Exploration of event-induced EEG phase synchronization patterns in cognitive tasks using a time–frequency-topography visualization system

EEG Correlations during WCST Performance in Female Adolescents with Sexual Abuse-Related Post-Traumatic Stress Disorder

“…CSA survivors can also manifest changes in brain functional synchronization or connectivity, which can be assessed by EEG coherence (coh) and correlation (r) analyses [9] [10]. Both analyses evaluate the degree of similarity between two EEG signals, though they use distinct mathematical calculations [11].…”

Classification and interactive segmentation of EEG synchrony patterns

“…This can be accomplished by running the EEG signals through a bank of bandpass quadrature filters. A common choice are the Gabor filters [26], which provide the best balance between time and frequency localization; however, because of their Gaussian frequency response, Gabor filters may have a significant response to negative frequencies when tuned at low frequencies, which may seriously distort the phase of the filter's output (for a detailed discussion, see [9]). This problem can be avoided if one forces the filters to have zero gain for negative frequencies.…”

“…Various measures have been proposed in the literature, including coherence [6], and measures based on certain statistics of the phase difference, such as the circular variance [7] [8], or the average magnitude [9]. Most of these measures yield values between 0 (no synchrony) and 1 (perfect synchrony); in practice, however, the differences between values at episodes of high synchrony and episodes of low synchrony are very subtle, thus a statistical analysis is usually required to determine the true significance of the observations.…”

“…This representation has been used extensively (see, for example, [5], [2], [10], [11]); however, it has two serious disadvantages: first, it may become too cluttered with complex synchronization patterns, and second, it does not show the dynamics of synchrony over time or frequency. The first disadvantage can be overcome with a multitoposcopic display (MD's) [9] (Figure 1b), in which one head diagram is plotted at each electrode position; each head diagram shows the spatial distribution of the synchronization between the corresponding electrode and every other site. Graph diagrams and MD's are both useful for 10 or 20 EEG channels, but they may become too cluttered with high-density EEG (64 or more electrodes), although it is possible to use MD's for high-density EEG by grouping the electrodes in regions and plotting one head diagram for each region [9].…”

“…The first disadvantage can be overcome with a multitoposcopic display (MD's) [9] (Figure 1b), in which one head diagram is plotted at each electrode position; each head diagram shows the spatial distribution of the synchronization between the corresponding electrode and every other site. Graph diagrams and MD's are both useful for 10 or 20 EEG channels, but they may become too cluttered with high-density EEG (64 or more electrodes), although it is possible to use MD's for high-density EEG by grouping the electrodes in regions and plotting one head diagram for each region [9]. A recent work [12] proposes a different solution, which consists in finding spatially connected maximal cliques (i.e., groups of mutually-synchronized electrodes); these cliques are referred to as functional units (FU's).…”