Topological and dynamical features of source localized EEG networks in presaccadic visual processing

Abstract

The topology and dynamics of cortical networks involved in the generation of presaccadic neural signals remain poorly understood. in particular, how they interact with simultaneously presented visual stimuli. Here, we used different approaches in graph theory analysis (GTA) and electroencephalography (EEG) to evaluate topology and dynamics of functional brain networks in the perisaccadic interval. EEG was recorded via 64 channels in two behavioral conditions (fixation or saccade). Participants (N=21) were pre-cued with a series of 1-3 grids (three horizontal lines, 10° by 10°) located 5° below the central fixation-point. 100ms later, a stimulus (three vertical lines; same size/location) was briefly presented (for 70ms). In the saccade condition, a left/right shift of the fixation-point during the interstimulus interval triggered a saccade after the second stimulus. Source localization (SL) was performed on the 200ms period following the saccade cue, or the equivalent time during fixation trials. Lagged coherences were calculated between all pairs of 84 Brodmann areas. SL/GTA both identified major network hubs near the frontal and parietal eye fields, with widespread cortical connectivity. Other GTA measures (clustering coefficient, global efficiency, energy, entropy) showed that network segregation, integration, synchronizability, and complexity were enhanced during the perisaccadic interval. Further, these network properties significantly interacted with stimulus repetition, altering both hubs and network topography. These data suggest a network mechanism for enhanced visual information processing and propagation in the presaccadic interval.