Different Approaches to Analysis of Brain Activity Topography in Stop-Signal Paradigm

Abstract

Different approaches to analysis of EEG/MEG data were applied for exploration of topological distribution of human cortical activity in conditions of Stop-Signal Paradigm [Logan et al., 1984]. SSP is an experimental method for studying brain mechanisms of human behavioral control. This method could be applied as for clinical studies particularly for attention deficit and impulsiveness syndrome as for non-clinical one. In our study EEG/MEG correlates of trait anxiety [Spielberger, 1970] were analyzed. Event-related spectral perturbations [Delorme and Makeig, 2004] were analyzed as measure of functional changes in brain oscillations in the response to target stimulus onset, stop-signal onset and button pressing. Also, ERSP was used for comparison of brain oscillations in two anxiety groups. Independent Component Analysis (ICA) was applied as for artifact correction as for separation of signal into several sub-processes [Delorme and Makeig, 2004]. Electromagnetic spatial-temporal ICA (EMSICA) was used for localization of activity sources in the brain cortex [Tsay et al., 2006]. As a result of EEG-analysis, we found several components which differed during situations with and without the stop-signal. These components have different time-frequency parameters, but similar localizations on the cortex surface. The factor analysis was used to explore cortical distribution of event-related synchronization/desynchronization [Levin, 2008]. Five cortical areas were detected as statistically independent functional areas for seven frequency bands. Same cortical areas were localized for both conditions as relevant to the task. However, time-dynamic in each area was quite different. A functional clustering (FC) method [Chiou and Li, 2007], k-centres FC, for longitudinal data was used for clusterization of EEG/MEG electrodes. The k-centres FC approach accounts for both the means and the modes of variation differentials between clusters by predicting cluster membership with a reclassification step. The cluster membership predictions are based on a non-parametric random-effect model of the truncated Karhunen-Loeve expansion, coupled with a non-parametric iterative mean and covariance updating scheme. Analysis of ERP-plot was used for functional interpretation of culter's properties. Three-cluster model is relevant for analysis of EEG-data, whereas the two-cluster model is more appropriate for MEG-data. 　　　In summary, our data support hypothesis of Band et al [2003], that the processes of behavioral activation and inhibition have same localization in cortex areas, but different time-frequency properties. During SSP there is competition between activation and inhibition processes for control of motor response.