Frequency and spatial characteristics of high‐frequency neuromagnetic signals in childhood epilepsy

Abstract

Purpose. Invasive intracranial recordings have suggested that high-frequency oscillation is involved in epileptogenesis and is highly localized to epileptogenic zones. The aim of the present study is to characterize the frequency and spatial patterns of high-frequency brain signals in childhood epilepsy using a non-invasive technology. Methods. Thirty children with clinically diagnosed epilepsy were studied using a whole head magnetoencephalography (MEG) system. MEG data were digitized at 4 000 Hz. The frequency and spatial characteristics of high-frequency neuromagnetic signals were analyzed using continuous wavelet transform and beamformer. Threedimensional magnetic resonance imaging (MRI) was obtained for each patient to localize magnetic sources. Results. Twenty-six patients showed highfrequency (100-1 000 Hz) components (26/30, 86%). Nineteen patients showed more than one high-frequency component (19/30, 63%). The frequency range of high-frequency components varied across patients. The highest frequency band was identified around 910 Hz. The loci of high-frequency epileptic activities were concordant with the lesions identified by magnetic resonance imaging for 21 patients (21/30, 70%). The MEG source localizations of high-frequency components were found to be concordant with intracranial recordings for nine of the eleven patients who underwent epilepsy surgery (9/11, 82%).Conclusion. The results have demonstrated that childhood epilepsy was associated with high-frequency epileptic activity in a wide frequency range. The concordance of MEG source localization, MRI and intracranial recordings suggests that measurement of high-frequency neuromagnetic signals might provide a novel approach for clinical management of childhood epilepsy.

Comparison of Brain Networks During Interictal Oscillations and Spikes on Magnetoencephalography and Intracerebral EEG

“…Thus, recent studies suggest that epileptic oscillations can be detected in the beta/gamma band (Guggisberg et al 2008;Rampp et al 2010) as in our study, and even in the fast ripple band (Xiang et al 2009). As high frequency activity (HFOs [ 80 Hz) has been shown to be a good indicator of the seizure onset zone on intracerebral EEG (Worrell et al 2008;Bagshaw et al 2009;Zijlmans et al 2011), a promising venue would be to combine network localization on beta/low gamma oscillations with localization of HFOs (Van Klink et al 2014).…”

Volumetric imaging of brain activity with spatial-frequency decoding of neuromagnetic signals

“…Building on previous reports (Dalal et al, 2008;Popescu et al, 2008;Xiang et al, 2009a), the present study aimed to utilize the frequency signature of neuromagnetic signals for localizing brain activity. To that end, we optimized wavelet algorithms by providing a dynamic sigma value for better detection of high-frequency brain signals.…”

“…The parameters used for establishing the threshold are: (a) sample size (number of LOB sensors); (b) alpha value (0.01); (c) Two-tailed T test; and (d) type of correlation coefficient: Pearson's correlation. Building on our previous study (Xiang et al, 2009a), we determined the threshold of alpha for estimating activity area by finding t values which reached the 99th percentile of confidence value (p < 0.01). Since our method supports "multiparameters for each location", each voxel in the volumetric source images reconstructed with our method includes corrected source strength (computed with Eq.…”

“…Previous studies have developed several algorithms for localizing brain activity with MEG signals in a low frequency range (Brookes et al, 2011;Cheyne et al, 2007;Huang et al, 2004;Mosher et al, 2009;Nagarajan et al, 2006;Robinson et al, 2012;Sekihara et al, 2004;Tadel et al, 2011;Vrba et al, 2010). Recent reports have demonstrated that high-frequency MEG signals can also be utilized to localize brain activity (Dalal et al, 2008;Xiang et al, 2009aXiang et al, , 2010.…”

“…Increasing evidence indicates that HFBS are not only new biomarkers for epilepsy but also provide a window for developing novel source localization methods (Gotman, 2010;Worrell et al, 2012;Xiang et al, 2009aXiang et al, , 2010. The combination of spatial filter (beamformer) and time-frequency analysis has the potential to analyze both low and high-frequency signals at source levels (Krause et al, 2010;Laaksonen et al, 2008).…”