Wavelet-Based Localization of Oscillatory Sources From Magnetoencephalography Data

Abstract

Transient brain oscillatory activities recorded with Eelectroencephalography (EEG) or magnetoencephalography (MEG) are characteristic features in physiological and pathological processes. This study is aimed at describing, evaluating, and illustrating with clinical data a new method for localizing the sources of oscillatory cortical activity recorded by MEG. The method combines time-frequency representation and an entropic regularization technique in a common framework, assuming that brain activity is sparse in time and space. Spatial sparsity relies on the assumption that brain activity is organized among cortical parcels. Sparsity in time is achieved by transposing the inverse problem in the wavelet representation, for both data and sources. We propose an estimator of the wavelet coefficients of the sources based on the maximum entropy on the mean (MEM) principle. The full dynamics of the sources is obtained from the inverse wavelet transform, and principal component analysis of the reconstructed time courses is applied to extract oscillatory components. This methodology is evaluated using realistic simulations of single-trial signals, combining fast and sudden discharges (spike) along with bursts of oscillating activity. The method is finally illustrated with a clinical application using MEG data acquired on a patient with a right orbitofrontal epilepsy.

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

“…The new method enables analysis of brain activity and activation at source space in addition to sensor space. Compared with reports with similar approaches (Liao et al, 2012;Lin et al, 2004;Lina et al, 2014), our method was optimized for detecting high-frequency MEG signals. Compared with outstanding MATLAB toolboxes (Dalal et al, 2004;Oostenveld et al, 2011), our implementaion of C/C++ seemed to be much faster in our tests for the same task (localizing auditory activation).…”

“…Since MEG signals are the spatiotemporal summation of synchronous activity from at least 10,000-50,000 neurons (Murakami and Okada, 2006), the frequency and spectral signatures of MEG data encode the spatiotemporal patterns of a group of neurons, which represent the functional organization of brain activity. The frequency and spectral signatures of MEG data in both low and high frequency ranges provide important information for computational reconstruction of functional brain activity (Lina et al, 2014).…”

Interictal High Frequency Oscillations Detected with Simultaneous Magnetoencephalography and Electroencephalography as Biomarker of Pediatric Epilepsy

“…NOTE: The MEM is an efficient technique that has been successfully used to determine the location and extent of sources of epileptic activity [62][63][64] . The wMEM is an extension of MEM that has been developed for localizing oscillatory activity as evaluated with realistic simulations 65 . It decomposes the signal in a discrete wavelet basis before performing MEM source localization on each time-frequency box.…”

Imaging brain source extent from EEG/MEG by means of an iteratively reweighted edge sparsity minimization (IRES) strategy

“…Another well-known Bayesian algorithm is the maximum entropy on the mean (MEM) approach (Grova et al, 2006) where the cortical surface is clustered in a data driven manner to obtain active and inactive regions on the cortex by regularizing the mean entropy. The idea of MEM has been further pursued by Chowdhury et al (Chowdhury et al, 2013) and Lina et al (Lina et al, 2014) in a parcelization framework, where the cortical surface is divided into segments and then it is determined if each parcel is within the active source patch or not (through statistical analysis). The proposed method in the present work is not defined within the Bayesian framework, but as a series of convex optimization problems, as will be discussed.…”