including segmented gray- and white-matter volumes and total CSF. Total lobar tissue volumes and segmented gray- and white-matter volumes also were examined.

Neuropsychological assessment

Patients and healthy controls were administered a comprehensive test battery that included standard clinical measures of intelligence (36), language [naming (37), fluency (38)], visuoperceptual/spatial skills [facial discrimination, spatial orientation (39)], memory [verbal (40) and nonverbal (41)], and executive functions [novel problem solving (42) and speeded psychomotor processing (43)]. Table 2 depicts the cognitive domains and specific abilities assessed as well as the test measures.

Table 2. Neuropsychological test battery
  • a

     Raw scores.

  • b

     Sum recalled over trials.

  • c

     Perseverative responses (one deck of cards).

  • d


WAIS-III Performance IQ
WAIS-III Full-scale IQ
LanguageConfrontation naming
Verbal fluency
Boston Naming Testa
Controlled Oral Word Fluencya
VisuoperceptualFacial discrimination
Spatial perception
Facial Recognition Testa
Judgment of Line Orientationa
MemoryVerbal memory
Nonverbal memory
Verbal Selective Reminding Testb
Nonverbal Selective Reminding Testb
Executive functionProblem solving
Speeded psychomotor processing
Wisconsin Card-Sorting Testc
Trail-making Tests A and Bd

Data were analyzed primarily by multivariate analyses of covariance to control for pertinent clinical and demographic differences, as well as to minimize the possibility of Type 1 error. Post hoc pair-wise comparisons were tested at p = 0.05 by using two-tailed tests.


Neuropsychological performance in childhood-onset temporal lobe epilepsy is characterized by a pattern of generalized cognitive compromise

To examine the effects of age at onset of temporal lobe epilepsy on cognitive status, the neuropsychological tests were analyzed by multivariate analysis of covariance (MANCOVA) with age and education as covariates. Table 3 provides the adjusted mean scores and results of pair-wise post hoc comparisons. Hotelling's T (F = 3.4, df = 24, 188, p < 0.001) was significant. Univariate effects were significant across all measures (all p values <0.003). Pairwise comparisons showed that the childhood-onset temporal lobe epilepsy group performed significantly worse across all tests compared with the controls. Compared with late-onset patients, childhood-onset patients were significantly worse on seven of 12 tests (performance and full-scale IQ; naming; spatial orientation; verbal and visual memory; problem solving), with similar trends (p < 0.10) across three additional measures (verbal IQ, fluency, facial perception). In contrast, there were few differences between the controls and late-onset patients, the latter group performing worse on only two of 12 tests (nonverbal memory and speeded processing).

Table 3. Neuropsychological performance
 Early onset
Late onset
controls (C)
EO vs. CeLO vs. CEO vs. LO
  • TLE, temporal lobe epilepsy.

  • a

     Raw scores.

  • b

     Sum recalled over trials.

  • c

     Perseverative responses (one deck of cards).

  • d


  • e

     p values.

Verbal IQ90.9 (2.2)98.3 (3.4)102.8 (1.7)0.0000.240.08
Performance IQ94.2 (2.1)102.2 (3.2)97.9 (1.6)0.0000.110.04
Full-scale IQ91.7 (2.1)100.1 (3.2)105.2 (1.6)0.0000.160.03
Naminga48.6 (0.9)52.3 (1.3)54.8 (0.7)0.0000.100.02
Fluencya28.3 (1.9)35.1 (2.9)35.9 (1.4)0.0020.790.06
Facial discriminationa43.7 (0.7)46.3 (1.1)46.6 (0.6)0.0020.800.06
Spatial perceptiona21.3 (0.7)26.4 (1.1)25.5 (0.5)0.0000.470.00
Verbal memorya43.9 (1.4)50.9 (2.1)51.7 (1.1)0.0000.740.008
Nonverbal memoryb45.5 (1.7)54.8 (2.7)61.6 (1.3)0.0000.020.002
Problem solvingc13.4 (1.2)8.4 (1.9)8.0 (0.9)0.0010.840.03
Speeded processing Ad33.3 (2.0)31.7 (3.2)24.6 (1.6)0.0010.040.69
Speeded processing Bd81.1 (5.3)69.7 (8.2)57.5 (4.1)0.0010.190.25

Results were unchanged when comparing controls with patients with ictal confirmation of temporal lobe seizure onset. The overall MANCOVA (age and education as covariates) was significant (F = 3.9, df = 24, 146; p < 0.001), and univariate effects were significant across all cognitive measures except facial discrimination (p = 0.08). Examination of age at onset effects showed the childhood-onset patients to perform significantly worse than controls across all 12 test measures (all p values <0.03); the childhood-onset patients performed significantly worse than late-onset patients across seven of 12 measures; and the late-onset group performed significantly worse than controls on only one of 12 measures. In summary, childhood-onset temporal lobe epilepsy patients performed more poorly across a diversity of cognitive abilities compared with controls and late-onset patients, with considerably fewer differences between late-onset patients and controls.

Childhood-onset temporal lobe epilepsy is associated with a significant reduction in total cerebral tissue volume, especially evident in white-matter tissue volume

To examine the effects of age at onset of temporal lobe epilepsy on brain structure, quantitative MRI volumetric data were first analyzed by MANCOVA with age, gender, and height as covariates. Hotelling's T was significant (F = 7.8, df = 6, 192; p = 0.001). All univariate effects were significant (all p values <0.01), and adjusted mean volumetric measurements are provided in Table 4. Post hoc pair-wise comparisons showed that, compared with healthy controls, childhood-onset patients had significantly smaller total cerebrum (supratentorial) tissue volume (p = 0.001), significantly reduced white matter (p < 0.001) and increased CSF (p = 0.01) volumes with a trend (p = 0.07) toward reduced total gray-matter volume. There were no significant MRI differences between controls and late-onset epilepsy patients. Compared with late-onset patients, childhood-onset epilepsy was associated with significantly reduced total cerebrum white-matter volume (p = 0.03). Figure 1 depicts the percentage reductions in adjusted tissue volumes across these regions of interest. As can be seen, the childhood-onset group exhibited a reduction in adjusted total white-matter volume (11.9%) along with significantly less total cerebrum tissue volume (6.9%) and a modest reduction in total gray-matter volume (3.3%). These results were unchanged in a more conservative analysis using total intracranial volume as the covariate, in which the childhood-onset group continued to exhibit a significantly reduced volume of total cerebral white matter (–6.1%) compared with both healthy controls (p < 0.001) and late-onset (p < 0.015) patients, the latter two groups not differing from one another.

Table 4. Mean adjusted volumes for epilepsy patients versus controls
 Early onset
Late onset
controls (C)
EO vs CaLO vs. CEO vs. LO
  • Standard error in parentheses. All volumes adjusted for age, gender, and height. Volumes in cc3.

  • TLE, temporal lobe epilepsy; CSF, cerebrospinal fluid.

  • a

     p values.

Total cerebrum tissue1,046.1 (17.0)1,097.1 (27.2)1,123.3 (13.4)0.0010.390.12
Total cerebrum gray matter643.4 (9.5)653.4 (15.2)665.9 (7.5)0.070.460.58
Total cerebrum white matter402.7 (9.7)443.8 (15.6)457.4 (7.7)0.0010.440.03
Total cerebrum CSF96.9 (6.0)91.5 (9.6)76.7 (4.7)
Figure 1.

Total and segmented tissue volumes in childhood and late onset temporal lobe epilepsy (age, gender and height as covariates).

Because childhood-onset patients had a longer duration of epilepsy than late-onset patients (23.6 vs. 16.2 years), MRI volumetric differences between early- and late-onset groups were compared with MANCOVA, with duration of epilepsy (as well as gender and height) as covariates. This was done to rule out the possibility that reduced volumes in early-onset patients were attributable to longer duration of epilepsy. Again, Hotelling's T was significant (F = 9.3, df = 3, 42; p < 0.001), and univariate effects revealed significantly smaller total cerebral white (p = 0.04) but not gray-matter volumes (p= 0.78) and larger total CSF (p = 0.04) volumes in childhood compared with late-onset patients. Thus, volumetric differences remained evident between childhood versus late-onset epilepsy groups even after controlling for duration of epilepsy.

Examining age-at-onset effects only among patients who underwent ictal monitoring with confirmed unilateral temporal lobe onset (n = 29), the results were again unchanged. Childhood-onset patients exhibited significantly reduced adjusted (age, gender, height) total tissue (p = 0.003) and white-matter volumes (p < 0.001) and increased CSF (p = 0.04) volumes compared with controls. Compared with late-onset patients, childhood-onset patients exhibited significantly reduced total tissue (p = 0.03) and white-matter volumes (p = 0.023). Late-onset patients did not differ significantly from controls. The differences between childhood- and late-onset patients in white-matter volume remained significant (p = 0.04) when duration (with gender and height) was used as a covariate, with a trend for CSF (p = 0.07).

Finally, standardized (z-scores) for total cerebral white matter corrected for total intracranial volume were derived and examined in relation to age at onset and duration of epilepsy. Earlier age at onset (Spearman correlation = 0.31; p = 0.025) but not duration of epilepsy (Spearman correlation = –0.11, NS) was significantly associated with reduced cerebral white-matter volume.

In summary, supplemental analyses confirmed the effect of childhood onset of recurrent seizures on quantitative volumetrics.

Reductions in brain tissue volume in childhood-onset temporal lobe epilepsy are generalized and not limited to the temporal lobe

It is possible that total brain volumes in childhood-onset patients were reduced merely because of focal temporal lobe atrophy. To determine the degree to which volumetric abnormalities extended to extratemporal regions, MANCOVA was used to compare epilepsy and control groups across total lobar measurements of segmented gray- and white-matter volumes by using age, gender, and height as covariates. Hotelling's T was significant (F = 1.9, df = 16, 182, p = .023) and one-way ANOVAs were significant across all total lobar white-matter volume measurements, whereas there were no comparable significant univariate effects across total lobar gray-matter volume measurements for any region (all p values >0.14). Subsequent pair-wise comparisons showed that childhood-onset patients exhibited significantly (p < 0.01) reduced total white-matter volumes compared with controls across all lobar regions. Compared with late-onset patients, childhood-onset temporal lobe epilepsy patients exhibited significantly smaller total white-matter volumes in the temporal (p = 0.027) and parietal (p = 0.015) regions, with trends for the frontal (p = 0.10) and occipital regions (p = 0.10; Fig. 2). Late-onset patients did not differ from controls in total white-matter (or gray-matter) volume across any lobar region.

Figure 2.

Total lobar white matter volume reductions in childhood and adult onset epilepsy (total intracranial volume as covariate).

These findings were essentially unchanged when a MANCOVA was computed with duration of epilepsy (with gender and height) as covariates to rule out the possibility that the effects were due to duration of epilepsy. Hotelling's T was significant (F = 2.4, df = 8, 27; p = 0.034), and univariate effects revealed significantly reduced white-matter volumes in the parietal (p = 0.03) and temporal (p = 0.004) lobes in the childhood-onset compared with late-onset patients. There were no significant differences in lobar gray-matter volumes. Finally, examining patients with ictally confirmed childhood onset of unilateral temporal origin seizures compared with controls, the degree of ipsilateral versus contralateral volume loss was examined. MANCOVA (intracranial volume as covariate) was significant (F = 6.7, df = 4, 72; p = 0.004), and examination of univariate effects revealed a 12% reduction in cerebral white matter in the hemisphere ipsilateral to side of temporal lobe onset with a 9.8% reduction contralaterally.

In summary, the reduction in cerebral white-matter volume in patients with childhood onset of temporal lobe epilepsy is evident both ipsilateral and contralateral to the side of temporal lobe seizure onset, generally reduced across extratemporal regions compared with healthy controls, and affecting predominantly the temporal and parietal lobes compared with late-onset patients.

Reductions in total white-matter volumes are clinically meaningful, as demonstrated by significant associations with neuropsychological performance

To examine the relation directly between volumetric measurements of segmented tissue volumes and cognitive status, partial correlations (age, gender, and height as covariates) were computed between neuropsychological measures and total (supratentorial) cerebral tissue as well as segmented gray- and white-matter volumes. The observed correlations (Table 5) reveal significant associations between reduced total white-matter volumes and poorer cognitive performance across multiple cognitive abilities (verbal, performance, and full-scale IQ; verbal fluency, spatial orientation, verbal and nonverbal memory, and response inhibition). There were a greater number of significant associations between cognitive performance and white-matter than gray-matter abnormalities.

Table 5.  Relation between cognitive performance and segmented gray- and white-matter volumes in temporal lobe epilepsy
  • a

     p < 0.05.

Verbal IQ0.31a0.240.32a
Performance IQ0.230.090.34a
Full-scale IQ0.30a0.190.36a
Facial recognition0.03−0.100.16
Spatial orientation0.260.200.27a
Nonverbal memory0.190.020.34a
Verbal memory0.12−0.070.30a
Trails A−0.12−0.06−0.16
Trails B−0.14−0.07−0.19
Problem solving−0.16−0.13−0.15

Additional analyses ruled out the confounding influence of other clinical factors on the core cognitive and volumetric findings

Within the childhood-onset group, MANCOVA with age, gender, and height as covariates revealed that the identified volumetric abnormalities were not significantly related to presence/absence of an identified initial precipitating injury (p = 0.14), history of status epilepticus (p = 0.66), presence of secondarily generalized seizures (p = 0.29), number of secondarily generalized seizures (p = 0.57), or number of antiepilepsy medications (p = 0.17). MANCOVA with age and education as covariates similarly revealed that neuropsychological performance was not associated with presence/absence of an initial precipitating injury (p = 0.70), presence (p = 0.89) or number (p = 0.67) of secondarily generalized seizures, number of antiepilepsy medications (p = 0.83), presence/absence of a history of lifetime-to-date major depression (p > 0.15), or comorbid medical disease (p > 0.15).


This investigation systematically examined both high-resolution quantitative MRI volumetrics and comprehensive neuropsychological status in patients with temporal lobe epilepsy as a function of the age at onset of recurrent seizures. The results demonstrate that childhood-onset temporal lobe epilepsy is associated with significant volumetric abnormalities in both temporal and extratemporal regions, with cerebral white-matter volume reduced both ipsilateral and contralateral to the side of temporal lobe seizure onset. Generalized reduction in neuropsychological function also is evident in childhood-onset patients compared with controls and late-onset patients, consistent with the generalized pattern of volumetric abnormalities. Late-onset patients exhibit considerably fewer volumetric and cognitive abnormalities compared with healthy controls despite a history of chronic temporal lobe epilepsy (16.2 years). The overall pattern of findings therefore raises the hypothesis that childhood-onset temporal lobe epilepsy is associated with a generalized adverse neurodevelopmental impact on brain structure and function. These major points are reviewed below.

Age at onset of temporal lobe epilepsy and neuropsychological status

As noted, many studies have reported an adverse effect of early-onset epilepsy on cognition (2–5,7,8). Many of these investigations examined age at onset/cognition relations without benefit of control groups, often relying on correlational approaches. Such correlational analyses do not provide insight into the nature, distribution, or severity of the cognitive alterations associated with onset of recurrent seizures at varying developmental stages. Compared with healthy controls representing family and friends of the patients, and controlling for education and age, we found neuropsychological abnormalities to be most evident in the childhood-onset temporal lobe epilepsy group (Table 3). Childhood-onset patients performed significantly worse across all cognitive domains compared with controls, including intelligence, language, visuoperception, memory, and executive function. These findings indicate that cognitive compromise in childhood-onset temporal lobe epilepsy is widespread and not limited to memory function, and the generalized nature of the identified cognitive effects is consistent with the generalized nature of the volumetric abnormalities to be reviewed later. In contrast, late-onset patients who also had chronic seizures (16.2 years on average) exhibited far fewer cognitive impairments compared with controls, significantly different from controls on only two of 12 neuropsychological measures. In addition, the childhood-onset group exhibited greater cognitive compromise compared with late-onset patients, attesting to the neurodevelopmental impact of early-onset epilepsy on cognition. In summary, although both childhood and late-onset groups had chronic temporal lobe epilepsy, the effects on cognition were different, with diffuse and generalized cognitive difficulties in childhood-onset patients compared with both healthy controls and late-onset patients.

Age at onset of temporal lobe epilepsy and volumetric abnormalities

Compared with healthy controls, volumetric abnormalities were most evident in patients with childhood-onset temporal lobe epilepsy. Patients with childhood-onset temporal lobe epilepsy exhibited significant reductions in volumetric measurements of total cerebrum (supratentorial) tissue and total white-matter volumes with increased total CSF, with a trend toward reduced gray-matter volume (Table 4). Most evident was the ∼6–12% reduction in total white-matter volume associated with childhood epilepsy onset (Fig. 1), the degree of white-matter reduction dependent on the analytic approach used. Childhood-onset temporal lobe epilepsy patients also were most distinguished from late-onset patients by significant reduction in total white-matter volume. The generalized nature of these volumetric abnormalities was confirmed by examination of lobar tissue volumes. Childhood-onset patients exhibited significant reductions in white but not gray matter across lobar regions compared with controls (Fig. 2). In contrast, patients with late onset of recurrent temporal lobe seizures did not exhibit significant MRI volumetric differences compared with controls. The identified volumetric anomalies in childhood-onset patients were robust and remained evident when duration of epilepsy was taken into account and were confirmed when analyses were restricted to patients with ictal EEG confirmation of unilateral temporal lobe onset. Further, additional analyses confirmed that the white-matter reduction was evident both ipsilateral and contralateral to the side of unilateral temporal lobe onset and remained significant after controlling for other pertinent clinical epilepsy variables (presence/absence of initial precipitating injury, presence and number of secondarily generalized seizures, monotherapy/polytherapy, psychiatric history, comorbid medical illness).

These neurodevelopmental findings are consistent with the only reports of quantitative MRI volumetrics of cerebral volume in children with mixed seizures (44) as well as focal temporal and frontal lobe epilepsy (45), in which total (but nonsegmented) cerebral volumes were examined and found to be significantly reduced compared with healthy controls.

White-matter volumetric abnormalities in temporal lobe epilepsy and their clinical significance

The finding of significant reduction in overall (supratentorial) white-matter volume in association with childhood-onset temporal lobe epilepsy is interesting. Other studies have noted abnormalities in ipsilateral temporal lobe white matter (20,22,46–48), primarily affecting collateral white matter in the parahippocampal gyrus and diminution in gray–white demarcation secondary to reduced myelin density. White-matter volume loss also was reported to extend into extratemporal lobe regions (21,25), but this study demonstrates a link to childhood onset of recurrent seizures, characterizes the extent of the volume loss (Figs. 1 and 2), and demonstrates the clinical significance of this abnormality through its association with compromised neuropsychological performance (Table 5).

In vivo cross-sectional quantitative MRI volumetric investigations of healthy children and adolescents reliably demonstrated age-related linear increases in white-matter volumes in the context of decreases in gray-matter volume (49–53). This linear increase in white-matter volume appears relatively robust even within the context of gender and regionally specific changes in gray-matter volumes reported in cross-sectional and prospective studies (51,54–56). For example, in 116 healthy controls (19 months to 80 years), gray-matter volume increased 13% from early childhood (19–33 months) to later childhood (6–9 years) and thereafter decreased linearly by ∼5% per decade. In contrast, white-matter volume increased 74% from early childhood (19–33 months) to adolescence (12–15 years) with ongoing but significantly slower rate of growth, reaching a plateau by the fourth decade and decreasing thereafter (49). The mean ages at onset of the childhood (7.8 years) and adult-onset (23.3 years) temporal lobe epilepsy groups studied here fall within and outside the boundaries of significant white-matter development, respectively. Perhaps the presence of epilepsy and its treatment with medications during a period of maximal white-matter growth could affect development of white matter. Alternatively, it cannot be ruled out that etiologic factors that led to the development of focal temporal lobe epilepsy may have affected brain development, nor can we exclude the possible long-term effects of medications per se.

It is widely appreciated that animal studies have shown the pathophysiologic consequences of seizures in the developing brain to differ from those in the mature brain. Whereas the immature brain is far more prone to seizures than is the mature brain, developing neurons appear less vulnerable to neuronal damage and cell loss, with different consequences of seizures in the mature compared with the immature brain (see 57–60 for reviews). However, there is animal research to support the hypothesis that early-onset seizures may affect cerebral white matter.

Dwyer and Wasterlain (27) examined the effects of electroconvulsive (EC)-induced seizures (two per day for 10 days) in rats at different developmental stages (2–11 days, 9–18 days, 19–28 days). Animals were later killed, and results showed that seizures in early development selectively impaired myelin accumulation out of proportion to their overall effect on brain growth. These effects varied regionally in that they were reversible in the cerebellum but not in the forebrain. Further, examination of cerebroside (CER) and proteolipid protein (PLP), relatively specific myelin lipids, were reduced ∼13% and 11% in rats who seized between days 2 through 11 and 9 through 18, but were not reduced in mature rats subjected to a comparable number of seizures. Examination of the ratio of lipid to DNA showed that CER and PLP were reduced to a greater degree than would be explained on the basis of cell loss alone, suggesting that early seizures affected myelin accumulation, findings that were subsequently replicated (61). In addition, Wasterlain et al. (26,28,61,62) more broadly demonstrated that in addition to the curtailment of myelin and selectively and permanently affected myelin-specific lipids, repeated EC-induced seizures in the immature rat reduced brain growth and resulted in a reduction of synaptic markers in the absence of markers of neuronal cell body loss. These effects on brain growth and development were dependent on the developmental stage of the animal when seizures were induced, again more severe when seizures occurred at younger ages, these brain changes were evident in the absence of histologic lesions.

Thus, there is evidence to suggest a vulnerability of the immature brain to insults that affect markers of growth and development. The findings reported here are consistent with the general theme of the reviewed findings. Early-onset epilepsy appears to be associated with adverse neurodevelopmental effects, as reflected by quantitative MRI volumetrics and exerting a generalized adverse effect on cognitive function.


The results of this investigation suggest that childhood-onset temporal lobe epilepsy is associated with greater than expected effects on brain structure and cognition. Rather than benefiting from increased plasticity with reorganization and protection of cognitive function, the presence of recurrent seizures in the developing brain appears to be associated with an adverse effect on both brain structure and function. The findings presented here are consistent with the suggestion that there is a neurodevelopmental consequence associated with early-onset focal seizures and their treatment, or with the factors that lead to the development of recurrent childhood-onset temporal lobe seizures, on brain structures distant from the region of primary epileptogenesis (21) with associated functional (cognitive) consequences. We proposed elsewhere that this early developmental impact on brain structure and function may be related to the risk of further adverse cognitive outcomes in the face of longstanding and intractable temporal lobe epilepsy (63). It is our hypothesis that a clearer understanding of the progressive effects of chronic epilepsy on brain structure and function is dependent on a finer appreciation of the neurodevelopmental impact of epilepsy on the developing brain.

Acknowledgment: This study was supported in part by NIH NS-37738 and MO1 RR03186.