The mismatch negativity (MMN) potential as a tool for the functional mapping of temporal lobe epilepsies

Epilepsy & Behavior 33 (2014) 87–93

Contents lists available at ScienceDirect

Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh

The mismatch negativity (MMN) potential as a tool for the functional mapping of temporal lobe epilepsies Ricardo Lopes a,⁎, Mário R. Simões a, Luís Ferraz b, Alberto J.R. Leal b,c a b c

Faculty of Psychology and Education Sciences, University of Coimbra, Rua do Colégio Novo, Apartado 6153, 3001-802 Coimbra, Portugal Instituto Universitário de Lisboa (ISCTE-IUL), Cis-IUL, Lisbon, Portugal Department of Neurophysiology, Centro Hospitalar Psiquiátrico de Lisboa, Lisbon, Portugal

a r t i c l e

i n f o

Article history: Received 18 November 2013 Revised 10 February 2014 Accepted 13 February 2014 Available online 13 March 2014 Keywords: Pediatric epilepsy MMN Auditory evoked potentials EEG Temporal lobe epilepsy

a b s t r a c t Temporal lobe epilepsies are associated with cognitive dysfunctions in memory which are important clues currently used clinically for the lateralization of the epileptic focus in evaluations for epilepsy surgery. Because these lobes also contain the primary auditory cortex, the study of auditory evoked potentials (AEPs) is a candidate, not yet established, complementary method to characterize epilepsy-induced dysfunction. We aimed to establish the clinical usefulness of auditory evoked potentials for the study of pediatric symptomatic temporal lobe epilepsies. A group of 17 patients (ages 4–16) with symptomatic epilepsies undergoing evaluation for epilepsy surgery epilepsy was submitted to auditory evoked potentials using 35-channel scalp EEG recordings. A control group of 10 healthy volunteers was studied with the same protocol. The P100 and mismatch negativity (MMN) potential latencies and normalized amplitudes were studied. We also performed a voxel-based lesion-symptom mapping (VLSM) to determine the anatomical areas associated with changes in the AEPs. Eleven patients had temporal lobe epilepsy, three had frontal lobe epilepsy, and three had occipital lobe epilepsy. Latencies for the P100 were normal in 15/17 and in 11/17 for the MMN, with no consistent correlation with the epilepsy type. The MMN amplitude was abnormal in 7/17 patients, all with temporal lobe epilepsies (sensitivity of 64%). Of these patients, four had a decreased MMN associated with a Heschl's gyrus lesion in the VLSM, and three had an increased MMN associated with hippocampal lesion. No extratemporal epilepsy showed MMN amplitude abnormalities (specificity of 100%). The P100 amplitude was abnormal in 3/17, two with temporal and one with frontal lobe epilepsies. The auditory MMN has a high specificity but a low sensitivity for temporal lobe epilepsy in symptomatic pediatric epilepsies. Amplitude decreases of the MMN are associated with homolateral Heschl's gyrus lesions, and MMN increases with hippocampal lesions. © 2014 Elsevier Inc. All rights reserved.

1. Introduction Epilepsies are not only associated with the occurrence of spontaneous seizures but also with different degrees of cognitive impairments in a wide range of functions, depending on the particular cortical networks involved in the pathological activity [1]. When the epileptogenic area involves or is in the neighborhood of the cortical representation of a particular cognitive function, a neurological dysfunction will only become apparent when the mechanisms of remapping of such function in the remaining cortex fail to compensate. Additionally, when the disease starts at an early age, brain plasticity is at its highest degree and so is the capability of the healthy cortex to compensate for the epileptic

⁎ Corresponding author. Tel.: +351 239 851450, +351 217819808; fax: + 351 217819809. E-mail address: [email protected] (R. Lopes).

http://dx.doi.org/10.1016/j.yebeh.2014.02.017 1525-5050/© 2014 Elsevier Inc. All rights reserved.

dysfunction, leading to atypical brain representations of cognitive functions and to restoration of function [2]. In this setting, the assumption of a strong relation between a cognitive dysfunction and a specific anatomical insult fails [3]. Thus, conventional neuropsychological evaluation, based only on the determination of the performance level of the patient in a series of tests, loses both sensitivity and localizing power [4]. The handicaps of neuropsychological testing for localizing the epileptogenic area are well established [5] and are particularly important when patients are referred for epilepsy surgery. This has motivated us to develop more sensitive methods that could complement the neuropsychological evaluation of such patients, such as recording of evoked potentials of visual stimuli to map the posterior cortex in patients with symptomatic occipital lobe epilepsy [6]. In the current study, we used a similar paradigm, employing auditory evoked potentials, to gain useful information for the localization of the epileptogenic area in patients with temporal lobe epilepsy undergoing evaluation for epilepsy surgery.

88

R. Lopes et al. / Epilepsy & Behavior 33 (2014) 87–93

The mismatch negativity (MMN) [7] is an endogenous auditory evoked potential that can be elicited by a discriminated change in an auditory stimulus. The major MMN generators are thought to be located in the superior temporal cortex near the primary auditory cortex [8,9]. The fact that MMN can be elicited independently from the attentional state of the subject [10] makes this ERP especially appropriate for clinical populations with limited cooperation such as children [11] and patients with schizophrenia [12] or dementia [13]. Moreover, the MMN seems to be quite stable throughout several developmental milestones, with no substantial latency and amplitude differences between school-age children and adults (reviewed in Cheour et al. [14]). Mismatch negativity has been used in epilepsy studies by several researchers: Boatman et al. [15] found, in a group of seven patients with benign childhood epilepsies, a prolonged MMN latency when compared with age-matched controls; Liasis et al. [16], in a similar group of patients, did not find a MMN response in the clinical group in contrast to a control group; Honbolygo et al. [17] used MMN as an index of language impairment on a 6-year-old boy with the Landau–Kleffner syndrome and found a MMN for phoneme difference but not for stress pattern difference; Gene-Cos et al. [18] compared patients with epileptic seizures, patients with nonepileptic seizures, and controls and found a different distribution and latency in the MMN elicited by tonedeviants; and Miyajima et al. [19] used MMN to investigate abnormalities in auditory processing in patients with temporal lobe epilepsies, which was delayed in this clinical group. We aimed to investigate the potential contribution of the MMN to the functional mapping of the temporal lobes in refractory focal epilepsies in a group of pediatric patients undergoing evaluation for surgery of epilepsy. The information obtained from the auditory evoked potentials was correlated with the anatomical lesions and the localization of the epileptic focus on an individual basis, allowing the determination

of the added value of the method to the decision-making process in the individual patient. 2. Methods and patient data 2.1. Patient data Seventeen patients (age range: 4 to 17 years) with symptomatic refractory focal epilepsy undergoing evaluation for epilepsy surgery were included in the present study (Table 1) after informed consent was obtained from the patient's parents. Seizures were recorded with a 35-channel long-term video-EEG complemented with a detailed brain MRI, including a high-resolution anatomical sequence. Eight of the patients underwent surgery. The large majority of patients had a structural lesion in the brain MRI (15/17). Of the remaining two, one had a focal hypometabolic area in a PET scan congruent with the EEG epileptic focus, and the other underwent successful surgery after invasive neurophysiological study. A control group of 10 healthy volunteers (age range: 4 to 16 years) was also submitted to the neurophysiological protocol. 2.2. EEG acquisition and processing Auditory evoked potentials (EPs) were obtained using a 35-channel montage including the 19 electrodes of the 10–20 system plus F11/12, F9/F10, T9/10, FC1/2, FC5/6, CP1/2, CP5/6, and P9/10 (Fig. 1a). Sintered Ag–Cl ring electrodes applied in a cap (EasyCap, Inc.) were used (Fig. 1b), with the resistance kept below 5 kΩ. The EEG was continuously sampled at 1000 Hz using a NuAmps digital amplifier (Neuroscan, Charlotte USA), with high and low pass filters of 0.05 and 70 Hz, controlled by the Scan 4.3 software (Neuroscan, Charlotte, USA).

Table 1 Patient data. Patient Sex Age Structural lesion (MRI)

Seizure onset area

Seizures

Age at Neurological onset examination

DR

F

8

Left supra sylvic lesion

Left frontotemporal area

2

ID

F

7

Left frontal lobe

JP

M

17

Left temporal lobe

MG

F

15

ACF BP DA

F M F

14 11 15

NA RM

M M

14 4

Sturge–Weber with left frontal atrophy Left thalamus plus mesial temporal lobe atrophy Right mesial temporal lobe sclerosis Normala Right first temporal gyrus Left temporal middle and lower gyrus Normalc Right perisylvian malformation

Motor seizure of right arm and face Right motor seizures

MR

M

6

GM MB

M F

14 7

Left occipital lobe Right occipital–temporal area Large inferior right Right occipital–temporal occipital–temporal lesion area Left inferior occipital lobe lesion Left occipital–temporal area Left mesial temporal lobe sclerosis Left temporal lobe

GG

M

12

Right temporal lobe lesion

DF

M

7

Lateral left occipital lesion

GB

M

8

BA

M

15

MLS

F

5

Large lower left frontal–parietal lesion Right temporal lobe neocortical lesions Left hippocampal lesion

Neuropathology

AEDsd PB, DPH

1

Low IQ and Cortical dysplasia behavioral problems Right hemiparesis

Partial complex

7

Normal

None

Right temporal lobe

Partial complex

9

Normal

Left frontotemporal area Right temporal lobe Left temporal lobe

Partial complex Partial complex Motor seizures of mouth and throat Partial complex Partial complex

1 9 8

Normal Normal Normal

DNETb

CBZ, VPA, DPH CBZ CLN, CBZ, LVT

8 3

Normal Left hemiparesis

Cortical dysplasia Cortical dysplasia

VPA VPA, VGB, LMT

Tonic head rotation to the left Partial complex Partial complex

1

Normal

7 3

Normal Normal

Partial complex with 7 left side jerking Left posterior temporal area Head right rotation and 1 visual illusions Left frontal lobe Head and trunk flexion 5

Normal

Left temporal lobe

Right temporal lobe Left temporal lobe

“Strange smell” and noise in the left side Partial complex

Normal Low IQ

Hippocampal sclerosis

LVT, CLB

LVT, CLB, VPA

CBZ, DPH Cortical dysplasia Hippocampal sclerosis

CBZ, LMT VPA CBZ

DNETb + cortical CBZ, PB, LVT dysplasia PB, DPH

8

Normal

None

3

Normal

CBZ, CLB

AED, antiepileptic drug; CBZ, carbamazepine; CLB, clobazam; CLN, clonazepam; DPH, phenytoin; LMT, lamotrigine; LVT, levetiracetam; PB, phenobarbital; VGB, vigabatrin; VPA, valproic acid. a PET focal hypometabolism in the left planum temporale. b Dysembryoplastic neuroepithelial tumor. c Seizure-free after a large inferior left occipital–temporal resection. d Antiepileptic drugs.

R. Lopes et al. / Epilepsy & Behavior 33 (2014) 87–93

89

Fig. 1. a) Butterfly plot of the 35-channel P100 (left) and MMN (right), with the mean global field power (MGFP) of a typical patient, below. b) MMN realistic standard boundary element model (BEM) with electrodes and fixed regional sources on Heschl's gyrus with EEG potential lines. c) Boxplot for both P100 (left) and MMN (right) MGFP peak latencies for the control group (N = 10), patients with extratemporal lobe epilepsies (Δ), and patients with temporal lobe epilepsies (○). d) Ratio between lesion and healthy hemisphere P100 (left) and MMN (right) for the control group and patients.

Auditory stimuli were applied through tube headphones using the Stim2 system (Neuroscan, Charlotte, USA). The patient was seated in a reclinable chair performing a visual “Where's Waldo?” distracting task. The optimal multiparametric protocol of Näätänen et al. [20] was used, with a sound intensity of 60 dB above hearing threshold, simultaneously on both ears. Three blocks of 600 stimuli were presented,

half of which were standard and the other half deviants from five different types (total of 900 standard stimuli and 180 for each deviant stimuli). The sequence was preceded by 15 standard stimuli in order to create a memory trace. The stimulus onset asynchrony was 500 ms, and each block lasted for 5.12 min. The standard harmonic tones had a 75-ms duration with a 5-ms ramp up and down and were composed

90

R. Lopes et al. / Epilepsy & Behavior 33 (2014) 87–93

by 3 different partials with 500, 1000, and 1500 Hz with a lower 3-dB and 6-dB intensity for the second and third parts, respectively. Deviant stimuli differ in five different dimensions: duration, location, intensity, gap, and frequency. Duration differed from the standard by reducing the stimuli to 25 ms long; location by using an 800-μs time interaural difference to the right channel and left channel; intensity by using an increasing or decreasing intensity of 10 dB; gap by introducing a 7-ms sound gap at the middle of the stimuli; and frequency by using a 10% higher or lower frequency partials. The EEG was visually inspected and the areas with artifact removed from further analysis. After additional high and low pass filtering at 3 and 40 Hz, respectively, epochs around (− 100 to 500 ms) the sound triggers were extracted. The P100 was obtained from the response to standard stimuli and the MMN from the difference between the response to the deviant and standard stimuli. We applied the protocol to the 10 healthy volunteers and found that the duration deviant elicited the highest signal/noise ratio and the most reliable potentials, an observation that agrees well with recent data from Hirose et al. [21], which found this deviant to be of higher amplitude and more sensitive to changes in patients with temporal lobe epilepsy. In the following, we applied the same acquisition protocol to all the patients and restricted data analysis to the MMN elicited by the duration deviant. A data reduction method similar to one previously published by our group [6] was used. Briefly, building on the knowledge of the main cortical generators of both the auditory P100 and MMN EPs [8,9,22], two regional dipole sources were placed symmetrically in the middle of the Heschl's gyrus of both hemispheres, using a standard realistic boundary element model (BEM) of the head [23] (Fig. 1b). The amplitudes of the left and right regional dipoles fitted at the P100 and MMN mean global field power (MGFP) peaks were taken as the contribution of each temporal lobe to the EPs, thus reducing the 35-channel recording to a single value per hemisphere, which was further reduced by calculating the ratio between the hemisphere with the lesion and the healthy one. A similar data reduction process was used for the P100 potential. For both latencies and interhemispheric ratios, we classified the results of the P100 and MMN potentials as pathological when they were outside the control range obtained by the ratio between the right and the left hemisphere for the control group. 2.3. Anatomical data In order to map the brain areas associated with abnormal auditory evoked potential responses, we conducted a voxel-based lesion-

symptom mapping (VLSM) on the 17 patients. The MRI lesion of each patient was manually drawn in a standard brain MRI, using the MRIcro software [24], to calculate a volume of interest (VOI). For the two patients with no identified MRI lesions, we used the focal hypometabolic volume in the PET scan in one and the surgical volume removed at surgery in the other. Because only 6 out of 17 patients had the lesion in the right hemisphere (Table 1), we decided to draw the mirror image of these lesions on the left hemisphere template, therefore merging all the anatomical data in this hemisphere (Table 2). Each patient with temporal lobe epilepsy was classified into one of three groups using the MMN on the hemisphere with the structural lesion: a group with normal MMN (Fig. 2c), a group with decreased MMN (Fig. 2a), and another group with increased MMN (Fig. 2b). For VLSM relating to changes in MMN, we used the non-parametric mapping (NPM) software [24] with binary predictors (MMN classification) and the Liebermeister statistic [25]. Only lesion voxels present in at least two patients were considered for analysis, and permutation thresholding was used for control of the multiple comparison problem [26]. The maps for both the increased and decreased MMN groups are shown in (Figs. 2d–e)(p=0.05). A similar lesion analysis could not be conducted for the P100 potential because too few patients showed abnormal results (Table 2). 3. Results 3.1. Neurophysiological data Most of our patients (15/17) had P100 latencies within the normative range (85 ms–135 ms) (Fig. 1c), failing to demonstrate a consistent delay in the processing of auditory stimuli, with a group latency average of 113 ms (Table 2, Fig. 1c). The exceptions were two patients with frontal lobe epilepsies (DR and GB) who had increased P100 latencies (Table 2). For MMN, the clinical group average was 149 ms (range of control group: 110–162 ms), with three patients with temporal lobe epilepsy (JP, MG, and ACF) as well as two patients with frontal (DR and ID) and one patient with occipital lobe epilepsy (GM), showing abnormally high latencies. These observations support a delay in the timing of processing of auditory information in 35% of patients with no specific lesion location correlation. The analysis of the P100 interhemispheric ratio, reflecting the normalized response of the lesioned hemisphere, demonstrated values outside the normal range of values (0.61–1.28) in 3 out of 17 patients (Table 2, Fig. 1d). Two of these patients had temporal lobe and one had occipital lobe epilepsies. The MMN hemispheric ratio (normal range: 0.69–1.41) demonstrated abnormal values in 7 out of 17 patients,

Table 2 Neurophysiological data. Patient

Lesion side

P100 latency (ms)

P100 L/H hemispheresa

MMN latency (ms)

RV %

MMN L/H hemispheresa

DR ID JP MG ACF BP DA NA RM MR BA GM MB GG DF GB MLS

L L L R L R L L R R R L L R L L L

139 106 123 92 113 108 110 105 114 102 128 112 86 119 88 153 115

1.01 0.62 1.20 1.19 1.72 0.93 1.16 1.15 1.00 0.91 1.01 1.06 0.95 0.42 2.56 0.83 1.06

172 163 164 174 176 133 127 116 136 140 160 173 154 121 133 149 143

18.8 17.0 11.0 10.6 9.1 5.6 19.9 8.8 10.0 22.7 13.4 7.6 9.9 9.8 25.9 16.4 11.9

1.37 1.35 2.70 1.81 0.05 0.70 0.61 0.98 0.61 1.04 0.94 1.23 0.83 0.69 0.88 0.71 1.79

a

Ratio of lesioned/healthy hemispheres. Values that are outside the control range are presented in bold.

R. Lopes et al. / Epilepsy & Behavior 33 (2014) 87–93

91

Fig. 2. a) Lesions of the four patients with decreased MMN ratio. b) Lesions of the three patients with increased MMN ratio. c) Lesions of the four patients with temporal lobe epilepsy and normal MMN ratio. d) Non-parametric mapping (NPM) for the group with reduced MMN (p = 0.05, corrected) in the superior temporal lobe. e) NPM for the group with increased MMN (p = 0.05, corrected) with lesions restricted to the hippocampus.

all of whom presented with temporal lobe epilepsy (Table 2, Fig. 1d). Four of these patients (ACF, DA, GG, and RM) had a decreased MMN on the temporal generator of the lesioned hemisphere, and all of them had neocortical temporal lobe lesions not including mesial structures but extending outside the temporal lobe in one patient (RM, Fig. 2a). The other three patients (MG, MLS, and JP) had an increased MMN on the temporal generator of the lesioned hemisphere (Table 2), all having in common lesions centered on the mesial temporal lobe structures (Fig. 2b). This seems to indicate a higher sensitivity of the MMN as compared to the P100 to detect the functional changes associated with our group of symptomatic epilepsies.

No MMN hemispheric ratio abnormalities were apparent in patients with frontal lobe epilepsy (N = 3) or in those with occipital lobe epilepsy (N = 3) (Fig. 1d). 3.2. Anatomical data The lesion distribution was heavily biased towards the temporal lobes, with 11 patients having lesions mainly localized in these structures and extending to extratemporal areas in three patients. Three patients had the lesion in the frontal lobes, while the remaining three had lesions in the occipital lobes. In all cases, there was a

92

R. Lopes et al. / Epilepsy & Behavior 33 (2014) 87–93

good correlation between the localization of the brain lesion and the neurophysiological epileptiform abnormalities detected from the scalp (Table 1). All patients with abnormal MMN had a lesion in the temporal lobe, which, in two cases, extended to other brain structures (Figs. 2a and b). This provides a high specificity for this neurophysiological finding as an indicator of a temporal lobe lesion in symptomatic epilepsies. On the other hand, only 7 out of 11 patients with temporal lobe lesions had an abnormal MMN, resulting in a sensitivity of 64%. The patients with temporal lobe lesion and normal MMN (N = 4, Fig. 2c) had lesions sparing the Heschl's gyrus, but one of them (MB) had a hippocampal sclerosis confirmed on surgery. This latter patient suggests that a hippocampal lesion may not be as predictive of MMN abnormality as Heschl's gyrus. The VLSM analysis revealed a statistically significant (p = 0.05, corrected) lesion area on the superior temporal area (Fig. 2d) for the group with reduced MMN (N = 4). For the group with an increased MMN, the statistically significant (p = 0.05, corrected) lesion area involved the hippocampus but not the amygdala or the other mesial structures (Fig. 2e). 4. Discussion The main finding of our study is that the asymmetry of the MMN auditory potential has a high specificity for temporal lobe epilepsy when abnormal (around 100%) but a low sensitivity (around 64%). In our sample of patients, it provided more informative results than the P100 potential or the abnormal latency of the MMN. The analysis of lesions associated with the changes in MMN revealed that the decrease in amplitude is associated with lesions involving the middle Heschl's gyrus and increased amplitude for the lesions involving the hippocampus. Our overall results concerning the latencies of the P100/MMN potentials do not support the conclusions from other authors [18,19,27] who have demonstrated latency delays for patients with epilepsy based on group averages. The average latency of our patients was within the control range both for the P100 and MMN. Nevertheless, 6 out of 17 patients (35%) showed abnormally delayed MMN latencies, supporting heterogeneity in auditory stimuli processing speed in symptomatic epilepsy, which, in our sample, is not specific for lesion location. Mismatch negativity normalized amplitude on the lesion hemisphere was abnormal in 7 out of 11 (64%) patients with temporal lobe epilepsy and within the control range in all extratemporal patients. The high specificity of this finding suggests a potential clinical role of MMN as a marker of temporal lobe dysfunction in temporal lobe epilepsies. The fact that the mean age of the patients with increased MMN (12.3 years) was similar to that of the patients with decreased MMN (11.3 years) does not support an age-related effect on the results. In contrast, the P100 potential was abnormal in a minority of patients (3/17) with no specific anatomical lesion correlation, demonstrating a low sensitivity for the pathology. The use of middle latency auditory evoked potentials, including the MMN potential, as an indicator of temporal lobe dysfunction has been described [28–30] using data from patients with vascular and tumoral pathologies. Patients with superior temporal gyrus lesions showed significant average reduction of the ERP amplitude when compared to healthy controls or patients with inferior parietal lobe lesions. Our study, based on individual analysis of a sample of patients with temporal lobe epilepsies (N = 11) with structural lesions involving more diverse areas within the temporal lobe, including mesial structures, allowed us to define sensitivity and specificity measures for the MMN abnormalities in temporal lobe epilepsy as well as to more precisely analyze, using VLSM, the effect of lesions within different temporal lobe compartments (Fig. 2). We found an effect of superior temporal

lobe lesions, producing reduction in MMN amplitude, opposite to that of mesial lesions, leading to pathological increase of MMN amplitude. The late finding agrees well with the data from Miyajima et al. [19] who described an increase of MMN amplitudes in a group with temporal lobe epilepsy (TLE) (n = 20) but is in contrast with the results of Lin et al. [27], who found no amplitude differences between the lesion and healthy hemispheres in 12 cases of mesial temporal lobe sclerosis, using MEG and a duration deviant. The latter author analyzed both patients with left and patients with right temporal lobe epilepsy together, which might have precluded the detection of lateralization effects on the MMN. Two of our three patients with frontal lobe epilepsy had a pathological increase in MMN latency, and all had borderline amplitudes but within the normal range (Table 2). This suggests a frontal lobe contribution to the MMN, which was also found in other studies [19,27] and also in patients with frontal lobe lesions [30]. The fact that our methods emphasize the superior temporal lobe component of the MMN may be the explanation for the normal values of the MMN amplitudes for these patients. The high specificity of MMN amplitude abnormalities for temporal lobe epilepsy in our study may be due to the use of the main temporal lobe generator of the MMN for the interhemispheric comparison and agrees well with the results of Hara et al. [31] which found changes in the temporal component of the MMN in patients with TLE, but not in the frontal component. Because these temporal lobe generators are largely dominant over the extratemporal ones [19,27], the abnormalities produced by extratemporal lobe epilepsies did not reach pathological values. The low sensitivity of the MMN amplitude abnormalities for temporal lobe epilepsy in our study (64%) may be due to the fact that only a subgroup of our patients had structural lesions involving the superior temporal planum or the mesial structures. Both these areas were demonstrated to have a critical role in producing the MMN abnormalities in the VLSM analysis we performed (Fig. 2). Most studies relating MMN abnormalities to temporal lobe lesions were done at the group level [19,27,30], and only one distinguished between the hippocampal and neocortical temporal lobe cortex. The latter study described a reduction in MMN amplitude for neocortical patients with superior temporal lobe lesions, in contrast with no changes for patients with mesial lesions. The results of the present study were obtained from a pediatric population, which could induce a bias in comparison with similar studies in the literature which have uniformly been based on adult populations. Studies comparing the MMN characteristics using frequency deviants in different age groups found either no significant differences between school-age children and adults [14] or a small latency decrease from children to adults [32]. None of these studies reports significant interhemispheric asymmetries in the temporal lobe MMN components. In this context, the asymmetries reported in our study, which constitute its main results, are unlikely to be due to maturational effects on the MMN. An important limitation of our study is that because of the small numbers in each group of patients, no robust statistical measures could be used, making the results exploratory and in need of confirmation by larger samples of patients. The reliability of the obtained potentials could also benefit from a more cooperative group of patients, and in this respect, adult patients with temporal lobe epilepsy may offer an advantage. Overall, our results demonstrate that amplitude abnormalities of the MMN are highly suggestive of temporal lobe epilepsy, with decreased amplitude in lesions affecting the superior temporal lobe planum and increased amplitudes in hippocampal lesions. Our focus in the main temporal lobe generator proved effective in reducing the confusional effect of extratemporal generators, therefore allowing the use of the MMN as a useful clinical tool for the functional mapping of temporal lobe epilepsies.

R. Lopes et al. / Epilepsy & Behavior 33 (2014) 87–93

Acknowledgments Ricardo Lopes has been supported by the Grant SFRH/BD/65617/ 2009 from the Portuguese Foundation for Science and Technology (FCT). Conflict of interest None of the authors has any conflict of interest to disclose. References [1] Baxendale S, Thompson P. Beyond localization: the role of traditional neuropsychological tests in an age of imaging. Epilepsia 2010;51(11):2225–30. [2] Chugani H, Müller R, Chugani D. Functional brain reorganization in children. Brain Dev 1996;18:347–56. [3] Vingerhoets G, Deblaere K, Backes W, Achten E, Boon P, Boon PA, et al. Lessons for neuropsychology from functional MRI in patients with epilepsy. Epilepsy Behav 2004;5:81–9. [4] Lassonde M, Sauerwein H, Gallagher A, Thériault M, Lepore F. Neuropsychology: traditional and new methods of investigation. Epilepsia 2006;47(2):9–13. [5] Jones-Gotman M, Smith M, Risse G, Westerveld M, Swanson S, Giovagnoli A, et al. The contribution of neuropsychology to diagnostic assessment in epilepsy. Epilepsy Behav 2010;18:3–12. [6] Lopes R, Cabral P, Canas N, Breia P, Foreid J, Calado E, et al. N170 asymmetry as an index of inferior occipital dysfunction in patients with symptomatic occipital lobe epilepsy. Clin Neurophysiol 2011;122:9–15. [7] Näätänen R, Gaillard A, Mantysalo S. Early selective-attention effect on evoked potential reinterpreted. Acta Psychol 1978;42:313–29. [8] Sherg M, Vajsar J, Picton T. A source analysis of the late human auditory evoked potentials. J Cogn Neurosci 1989;1:336–55. [9] Giard M, Perrin F, Pernier J, Bouchet P. Brain generators implicated in processing of auditory stimulus deviance: a topographic event-related potential study. Psychophysiology 1990;27:627–40. [10] Näätänen R, Kujala T, Escera C, Baldeweg T, Kreegipuu K, Carlson S, et al. The mismatch negativity (MMN) — a unique window to disturbed central auditory processing in ageing and different clinical conditions. Clin Neurophysiol 2012;123:424–58. [11] Cheour M, Shestakova A, Alku P, Ceponiene R, Näätänen R. Mismatch negativity (MMN) shows that 3–6-year-old children can learn to discriminate non-native speech sounds within two months. Neurosci Lett 2002;325:187–90. [12] Salisbury D, Shenton M, Griggs C, Bonner-Jackson A, McCarley R. Mismatch negativity in chronic schizophrenia and first-episode schizophrenia. Arch Gen Psychiatry 2002;59:686–94. [13] Brønnick K, Nordby H, Larsen J, Aarsland D. Disturbance of automatic auditory change detection in dementia associated with Parkinson's disease: a mismatch negativity study. Neurobiol Aging 2010;31:104–13.

93

[14] Cheour M, Leppänen P, Kraus N. Mismatch negativity (MMN) as a tool for investigating auditory discrimination and sensory memory in infants and children. Clin Neurophysiol 2000;111:4–16. [15] Boatman D, Trescher W, Smith C, Ewen J, Los J, Wied H. Cortical auditory dysfunction in benign rolandic epilepsy. Epilepsia 2008;49:1018–26. [16] Liasis A, Bamiou D, Boyd S, Towell A. Evidence for a neurophysiologic auditory deficit in children with benign epilepsy with centro-temporal spikes. J Neural Transm 2006;113:939–49. [17] Honbolygo F, Csepe V, Fekeshazy A, Emri M, Marian T, Sarkozy G. Converging evidences on language impairment in Landau–Kleffner syndrome revealed by behavioral and brain activity measures: a case study. Clin Neurophysiol 2006;117:295–305. [18] Gene-Cos N, Pottinger R, Barret G, Trimble M, Ring H. A comparative study of mismatch negativity (MMN) in epilepsy and non-epileptic seizures. Epileptic Disord 2005;7(4):363–72. [19] Miyajima M, Ohta K, Hara K, Lino H, Maehara T, Hara M. Abnormal mismatch negativity for pure tone sounds in temporal lobe epilepsy. Epilepsy Res 2011;94:149–57. [20] Näätänen R, Pakarinen S, Rinne T, Takegata R. The mismatch negativity (MMN): towards the optimal paradigm. Clin Neurophysiol 2004;115:140–4. [21] Hirose Y, Hara K, Miyajima M, Matsuda A, Maehara T, Hara M, et al. Changes in the duration and frequency of deviant stimuli engender different mismatch negativity patterns in temporal lobe epilepsy. Epilepsy Behav 2014;31:136–42. [22] Knight R, Hillyard S, Woods D, Neville H. The effects of frontal and temporal–parietal lesions on the auditory evoked potentials in man. Electroencephalogr Clin Neurophysiol 1980;50:112–21. [23] Fuchs M, Kastner J, Wagner M, Ebersole J. A standardized boundary element method volume conductor volume. Clin Neurophysiol 2002;113(5):702–12. [24] Rorden C, Karnath H, Bonilha L. Improving lesion-symptom mapping. J Cogn Neurosci 2007;19(7):1081–8. [25] Seneta E, Phipps M. On the comparison of two observed frequencies. Biom J 2001;43(1):23–43. [26] Kimberg D, Coslett H, Schwartz M. Power in voxel-based lesion-symptom mapping. J Cogn Neurosci 2007;19(7):1067–80. [27] Lin Y, Hsiao F, Shih Y, Yiu C, Yen D, Kwan S. Plastic phase-locking and magnetic mismatch response to auditory deviants in temporal lobe epilepsy. Cereb Cortex 2007;17:2516–25. [28] Woods D, Knight R, Neville H. Bitemporal lesions dissociate auditory evoked potentials and perception. Electroencephalogr Clin Neurophysiol 1984;57:208–20. [29] Knight R, Scabini D, Woods D, Clayworth C. The effects of lesions of superior temporal gyrus and inferior parietal lobe on temporal and vertex components of the human AEP. Electroencephalogr Clin Neurophysiol 1988;70:499–509. [30] Alain C, Woods C, Knight R. A distributed cortical network for auditory sensory memory in humans. Brain Res 1998;812:23–37. [31] Hara K, Ohta K, Miyajima M, Hara M, Lino H, Matsuda A, et al. Mismatch negativity for speech sounds in temporal lobe epilepsy. Epilepsy Behav 2012;23:335–41. [32] Gomot M, Giard M, Roux S, Barthélémy C, Bruneau N. Maturation of frontal and temporal components of mismatch negativity (MMN) in children. Neuroreport 2000;14:3109–12.