Stochastic Behavior of Phase Synchronization Index and Cross-Frequency Couplings in Epileptogenic Zones during Interictal Periods Measured with Scalp dEEG

Abstract

The stochastic behavior of the phase synchronization index (SI) and cross-frequency couplings on different days during a hospital stay of three epileptic patients was studied for non-invasive localization of the epileptogenic areas from high density, 256-channel, scalp EEG (dEEG) recordings. The study was performed with short-duration (0–180 s), seizure-free, epileptiform-free, and spike-free interictal dEEG data on different days of three subjects. The seizure areas were localized with subdural recordings with an 8 × 8 macro-electrode grid array and strip electrodes. The study was performed in theta (3–7 Hz), alpha (7–12 Hz), beta (12–30 Hz), and low gamma (30–50 Hz) bands. A detrended fluctuation analysis was used to find the long range temporal correlations in the SI that reveals the stochastic behavior of the SI in a given time period. The phase synchronization was computed after taking Hilbert transform of the EEG data. Contour plots were constructed with 20 s time-frames using a montage of the layout of 256 electrode positions. It was found that the stochastic behavior of the SI was higher in epileptogenic areas and in nearby areas on different days for each subject. The low gamma band was found to be the best to localize the epileptic sites. Also, a stable higher pattern of SI emerged after 60–120 s in the epileptogenic areas. The cross-frequency couplings of SI in theta–gamma, beta–gamma, and alpha–gamma bands were decreased and spatial patterns were fragmented in epileptogenic areas. Combinations of an increase in the stochastic behavior of the SI and decrease in cross-frequency couplings are potential markers to assist in localizing epileptogenic areas. These findings suggest that it is possible to localize the epileptogenic areas non-invasively from a short-duration (∼180 s), seizure-free and spike-free interictal scalp dEEG recordings.

Spatiotemporal phase clusters and phase synchronization patterns derived from high density EEG and ECoG recordings

“…The stochastic behavior of the SI is given in Figure 4. It was computed from 256-channel EEG data of an epilepsy patient [27 ]. The seizure area is marked with a rectangle that was determined with invasive subdural grid and strip electrode recordings.…”

“…Spatiotemporal plots of analytic amplitude and phase carry a wealth of information in all EEG bands. Generally, EEG bands are classified as: theta (3-7 Hz), alpha (7-12 Hz), beta (12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30), low gamma and high gamma (50-80 Hz). Two nearby channels slip in and out of phase in the temporal direction and form spatial structures.…”

What graph theory actually tells us about resting state interictal MEG epileptic activity

“…Besides, van Diessen et al (2013b) have for the first time developed a multivariable diagnostic prediction model based on functional network characteristics. However, despite extensive research on the characterization of epileptic functional brain networks from M/EEG ( Otte et al, 2012 ; van Dellen et al, 2012 ; Bartolomei et al, 2013 ; Quraan et al, 2013 ; Ramon and Holmes, 2013 ; van Diessen et al, 2013a ; Burns et al, 2014 ; Chiang and Haneef, 2014 ) or functional magnetic resonance data (e.g., ( Zhang et al, 2011 ; Centeno and Carmichael, 2014 )), there is no consensus on which measure (or set of them) is most suitable for the purpose, with different studies using difference measures. Besides, there is theoretic and empirical evidence ( Lynall et al, 2010 ; Li et al, 2011 ; Hahn et al, 2015 ) that many of them are strongly correlated with each other.…”

Cambios en la variabilidad temporal de la conectividad funcional electroencefalográfica durante el envejecimiento

“…En sujetos sanos y jóvenes, esta hipótesis de estacionariedad podría ser cierta bajo la hipótesis de que, sólo factores aleatorios afectan a las medidas EEG y, los factores cognitivos / psicológicos y neurofisiológicos permanecen poco alterados durante el período de registro. Sin embargo, incluso en estos casos, la evidencia reciente (Botcharova, Farmer, y Berthouze, 2014;Kitzbichler, Smith, Christensen, y Bullmore, 2009) sugiere claramente que la sincronización cerebral, según la evaluación de las señales neurofisiológicas, no es constante, sino que presenta una variabilidad temporal significativa o no trivial, que puede estar alterada en ciertas condiciones neuropatológicos tales como la epilepsia (Ramón y Holmes, 2013). Por tanto, es razonable pensar que tal variabilidad no trivial puede ser capturada por las medidas de CF de EEG, y que esta variabilidad también puede cambiar durante el estado de reposo de los sujetos mayores.…”