SEARCH

SEARCH BY CITATION

Keywords:

  • Idiopathic generalized epilepsy;
  • Auditory event-related potentials;
  • Language;
  • Children

Abstract

  1. Top of page
  2. Abstract
  3. METHODS
  4. Participants
  5. Patient group
  6. Control group
  7. Auditory event-related potential recordings
  8. Auditory stimuli
  9. Procedure
  10. Data analysis
  11. RESULTS
  12. DISCUSSION
  13. Acknowledgments
  14. REFERENCES

Summary: Purpose: Auditory processing of increasing acoustic and linguistic complexity was assessed in children with idiopathic generalized epilepsy (IGE) by using auditory event-related potentials (AERPs) as well as reaction time and performance accuracy.

Methods: Twenty-four children with IGE [12 with generalized tonic–clonic seizures (GTCSs), and 12 with absence seizures (ASs)] with average intelligence and age-appropriate scholastic skills, uniformly medicated with valproic acid (VPA), and 20 healthy controls, performed oddball discrimination tasks that consisted of the following stimuli: (a) pure tones; (b) nonmeaningful monosyllables that differed by their phonetic features (i.e., phonetic stimuli); and (c) meaningful monosyllabic words from two semantic categories (i.e., semantic stimuli).

Results: AERPs elicited by nonlinguistic stimuli were similar in healthy and epilepsy children, whereas those elicited by linguistic stimuli (i.e., phonetic and semantic) differed significantly in latency, amplitude, and scalp distribution. In children with GTCSs, phonetic and semantic processing were characterized by slower processing time, manifested by prolonged N2 and P3 latencies during phonetic processing, and prolongation of all AERPs latencies during semantic processing. In children with ASs, phonetic and semantic processing were characterized by increased allocation of attentional resources, manifested by enhanced N2 amplitudes. Semantic processing also was characterized by prolonged P3 latency. In both patient groups, processing of linguistic stimuli resulted in different patterns of brain-activity lateralization compared with that in healthy controls. Reaction time and performance accuracy did not differ among the study groups.

Conclusions: AERPs exposed linguistic-processing deficits related to seizure type in children with IGE. Neurologic follow-up should therefore include evaluation of linguistic functions, and remedial intervention should be provided, accordingly.

Auditory event related potentials (AERPs) are measures of electrical brain activity that may be used to assess higher-level cognitive processing (1). They are considered markers of specific aspects or stages of information processing and can therefore provide insight into the timing, ordering, and interactions of intermediate processes that are engaged in auditory processing (2–4). For this reason, AERPs have been used extensively to study brain mechanisms underlying cognition and to characterize information processing in healthy and cognitively impaired populations.

One group of individuals considered to be cognitively impaired and yet to have been understudied with AERPs are patients with epilepsy. Furthermore, the few AERP studies that have been conducted have been limited primarily to adults with epilepsy and to pure tones as the eliciting stimulus (5–9). Among the scarce studies that investigated AERPs in childhood epilepsy is that of Konishi et al. (10), which elicited AERPs by using pure tones in 129 children with different types of epilepsy. They found a significant prolongation of the P3 component that is believed to reflect stimulus encoding, recognition, and classification (4) in children with symptomatic partial epilepsy. In children with idiopathic generalized and partial epilepsies, however, P3 was not significantly different from that of healthy controls. In another study in children, the P3 component has proved to be a sensitive measure for monitoring the cognitive adverse side effects of antiepileptic drugs (AEDs) before and after the initiation of treatment (11).

Although studies of AERP with pure tones provide insight into auditory processing of simple stimuli, this information cannot be generalized to more complex auditory stimuli such as speech. Furthermore, it is assumed that in epilepsy patients with suspected cognitive deficits, AERP elicited by linguistic stimuli might provide important information regarding auditory processing that is not available from behavioral speech-perception studies. To our knowledge, linguistic stimuli have not been used to elicit AERPs in epilepsy patients. The main goal of the present study was therefore to use AERPs to linguistic stimuli to gain insight into the brain processes underlying linguistic processing in epilepsy patients.

The nature and severity of the cognitive deficits in children with epilepsy are often complex and unclear because many interrelated factors may be involved. These include the occurrence of seizures of various types; the pathophysiology underlying epilepsy; the brain pathology, either causative or secondary to the epilepsy; AEDs; and subclinical discharges causing transitory cognitive impairment (TCI) (12). Thus the evaluation of the exclusive effects of the underlying “seizure condition” on linguistic processing requires that the children in the epilepsy group have (a) idiopathic epilepsy; (b) average intelligence; (c) monotherapy, because polytherapy has deleterious effects on cognitive function (13,14); and (d) infrequent seizures, receiving AEDs (at therapeutic levels) that are not known to cause cognitive impairment. Theoretically, this information may contribute to the elucidation of the influence of the underlying “seizure condition” on linguistic processing. Clinically, it may allow the identification of the impaired linguistic levels and thus contribute to the planning of remediation programs for these patients.

Patient group

  1. Top of page
  2. Abstract
  3. METHODS
  4. Participants
  5. Patient group
  6. Control group
  7. Auditory event-related potential recordings
  8. Auditory stimuli
  9. Procedure
  10. Data analysis
  11. RESULTS
  12. DISCUSSION
  13. Acknowledgments
  14. REFERENCES

Of a cohort of 60 children diagnosed with IGE at the epilepsy unit at Schneider Children's Medical Center of Israel (CMCI), 24 (14 girls and 10 boys) were included in the study based on the following inclusion criteria: (a) average performance (10 ± 3 standard score) in four subtests of the WISC-R [two verbal (vocabulary; information) and two nonverbal (block design; picture completion) subtests]; (b) attendance in ordinary schools with no reported learning disability or any other difficulties in scholastic skills; (c) normal hearing based on audiometric threshold evaluation for octave frequencies between 250 and 8,000 Hz; and (d) well-defined right-handedness according to the translated Edinburgh Inventory for Handedness (15). Of the 60 children with epilepsy, 36 were excluded from the study for the following reasons: coexistence of other pathologies in addition to epilepsy (e.g., neurofibromatosis, arthritis, growth disorders, anorexia); history of neonatal hyperbilirubinemia necessitating exchange transfusion; low birth weight (<2,000 g); learning disability; speech and language disorders; hearing loss; and mental retardation.

All children in the patient group received valproic acid (VPA; Depalept) monotherapy, and their contrast-enhanced brain computed tomography (CT) was normal. Twelve children (six boys and six girls) had generalized tonic–clonic seizures only (GTCS group), and their demographics and seizure characteristics are shown in Table 1. Complete seizure control was achieved in eight, whereas the remaining four children were seizure free for ≥2 months before the experimental session. The average number of seizures per year was 2.3 (range, one to 10). Interictal EEG was normal in six of the eight seizure-free children. In all other children (six), interictal EEG showed three to four per second spike–wave discharges. The remaining 12 children (eight girls and four boys) had absence seizures only (AS group), and their demographics and seizure characteristics are shown in Table 1. Six of them were seizure free 1–12 months before the experimental session, and their interictal EEG was normal. The other six children with ASs reported having one to four seizures every 2–3 days, and their interictal EEGs showed generalized three-per-second spike–wave discharges.

Table 1. Patient clinical characteristics
GroupAge at testing (yr) Mean RangeAge at seizure onset (yr) Mean RangeSerum level of valproate (μg/ml) Mean Range
  1. GTCS, generalized tonic–clonic seizure; AS, absence seizure.

GTCS14.3  12–16.312.3  6–1588.1  65–100
AS14.4  11–16.3 7.2  4–1173  65–90 

Control group

  1. Top of page
  2. Abstract
  3. METHODS
  4. Participants
  5. Patient group
  6. Control group
  7. Auditory event-related potential recordings
  8. Auditory stimuli
  9. Procedure
  10. Data analysis
  11. RESULTS
  12. DISCUSSION
  13. Acknowledgments
  14. REFERENCES

Twenty healthy children (12 girls and eight boys) with a mean age of 14.4 years (range, 10.6–16.6 years) were recruited for the study. Medical history, as reported by the parents, was unremarkable. The four inclusion criteria used for the patient group also were used for this group.

An informed consent was obtained from the parents of all participants, and the study was approved by the Institutional Review Board of the Schneider CMCI.

Auditory event-related potential recordings

  1. Top of page
  2. Abstract
  3. METHODS
  4. Participants
  5. Patient group
  6. Control group
  7. Auditory event-related potential recordings
  8. Auditory stimuli
  9. Procedure
  10. Data analysis
  11. RESULTS
  12. DISCUSSION
  13. Acknowledgments
  14. REFERENCES

Brain electrical activity was recorded from 11 sites on the scalp by using electrocap tin electrodes. The electrodes were placed according to the 10-20 system (16) at the following locations: F3, Fz, F4, T3, C3, Cz, C4, T4, P3, Pz, and P4, referenced to the chin. The impedance measured for each electrode was <5 kOhm. A ground electrode was placed on the right mastoid process. Eye movements were monitored by electrodes placed above and below the right eye.

Brain electrical activity was sampled with a Brain Performance Measurement (BPM) System (Orgil) at 256 Hz. The resolution of the 12-bit analog-to-digital converter was 0.1 μV per bit, and the system bandpass was 0.1– 100 Hz. A correction procedure for eye-movement artifacts was performed offline by using the EMC (Eye Movement Correction) algorithm of the system. The procedure identifies and corrects ocular artifacts that contaminate AERP recordings. It includes detection of an epoch in which an eye-movement artifact contaminated a single AERP trial, and correction for each epoch and each channel separately, based on the eye-artifact potential as recorded at the electrode below the eye. Estimation and correction of the contamination are performed by linear regression between each electrode within the detected epoch and the electrode below the eye. In addition to AERP recordings, reaction time was measured in a separate channel online, and performance accuracy was calculated offline.

Auditory stimuli

  1. Top of page
  2. Abstract
  3. METHODS
  4. Participants
  5. Patient group
  6. Control group
  7. Auditory event-related potential recordings
  8. Auditory stimuli
  9. Procedure
  10. Data analysis
  11. RESULTS
  12. DISCUSSION
  13. Acknowledgments
  14. REFERENCES

Four sets of stimuli were used: one set of tonal stimuli, two sets of “phonetic” stimuli (i.e., differed by their phonetic features), and one set of semantic stimuli. All sets of stimuli were presented in the “oddball” paradigm. The tonal set consisted of 1,400- and 1,000-Hz pure tones with a duration of 285 ms and rise/decay time of 30 ms. The phonetic stimuli consisted of nonmeaningful consonant–vowel–consonant (CVC) monosyllabic words. One set of stimuli (/bal/vs./bav/) differed by two significant speech contrasts (place of articulation and manner in final position) and was therefore considered the “easy” phonetic task. The second set of phonetic stimuli (/bam/vs./ban/) differed by only one significant speech contrast (place of articulation in final position) and was thus considered the “difficult” phonetic task. These stimuli were chosen to evaluate the effect of acoustic complexity (i.e., stimuli representing the same linguistic level, but differing in the composition of acoustic cues). The assumption was that listeners had to use acoustic–phonetic information that was available to them at the final phoneme of the word, before making their decision.

The semantic set of stimuli included six meaningful monosyllabic CVC words from two semantic categories: Hebrew boys' names [/Gal/,/Dan/,/Ron/] and body parts [/gav/(back),/dam/(blood),/rosh/(head)]. These words were found to have a similar frequency of occurrence as ranked by 50 Hebrew-speaking adults. The subject's task was to listen to a series of stimuli that consisted of words from both categories but to respond to stimuli from only one targeted category. Therefore, the structure of the stimuli was such that listeners always had to use acoustic–phonetic information derived from the final phoneme before making a semantic decision. Furthermore, final phonemes of the stimuli differed by one or two speech contrasts between semantic categories. These were the same differences on which listeners based their decisions in the phonetic tasks, thus allowing a more controlled evaluation of the effect of linguistic complexity [i.e., stimuli representing different linguistic levels (phonetic vs. semantic)] on AERPs. In addition, stimuli of the phonetic and semantic tasks included voiced plosives (/b/,/d/, and /g/) of equal length in the initial position, and the vowel /a/ to minimize coarticulation effects.

All speech stimuli were produced by an adult female, a native Hebrew speaker. They were digitally recorded at 44-kHz sampling rate and 16-bit quantization by using a commercial speech-analysis program (Sensimetric Speech Station Ver.2.1, Ariel Corporation). Stimuli were controlled for intensity and duration of the vowels and consonants. Duration of all speech stimuli was 500 ms.

The tonal and phonetic sets consisted of 200 stimuli with probabilities of target and nontarget stimuli set at 20 and 80%, respectively. The semantic set included 240 stimuli with probabilities of target and nontarget set at 25 and 75%, respectively. The contrasting stimuli /Ron/ and /rosh/ were included in the semantic task to increase the number of different items in each category. However, because their acoustic features differed significantly from those of the phonetic stimuli, their data were excluded from the averaging procedure for this task.

Stimuli were presented binaurally every 2 s at 65 dBHL by using TDH-39 calibrated earphones.

Procedure

  1. Top of page
  2. Abstract
  3. METHODS
  4. Participants
  5. Patient group
  6. Control group
  7. Auditory event-related potential recordings
  8. Auditory stimuli
  9. Procedure
  10. Data analysis
  11. RESULTS
  12. DISCUSSION
  13. Acknowledgments
  14. REFERENCES

All experimental sessions were conducted in the morning. After the administration of the neuropsychological tests, electrodes were applied, and subjects were seated in a comfortable armchair in a quiet room. Subjects were instructed to respond as quickly as possible to the infrequent events (targets) by pressing a button with their right index finger. Presentation order of the different sets of stimuli was counterbalanced across subjects to prevent the effect of fatigue on the experimental measures. Brief intermissions were provided between stimulus sets. Subjects were instructed to fixate on a colored spot located on the wall in front of them while listening to the prerecorded stimuli, to avoid excessive eye movements. Each experimental session lasted 2 h.

Data analysis

  1. Top of page
  2. Abstract
  3. METHODS
  4. Participants
  5. Patient group
  6. Control group
  7. Auditory event-related potential recordings
  8. Auditory stimuli
  9. Procedure
  10. Data analysis
  11. RESULTS
  12. DISCUSSION
  13. Acknowledgments
  14. REFERENCES

Averaged AERPs waveform peaks were identified for each subject and for each experimental condition, based on the order of their appearance: N1 was defined as the first largest negative peak; N2, as the second largest negative peak after P2; and P3 was defined as the largest positive peak after N2. Peak latency was defined as the time (in milliseconds) from stimulus onset to the point of the most negative or positive peak in a typical time window for each of the tasks: N1 at 82–324 ms, N2 at 160–718 ms, and P3 at 305–1,093 ms. Amplitude was defined as the voltage difference between the peak and baseline (average of the voltage at a prestimulus interval of 200 ms). One-way multivariate analysis of variance (MANOVA) for repeated measures evaluating between-group factor (healthy–AS–GTCS) with four levels of task (tonal, phonetic “easy,” phonetic “difficult,” semantic), three levels of electrode laterality (right, midline, left), and three levels of electrode frontality (frontal, central, parietal) were used on the latency and amplitude of the AERPs.

Reaction time was defined as the time (in milliseconds) from stimulus onset to the subject's button press. The effects of group and task on reaction time and performance accuracy were evaluated by using one-way MANOVA for repeated measures. Level of significance was set at p < 0.05.

RESULTS

  1. Top of page
  2. Abstract
  3. METHODS
  4. Participants
  5. Patient group
  6. Control group
  7. Auditory event-related potential recordings
  8. Auditory stimuli
  9. Procedure
  10. Data analysis
  11. RESULTS
  12. DISCUSSION
  13. Acknowledgments
  14. REFERENCES

AERPs of children with epilepsy compared with those of healthy cohorts varied significantly as the acoustic and linguistic complexity of the stimuli increased. However, reaction time and performance accuracy did not differ significantly among the study groups.

Reaction time and performance accuracy

Table 2 shows the reaction time and performance accuracy in the different tasks for each group of children. No significant differences were found among groups. A significant main effect of task on reaction time was found [F(3, 39) = 176.7, p = 0.0001]. The reaction times in all linguistic tasks (phonetic and semantic) were significantly longer compared with those in the nonlinguistic tonal task (p = 0.0001). The reaction times of the two phonetic tasks were not different. However, reaction time in the semantic task was significantly longer than that in the phonetic tasks (p = 0.0001). The interaction between task and group was not significant.

Table 2. Mean reaction times and performance accuracy of the healthy, absence seizure, and generalized tonic–clonic seizure groups in the different tasks
MeasureTonal Mean (SEM)Phonetic “easy” Mean (SEM)Phonetic “difficult” Mean (SEM)Semantic Mean (SEM)
  1. AS, absence seizures; GTCS, generalized tonic–clonic seizure.

Reaction time (ms)437.3 (18.8)   
 Healthy467.4 (37.5)716.3 (21.1)704.6 (18.1)835.2 (24.1)
 AS435 (28.3)754.5 (33.5)725.1 (25.1)  835 (29.1)
 GTCS 721.6 (16.4)709.5 (17.5)847.3 (20.3)
Performance accuracy (%)    
 Healthy 97.5 (1.1) 99.1 (0.2)  99 (0.2)98.2 (0.6)
 AS 97.3 (1.8) 96.8 (1.1)97.4 (1.1)94.6 (2.4)
 GTCS 99.5 (0.4) 99.4 (0.3)98.3 (0.8)98.4 (0.3)

Performance accuracy results (Table 2) were considerably high for healthy children and children with epilepsy, and the main effect of task was nonsignificant.

Auditory event-related potential

Latency

The main effect of task on all AERPs components was significant [F(3, 39) = 165.8, 111.6, and 571.4 for N1, N2, and P3, respectively; p = 0.0001]. Specifically, AERPs components elicited by the nonlinguistic tonal task appeared significantly earlier than those elicited by the linguistic tasks. The mean latencies of N1, N2, and P3 elicited by the different tasks in the healthy, AS, and GTCS groups are shown in Table 3. Figures 1–4 show the grand average waveforms of each of the study groups elicited by the tonal, phonetic “easy,” phonetic “difficult,” and semantic tasks, respectively.

Table 3. Mean latencies of N1, N2, and P3 of the healthy, absence seizure, and generalized tonic–clonic seizure groups in the different tasks
 HealthyASGTCS
TaskMean (SEM)Mean (SEM)Mean (SEM)
  1. Values expressed in milliseconds (SEM). Peak latencies were measured at the midline electrodes where amplitude was maximal (N1 at Cz, N2 at Fz, and P3 at Pz).

  2. ap < 0.05.

  3. bp < 0.01.

Tonal   
 N1103.9 (3.5)116.9 (8.0) 117.2 (6.6) 
 N2246.3 (3.8)254.6 (6.2) 244.5 (4.4) 
 P3 378.1 (10.7)366.5 (12.3)396.2 (41.6)
Phonetic “easy”   
 N1185.1 (5.1)186.1 (6.3) 208.3 (4.0) 
 N2 416.2 (12.7)411.6 (18.9)397.8 (13.1)
 P3642.7 (8.7)665.5 (17.2)688.9 (34.3)
Phonetic “difficult”   
 N1185.7 (4.0)184.3 (4.1) 191.1 (6.3) 
 N2 409.6 (12.6)421.8 (23.2) 468.4 (32.1)a
 P3 624.9 (10.1)648.4 (20.8) 749.3 (49.1)b
Semantic   
 N1188.4 (7.4)194.6 (8.0)  225.6 (9.9)a
 N2455.8 (9.7)415.4 (20.4) 482.7 (20.9)a
 P3736.3 (9.8) 855.8 (52.2)a 833.9 (43.6)a
image

Figure 1. Superimposition of the target grand mean waveforms elicited by the tonal task in the healthy (black thin line), absence seizure (gray line), and generalized tonic–clonic seizure (black thick line) groups at all recording electrodes and electrooculogram channel (X2). Also presented is an enlargement of the waveform recorded at Pz including labeling of N1, N2, and P3. Upward arrow, stimulus onset (0 ms).

Download figure to PowerPoint

image

Figure 2. Superimposition of the target grand mean waveforms elicited by the phonetic “easy” task in the healthy (black thin line), absence seizure (gray line), and generalized tonic–clonic seizure (black thick line) groups at all recording electrodes and electrooculogram channel (X2). Also presented is an enlargement of the waveform recorded at Pz, including labeling of N1, N2, and P3. Upward arrow, stimulus onset (0 ms).

Download figure to PowerPoint

image

Figure 3. Superimposition of the target grand mean waveforms elicited by the phonetic “difficult” task in the healthy (black thin line), absence seizure (gray line), and generalized tonic–clonic seizure (black thick line) groups at all recording electrodes and electrooculogram channel (X2). Also presented is an enlargement of the waveform recorded at Pz, including labeling of N1, N2, and P3. Upward arrow, stimulus onset (0 ms).

Download figure to PowerPoint

image

Figure 4. Superimposition of the target grand mean waveforms elicited by the “semantic” task in the healthy (black thin line), absence seizure (AS, gray line), and generalized tonic–clonic seizure (black thick line) groups at all recording electrodes and electrooculogram channel (X2). Also presented is an enlargement of the waveform recorded at Pz, including labeling of N1, N2, and P3. Upward arrow, stimulus onset (0 ms). Note that although P3 latency prolongation in the AS group is not prominent at the parietal electrodes, individual data as well as statistical analysis indicated a significant prolongation.

Download figure to PowerPoint

The main effect of group on P3 and the task × group interaction were significant [F(2, 41) = 4.2, p = 0.02; F(6, 78) = 2.4, p = 0.037, respectively]. Specifically, P3 latencies elicited by the nonlinguistic tonal task were not significantly different; however, those elicited by the linguistic tasks were found to differ significantly as the acoustic and linguistic complexity of the tasks increased. In the phonetic “easy” task, nonsignificant latency differences among groups were found. However, in the phonetic “difficult” task, P3 latencies of the GTCS group were significantly prolonged compared with those of the healthy and AS groups [F(2, 41) = 5.97, p = 0.005]. In the semantic task, a significant P3 latency prolongation was evident in both epilepsy groups compared with the healthy group [F(2, 41) = 4.2, p = 0.02].

A significant main effect of Group on N1 latency [F(2, 41) = 4.7, p = 0.01] and a marginally significant main effect on N2 latency [F(2, 41) = 2.9, p = 0.06] were found. Specifically, in the GTCS group, N2 latency was prolonged in the phonetic “difficult” task, and both N1 and N2 latencies were prolonged in the semantic task. In the AS group, N1 and N2 latencies did not differ from those of the healthy group.

Amplitude

The mean amplitudes of N1, N2, and P3 elicited by the different tasks in the healthy, AS, and GTCS groups are shown in Table 4. The main effect of group on P3 was not significant. However, a significant main effect of task [F(3, 39) = 13.2, p = 0.0001] indicated greater P3 amplitudes in the tonal and phonetic “easy” tasks compared with the amplitudes in the phonetic “difficult” and semantic tasks. Thus a progressive decrease was evident in P3 amplitude as acoustic and linguistic complexity increased. No significant interaction was found between group and task.

Table 4. Mean amplitudes in microvolts of N1, N2, and P3 of the healthy, absence seizure, and generalized tonic–clonic seizure groups in the different tasks
 HealthyAS MeanGTCS Mean
TaskMean (SEM)(SEM)(SEM)
  1. Amplitudes were measured at the midline electrodes where values were maximal (N1 at Cz, N2 at Fz, and P3 at Pz).

  2. ap < 0.05.

Tonal
 N1−7.85 (1.36)−8.81 (2.54) −6.81 (1.52)
 N2−8.82 (1.28)−10.17 (2.45)  −8.67 (1.93)
 P315.76 (1.14)18.35 (1.66)  17.0 (2.04)
Phonetic “easy”
 N1−8.39 (1.12)−7.70 (1.48) −10.35 (1.08) 
 N2−10.68 (0.89) −11.20 (1.82)  −7.40 (1.84)
 P314.44 (1.22)13.83 (2.40) 15.28 (1.72)
Phonetic “difficult”
 N1−7.79 (1.41)−8.93 (2.07) −8.76 (2.26)
 N2−9.95 (1.07)−15.83 (1.31)a−9.23 (1.95)
 P312.39 (1.18)11.72 (1.73) 11.22 (1.58)
Semantic
 N1−6.38 (1.00)−7.22 (1.20) −9.77 (2.03)
 N2−7.96 (0.91)−13.3 (1.28)a−9.14 (1.85)
 P311.94 (0.96)11.53 (1.16) 10.16 (1.41)

The main effect of group on N2 amplitude was significant [F(2, 41) = 4.0, p = 0.02] and showed that in the AS group, N2 amplitudes were significantly enhanced in the phonetic “difficult” and semantic tasks compared with those of the healthy and GTCS groups. Figure 5 shows N2 amplitudes at the frontal electrodes (F3, Fz, F4) of each of the study groups in the phonetic “difficult” and semantic tasks.

image

Figure 5. N2 amplitude (+SEM) at the frontal electrodes (F3, Fz, F4) of the healthy, absence seizure, and generalized tonic–clonic seizure groups in the (A) phonetic “difficult” task, and (B) semantic task.

Download figure to PowerPoint

The main effects of group and task on N1, as well as the group × task interaction, were not significant.

Scalp distribution

P3 laterality patterns elicited by the linguistic tasks were significantly different in children with epilepsy compared with the healthy cohorts. A significant main effect of laterality [F(2, 40) = 18.1, p = 0.0001] as well as a significant laterality × task × group interaction [F(12, 240) = 1.9, p = 0.03] was found. In the healthy group, linguistic processing was characterized by significantly greater P3 amplitudes over the left compared with right scalp. In the epilepsy groups, however, no significant differences were found between amplitudes over the left and right scalp. To present this interaction, the P3 laterality ratio was calculated by dividing P3 amplitude at the P3 electrode (left) by that at the P4 electrode (right). Figure 6 shows P3 laterality ratio in the different tasks in each of the study groups. As can be seen, in the healthy group, ratios in the linguistic tasks are significantly greater than those in the nonlinguistic task. In the epilepsy groups, ratios were reduced and were similar in the nonlinguistic and linguistic tasks.

image

Figure 6. P3 laterality ratio in the healthy, absence seizure, and generalized tonic–clonic seizure groups in the different tasks. The laterality ratio was calculated by dividing P3 amplitude at the P3 electrode (left) by that at the P4 electrode (right).

Download figure to PowerPoint

A significant main effect of electrode frontality on N2 [F(2, 40) = 53.9, p = 0.0001] and P3 [F(2, 40) = 130.1, p = 0.0001] amplitudes was found. No significant frontality × group interaction occurred. In all groups, scalp distribution was centroparietal for N1, frontocentral for N2, and parietal for P3.

DISCUSSION

  1. Top of page
  2. Abstract
  3. METHODS
  4. Participants
  5. Patient group
  6. Control group
  7. Auditory event-related potential recordings
  8. Auditory stimuli
  9. Procedure
  10. Data analysis
  11. RESULTS
  12. DISCUSSION
  13. Acknowledgments
  14. REFERENCES

AERPs elicited by linguistic stimuli of children with IGE differed significantly from those of healthy controls and were affected by seizure type (i.e., GTCS, AS). Specifically, as the acoustic and linguistic complexity of the eliciting stimulus increased, significant differences in AERP latency, amplitude, and scalp distribution between healthy children and those with epilepsy were evident. In agreement with previous reports (5–10) AERPs elicited by nonlinguistic stimuli (i.e., pure tones) were found to be similar among the study groups. This is the first report in which linguistic stimuli were used to elicit AERPs in children with IGE.

Auditory event-related potentials during phonetic processing

Only children with GTCSs required longer processing time as the acoustic complexity increased. The discrimination between stimuli that differed by one phonetic contrast only (phonetic “difficult” task) was associated with significant prolongation of N2 and P3 compared with that in healthy controls. N2 and P3 latencies are assumed to reflect stimulus-evaluation time, with N2 reflecting early stages of stimulus evaluation and classification (1), and P3 indicating the closure or completion of that processing (3,4). Thus these results suggest thatchildren with GTCSs process auditory phonetic information slower than do healthy children.

The finding of a shorter processing time for “easy” compared with “difficult” stimuli has theoretic implications regarding the underlying theory of phonetic processing. It has been suggested that discrimination between stimuli that differ by two speech contrasts is easier than the discrimination of stimuli that differ by one contrast because greater acoustic separation allows easier detectability and recognition (17–19). One could argue, however, that processing additional information may be time-consuming and thus increase the time required for the listener to reach a decision. The results of the present study support the first of the two hypotheses. Similar to hearing-impaired (17) and aphasic patients (18), children with GTCSs take advantage of salient acoustic information to reach linguistic decisions.

The finding that children with GTSCs showed prolongation of N2 and P3 latency but not of N1 suggests that the transmission of auditory phonetic information and its “arrival” at the primary auditory cortex was intact in this group of children. This is based on the evidence that N1 is a mixed component of mostly exogenous with some attention and complexity-sensitive contributions (20,21). Because all phonetic stimuli used in the present study had identical initial consonant–vowel (/ba/) but differed only in their final consonant, N1 latency did not appear to be affected. The significant prolongation of the remaining waveform components, however, suggests that perceptual processing was impaired at more central levels. This assumption can be further explored in studies with earlier potentials such as auditory brain stem response (ABR) and auditory middle latency response (AMLR), which measure sound transmission at even lower-level sites along the auditory pathway.

The present study showed different auditory linguistic-processing patterns in children with ASs compared with GTCSs. The AERPs latencies of children with ASs were similar to those of healthy children; nevertheless, a significant enhancement of N2 amplitude was evident in the phonetic “difficult” task. Similar results were reported for children with attention-deficit hyperactive disorder (ADHD) (22). This result may be related to the fact that children with epilepsy, irrespective of type, have a form of attention disorder related to impaired alertness (23,24). Evidence for an attention deficit originating in the thalamic or early cortical levels may be derived from the fact that the cortex is believed to be the main generator of the three-per-second spike-and-wave paroxysmal activity with a special role attributed to thalamocortical oscillating circuits (25). Event-related potential data from healthy subjects indicate that selective listening exerts its control on auditory processing at thalamic or early cortical levels, and support early selection of inputs before perceptual processing is complete (26). We therefore assume that the larger N2 amplitudes found in children with AS may reflect enhanced allocation of attentional resources serving as a functional strategy for sustaining accurate task performance. Further support for this notion was the poorer performance of these children compared with healthy controls, in verbal and nonverbal attention tasks (data not shown).

Auditory event-related potentials during semantic processing

Processing time was found to be longer for semantic compared with phonetic stimuli in both healthy children and those with epilepsy. This prolongation may be explained by the complexity of the task (19) and the involvement of additional associative brain areas that are activated during semantic, but not during phonetic processing, as demonstrated with positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) (27,28).

Our results indicate that processing time of semantic information was significantly prolonged in children with IGE compared with that in healthy controls. Specifically, in children with GTCSs, a significant prolongation of all AERPs suggests a semantic processing deficit that is evident at the initial stages of processing (i.e., N1). We assume that the higher complexity of the semantic task (compared with that of the phonetic tasks), as well as evidence of N1 being a mixed component with some attention- and complexity-sensitive contributions (20,21), might be the cause for the prolongation starting at N1. A significant prolongation of N1 also was reported in a group of children with severe language impairment of unknown etiology (29). In contrast, in children with ASs, only P3 was significantly prolonged, suggesting a differential semantic deficit that affected the later stages of perceptual processing. In addition, the significantly greater N2 amplitudes that are assumed to reflect enhanced allocation of attentional resources, found in children with ASs but not in those with GTCSs, provide further evidence for a different semantic deficit.

The relatively few studies that addressed the issue of cognitive function in well-defined homogeneous groups of children with IGE, and the lack of studies regarding linguistic processing, make it difficult to draw further conclusions as to the origin of these deficits. Nevertheless, a possible explanation may be inter- and intrahemispheric functional reorganization of language representation that also has been suggested for patients with a left temporal epileptic focus and demonstrated with fMRI (30). Clearly this issue requires further investigation.

Further evidence of different patterns of language organization in children with IGE can be found in the P3-amplitude data. In healthy children the nonlinguistic task elicited similar P3 amplitudes over the left and right scalp, whereas the linguistic tasks (phonetic and semantic) elicited significantly larger amplitudes over the left scalp. These results are compatible with enhanced synchronous activity over the left hemisphere that is more involved in linguistic processing, as shown in studies using AERPs (31,32), PET (27,33), and fMRI (34). In children with IGE, however, nonsignificant differences were found between P3 amplitudes over the right and left scalp during nonlinguistic and linguistic processing. These different patterns of lateralization may be the reflection of a nonsuppressed right hemisphere evolving from the generalized hyperexcitability of cortical neurons present in generalized epilepsy, particularly associated with three-per-second spike–wave paroxysms (25). Overactivation or dysinhibition may interfere with normal, predominantly left hemisphere language processing in IGE. If the activity of one hemisphere is not suppressed, competing information may be available from the interaction of processing in the two hemispheres, resulting in delays and misidentifications (35). Clearly, additional studies are needed with both AERPs and fMRI during linguistic processing in IGE.

Reaction time and performance accuracy

The finding that performance accuracy and reaction time did not differ significantly among the groups suggests that these tasks may not have been sensitive enough to reveal auditory-processing deficits. Furthermore, because reaction time reflects the full time span between stimulus and response, whereas AERPs reflect fewer stages related to processing, reaction-time measures have a larger variability and thus may not show significant effects as AERPs do, with their smaller variability. It also is possible thatepilepsy subjects allocated greater mental effort to achieve performances similar to those of controls, and consequently reaction time and performance accuracy did not differ among groups. In line with the results of the present study, auditory reaction time (simple and binary choice) was not significantly different in children with epilepsy receiving VPA compared with that in controls (36,37). Significant prolongation of reaction time was present only in tasks with greater complexity. For example, Brandt (38) reported that patients with IGE performed similarly to healthy subjects in a binaural continuous performance test (CPT), whereas in the dichotic version of the test that requires attending only to the left or right ear, significantly inferior performance was evident in both ears. It is possible that by increasing the difficulty of the tasks such as limiting the response time allocated to the listeners, adding background noise or competing stimuli may show differences between the groups. This clearly requires further investigation.

In summary, with AERPs, linguistic-processing deficits in children with IGE were exposed and could be related to seizure type. Specifically, phonetic processing was characterized by (a) slower processing time, manifested by prolonged N2 and P3 latencies, in children with GTCSs; and (b) increased allocation of attentional resources, manifested by enhanced N2 amplitudes, in children with ASs. Semantic processing was characterized by (a) slower processing time, manifested by the prolongation of all AERPs in GTCSs, and only P3 latency in ASs; and (b) increased allocation of attentional resources in children with ASs. In addition, processing of linguistic stimuli (phonetic and semantic) resulted in different patterns of brain-activity lateralization in IGE (GTCSs and ASs) compared with healthy controls. The behavioral measures (reaction time and performance accuracy) that were used in the present study did not provide satisfactory insight into the events occurring during linguistic processing. Although AERPs differences between groups were found only for the linguistic tasks, additional studies using complex nonlinguistic stimuli are required to support a language-specific deficit rather than a general auditory deficit.

Finally, the results of the present study suggest that children with IGE may have average intelligence and age-appropriate academic achievements despite linguistic-processing deficits. Although these results indicate statistically significant differences among groups, they cannot be applied to individual patients. It is apparent from this study that neurologic follow-up of such children should include evaluation of linguistic functions to provide appropriate remedial intervention when needed.

Acknowledgments

  1. Top of page
  2. Abstract
  3. METHODS
  4. Participants
  5. Patient group
  6. Control group
  7. Auditory event-related potential recordings
  8. Auditory stimuli
  9. Procedure
  10. Data analysis
  11. RESULTS
  12. DISCUSSION
  13. Acknowledgments
  14. REFERENCES

Acknowledgment:  This research was funded by the Schneider Fund and Tel Aviv University. We thank the subjects and their families for their participation, Esther Shabtay for the statistical analysis, and engineer Emma Sachartov and Dr. Shlomo Gilat for the technical assistance.

REFERENCES

  1. Top of page
  2. Abstract
  3. METHODS
  4. Participants
  5. Patient group
  6. Control group
  7. Auditory event-related potential recordings
  8. Auditory stimuli
  9. Procedure
  10. Data analysis
  11. RESULTS
  12. DISCUSSION
  13. Acknowledgments
  14. REFERENCES
  • 1
    Hillyard SA, Picton TW. Electrophysiology of cognition. In: PlumF, ed. Handbook of physiology, Section 1: the nervous system: higher functions of the nervous system. Bethesda , MD : American Physiological Society, 1987: 51984.
  • 2
    Hillyard SA, Kutas M. Electrophysiology of cognitive processing. Annu Rev Psychol 1983;34: 3361.
  • 3
    Picton TW, Hillyard SA. Endogenous event-related potentials. In: PictonTW, ed. Handbook of electroencephalography and clinical neurophysiology: human event-related potentials. Amsterdam : Elsevier Science Publications, 1988: 361426.
  • 4
    Kutas M, Dale A. Electrical and magnetic readings of mental functions. In: RuggMD, ed. Cognitive neuroscience. Sussex, United Kingdom : Psychology Press, 1997: 197242.
  • 5
    Drake ME Jr, Bugess RJ, Gelety TJ, et al. Long-latency auditory event-related potentials in epilepsy. Clin Electroencephalogr 1986;17: 103.
  • 6
    Rodin E, Khabbazeh Z, Twitty G, et al. The cognitive evoked potential in epilepsy patients. Clin Electroencephalogr 1989;20: 17682.
  • 7
    Fukai M, Motomura N, Kobayashi S, et al. Event-related potential (P300) in epilepsy. Acta Neurol Scand 1990;82: 197202.
  • 8
    Triantafyllo NI, Zalonis I, Kokotis P. Cognition in epilepsy: a multi-channel event related potential (P300) study. Acta Neurol Scand 1992;86: 4625.
  • 9
    Soysal A, Ataki D, Atay T, et al. Auditory event related potentials (P300) in partial and generalized epileptic patients. Seizure 1999;8: 10710.
  • 10
    Konishi T, Naganuma Y, Hongou K, et al. Changes of P300 latency with age in childhood epilepsy. Pediatr Neurol 1995;12: 1325.
  • 11
    Chen JY, Kang WM, So WMK. Comparison of antiepileptic drugs on cognitive function in newly diagnosed epileptic children: a psychometric and neurophysiological study. Epilepsia 1996;37: 816.
  • 12
    Binnie CD. Cognitive performance, subtle seizures, and the EEG. Epilepsia 2001;42: 168.
  • 13
    Trimble MR. Antiepileptic drugs, cognitive function and behavior in children: evidence from recent studies. Epilepsia 1990;31: S304.
  • 14
    Devinsky O. Cognitive and behavioral effects of antiepileptic drugs. Epilepsia 1995;36: S4665.
  • 15
    Oldfield RC. The assessment and analysis of handedness: the Edinbourgh inventory. Neuropsychologia 1971;9: 97113.
  • 16
    Jasper HH. The ten-twenty electrode system of the International Federation. Electroencephalogr Clin Neurophysiol 1958;10: 3715.
  • 17
    Boothroyd A. Auditory perception of speech contrasts by subjects with sensorineural hearing loss. J Speech Hear Res 1984;27: 13444.
  • 18
    Blumstein SE. The neurobiology of the sound structure of language. In: GazzanigaMS, ed. The cognitive neurosciences. Cambridge, Mass. : MIT Press, 1995: 91529.
  • 19
    Henkin Y, Kishon-Rabin L, Gadoth N, et al. Auditory event related potentials during phonetic and semantic processing in children. Audiol Neuro-Otol 2002;7: 22839.
  • 20
    Näätänen R, Picton TW. The N1 wave of the human electric and magnetic response to sound: a review and an analysis of the component structure. Psychophysiology 1987;24: 375425.
  • 21
    Hyde M. The N1 response and its applications. Audiol Neuro-Otol 1997;2: 281307.
  • 22
    Satterfield JH, Schell AM, Nicholas TW, et al. Ontogeny of selective attention effects on event related potentials in attention-deficit hyperactivity disorder and normal boys. Biol Psychiatry 1990;28: 879903.
  • 23
    Stores G, Williams PL, Styles E, et al. Psychological effects of sodium valproate and carbamazepine in epilepsy. Arch Dis Child 1992;67: 13307.
  • 24
    Zametkin AJ, Ernst M. Problems in the management of attention-deficit-hyperactivity disorder. N Engl J Med 1999;340: 406.
  • 25
    Niedermeyer E. Primary (idiopathic) generalized epilepsy and underlying mechanisms. Clin Electroencephalogr 1996;27: 121.
  • 26
    Hillyard SA, Mangun GR, Woldorff MG, et al. Neural systems mediating selective attention. In: GazzanigaMS, ed. The cognitive neurosciences. Cambridge, Mass. : MIT Press, 1995: 66581.
  • 27
    Demonet JF, Chollet F, Ramsay S, et al. The anatomy of phonological and semantic processing in normal subjects. Brain 1992;115: 175368.
  • 28
    Poldrack RA, Wagner AD, Prull MW, et al. Functional specialization for semantic and phonological processing in the left inferior prefrontal cortex. Neuroimage 1999;10: 1535.
  • 29
    Tonnquist-Uhlen I. Topography of auditory evoked cortical potentials in children with severe language impairment. Scand Audiol 1996;116: 6809.
  • 30
    Billingsley RL, McAndrews MP, Crawley AP, et al. Functional MRI of phonological and semantic processing in temporal lobe epilepsy. Brain 2001;124: 121827.
  • 31
    Ahonniska J, Cantell M, Tolvanen A, et al. Speech perception and brain laterality: the effect of ear advantage on auditory event-related potentials. Brain Lang 1993;45: 12746.
  • 32
    Kayser J, Tenke CE, Bruder GE. Dissociation of brain ERP topographies for tonal and phonetic oddball tasks. Psychophysiology 1998;35: 57690.
  • 33
    Zattore RJ, Evans AC, Meyer E, et al. Lateralization of phonetic and pitch discrimination in speech processing. Science 1992;256: 8469.
  • 34
    Kent RD. Neuroimaging studies of brain activation for language, with an emphasis on functional magnetic resonance imaging: a review. Folia Phoniatr Logop 1998;50: 291304.
  • 35
    Brunswick N, Rippon G. Auditory event-related potentials, dichotic listening performance and handedness as indices of lateralization in dyslexic and normal readers. Int J Psychophysiol 1994;18: 26575.
  • 36
    Blennow G, Heijbel J, Sandstedt P, et al. Discontinuation of antiepileptic drugs in children who have outgrown epilepsy: effects on cognitive function. Epilepsia 1990;31: S503.
  • 37
    Alpherts WCJ, Aldenkamp AP. Computerized neuropsychological assessment of cognitive functioning in children with epilepsy. Epilepsia 1990;31: S3540.
  • 38
    Brant J. Defective stimulus set attention in generalized epilepsy. Brain Cogn 1984;2: 14051.