Comparison of brain effective connectivity in different states of attention and consciousness based on EEG signals

Abstract

Evidence from recent studies on attention and consciousness has raised many questions about their inherent natures and the existence of a possible relationship between them. The main goal of this study is to investigate the significant difference among different states of attention and consciousness, through the connectivity method. These two cognitive phenomena have been studied in recent years with different approaches, but none of them has employed connectivity methods to study the attention and consciousness relationship. Therefore, the connectivity approach is a novel effort in this field. The 8-channel Electroencephalogram signal (EEGs) was used from 45 participants performing a psychophysical visual task. To track the variation of connectivity during time, the EEG recorded data were organized in a particular way, providing 9 1 s signals, including four stimuli classes: attention-consciousness class, attention-unconsciousness class, inattention-consciousness class, and inattention-unconsciousness class. In order to reduce the volume conduction effect, source signals were extracted with the ICA method. The effective connectivity network was computed using the Granger causality method. For the statistical analysis, connectivity indices were presented. This approach suggests that attention-consciousness and attention-unconsciousness classes, and attention-unconsciousness and inattention-unconsciousness classes are the most and least distinctive combinations, respectively. In a more general view provided by averaging the connectivity matrices, all the presented connectivity indices showed significant differences with p-value<0.05 for two combinations: 1-attention-consciousness and attention-unconsciousness classes, and 2- inattention-consciousness and inattention-unconsciousness classes.