Changes in background electroencephalographic activity in benign childhood epilepsy with centrotemporal spikes after oxcarbazepine treatment: a standardized low-resolution brain electromagnetic tomography (sLORETA) study

Benign childhood epilepsy with centrotemporal spikes (BCECTS), also known as rolandic epilepsy, is the most common form of benign childhood epilepsy. The seizures are characterized by hemifacial motor seizures, and frequently involve the hand or both the hand and leg on the side ipsilateral to the involved facial side [1]. By definition, centrotemporal spikes observed on electroencephalography (EEG) are the hallmark of BCECTS [2]. The prognosis is usually favorable [1]. Oxcarbazepine was established as an effective initial monotherapy for BCECTS by the Commission of Treatment Guidelines of the International League Against Epilepsy (ILAE) [3].

Several neuroimaging studies using EEG and functional magnetic resonance imaging (MRI) have reported neurophysiological alterations in BCECTS, but the reported outcomes have been inconsistent, and the progress of alterations in the brain remains unresolved. Some studies on dipole source localization have demonstrated that centrotemporal spikes can be reliably modelled by single tangential dipole sources oriented from the central to the frontal lobes and localized in the central regions, namely the rolandic area [4‐7]. However, there is growing evidence from neuroimaging studies that BCECTS may functionally and structurally affect a larger portion of the brain [8‐12]. Moreover, altered EEG characteristics in BCECTS are mostly studied with respect to qualitative aspects of epileptic activity, whereas background activity in the interictal period is often not quantified [13].

Distributed models for EEG source localization are increasingly sophisticated. Furthermore, a theoretical advantage held by distributed modelling is that its algorithms address the inverse problem with few lead-in assumptions [14]. Standardized low-resolution brain electromagnetic tomography (sLORETA) is one such distributed modelling method. sLORETA can yield 3-dimensional images of electrical neuronal activity, with maximum similarities of orientation and strength between neighbouring neuronal populations [15]. Therefore, it is an excellent tool for investigating the electrophysiological and neuroanatomical features of BCECTS [15‐17]. However, few studies have investigated the electrophysiological characteristics of BCECTS using sLORETA [12, 13].

Thus, to better understand the neurophysiological alterations and their development in BCECTS, the present study investigated changes in background EEG activity after six months of oxcarbazepine treatment, which is considered the drug of choice in BCECTS [3]. We performed a comparative analysis of background EEG activity after oxcarbazepine treatment using sLORETA.

Comparative analysis using SnPM of the sLORETA data revealed significant current density differences for the delta frequency band in the frontal lobe (middle frontal gyrus, inferior frontal gyrus, medial frontal gyrus, subcallosal gyrus, rectal gyrus, and orbital gyrus), the temporal lobe (superior temporal gyrus, middle temporal gyrus, inferior temporal gyrus, fusiform gyrus, and subgyral), and the limbic lobe (parahippocampal gyrus and anterior cingulate) of the left hemisphere after six months of oxcarbazepine treatment (threshold log-F-ratio = ±2.729, P < 0.01) (Figs. 1 and 2). Otherwise, there were no significant differences in the background EEG activity between the untreated and treated conditions.

Figures 1 and 2 show the statistical maps of the spatial extent of voxels within the areas of significant current density differences in the three-dimensional fiducial cortical surface and brain MRI template axial slices, respectively. The significant current density differences for the delta frequency band extend over the frontal, temporal, and limbic cortices of the left hemisphere. As a result, background delta frequency EEG activity after oxcarbazepine treatment was significantly decreased throughout these extended regions. The maximum current density decrease for the delta frequency band was found in the parahippocampal gyrus of the left limbic lobe (MNI coordinate [x, y, z = 25, − 20, − 10], Brodmann area 28) (log-F-ratio = 3.081, P < 0.01) (Fig. 3).

The essence of our results is that delta frequency activity was significantly reduced in the fronto-temporal and limbic cortices after oxcarbazepine treatment. Several studies have reported enhanced low-frequency band power in the background EEG activity of children with epilepsy compared to that of healthy controls [13, 27, 28]. These results were interpreted as reflecting dysfunction of the relevant cortical regions in children with epilepsy. In addition, later studies showed that antiepileptic drugs reduced background EEG activity in the low-frequency band [20, 29, 30]. These findings are consistent with previous studies. In addition, our results suggest that the fronto-temporal and limbic regions might be involved in BCECTS, and antiepileptic drugs could reduce the low-frequency band power of the background EEG activity in these regions.

The mechanisms underlying these results are not clear, but several possibilities exist. One explanation is that antiepileptic drugs decrease activity in the cortical areas of increased excitability or synchronization that are responsible for seizures [31, 32]. Recent studies have demonstrated that BCECTS was involved in widespread cortical regions over the Rolandic area, but no clearly defined cortical areas have been identified [8‐12]. Our previous study using a distributed source model showed that source localization of interictal epileptiform discharges in BCECTS included the central, frontal, and temporal regions surrounding the rolandic area [12]. As mentioned above, several studies have reported enhanced low-frequency activity in the background EEG activity of children with epilepsy [13, 27, 28]. Moreover, increased background EEG activity for the low-frequency band was also demonstrated in the temporal lobe and parietal lobe of BCECTS patients [13]. In an empirical study, oxcarbazepine treatment modified the extent of interictal spikes associated with seizure control [33]. We found the decreased EEG activity seen in those studies closely overlapped regions in our results in the frontal, temporal, parietal, and limbic lobes. This anatomical overlap between the specified cortical areas and the oxcarbazepine effect supports the possibility that oxcarbazepine decreases EEG activity in the cortical neural networks that are involved in BECTS [30].

However, this anatomical distribution cannot be the determinant of the antiepileptic drug effect. Seizure control and epileptogenic activity modification might be linked to decreased EEG activity in the low-frequency band. These observations call for caution in the neurophysiological interpretation of the sLORETA findings, but do not contradict the statement that antiepileptic drugs could decrease background EEG activity in the areas that are involved in BCECTS [30]. The three-dimensional sLORETA results encourage further detailed surveys of our findings. Importantly, further experimental studies may help when interpreting our sLORETA results. For example, a more extended study that includes participants with a range of disease courses undergoing various treatment protocols might answer this question.

It is widely accepted that BCECTS originates from the rolandic area [2, 34]. However, in recent years, widespread cortical involvement in children with BCECTS has been increasingly reported [35‐38]. These reports support the view that BCECTS, although initially described as a benign epilepsy syndrome, represents a deviation from normal development during a critical period of brain maturation [36]. BCECTS has been associated with various cognitive function impairments, as well as behavioural problems [1, 2, 13]. Children with BECTS may be more inclined to need antiepileptic drug treatment and special medical attention, instead of the often-applied “wait-and-see” strategy [36].

Our results can be discussed from the lens of epileptogenic zone, seizure-onset zone, symptomatogenic zone, and functional deficit zone. The seizure onset zone is the area of the cortex from which clinical seizures are actually generated, as opposed to the epileptogenic zone, which is the area of the cortex that is indispensable for generation of epileptic seizures. The symptomatogenic zone is the cortical area that is invaded by electrical spread from the seizure-onset zone, and is in the neighbourhood of or directly connected to the seizure-onset zone. The functional deficit zone is the cortical area that does not function normally in the interictal period. The area of functional cortical abnormality frequently extends outside the limits of the epileptogenic zone [39, 40]. However, this area can be defined by a number of tests, including neuropsychological examinations and functional imaging, such as EEG and functional MRI [40]. We did not observe inclusion of the fronto-temporal limbic cortex in the functional deficit zone in this study. Neuropsychological examinations to evaluate the comorbid impairments in children with BCECTS were not routinely included in this study. Further studies including such examinations may provide more data on the functional deficit zone in BCECTS.

An additional limitation of this study is the small number of patients. A larger sample size is desirable to improve the statistical power and generalizability of the results, and would allow the additional possibility of parametric testing [41]. However, the SnPM method adopted here is well suited to the limitations of our study design and sample size [15, 22, 42]. In fact, one of the successful examples of the use of the SnPM method offered by Nichols and Holmes is a functional MRI study between two conditions with 12 subjects, which directly reflects the constraints of the present study [22]. Therefore, we believe that there is a justification for reporting our results. Meanwhile, this study design was not able to make a comparison of the background EEG activity between the untreated patients and age/gender-matched healthy controls. Several previous publications have reported the comparative study of untreated patients with healthy controls for this purpose. Our results are consistent with the previous studies with regards to the affected frequency band, and contribute to further understanding of the progression of brain alterations in BCECTS during antiepileptic drug treatment. A more expanded study with age/gender-matched healthy controls is needed, and such a study might be able to confirm our results. Although sLORETA has no bias location, it tends to show diffused results. sLORETA is an excellent tool with a theoretical advantage of distributed modelling, but some other brain mapping algorithms, including adaptive or Bayesian-based beamformers, for example, Multiple Sparse Priors (MSP) and Champagne, could also be considered. Thus, further studies with larger sample sizes, age−/gender-matched healthy controls, neuropsychological examinations, and other analysis tools are needed to validate and expand on the results of our study.

In conclusion, this study confirms the involvement of the fronto-temporal and limbic regions in BCECTS, and demonstrates antiepileptic drugs could reduce the delta frequency activity of the background EEG in these regions. It has been suggested that widespread cortical regions over the rolandic area are involved in BCECTS. Our results contribute to further understanding of brain alterations and their progression in BCECTS. In the future, we suggest more expanded studies including a larger population, age−/gender-matched healthy controls, neuropsychological examinations, and other analysis tools to further extend our understanding.