Brain Connectivity Variation Topography Associated with Working Memory.

Abstract

Brain connectivity analysis plays an essential role in the research of working memory that involves complex coordination of various brain regions. In this research, we present a comprehensive view of trans-states brain connectivity variation based on continuous scalp EEG, extending beyond traditional stimuli-lock averaging or restriction to short time scales of hundreds of milliseconds after stimulus onset. The scalp EEG was collected under three conditions: quiet, memory, and control. The only difference between the memory and control conditions was that in the memory condition, subjects made an effort to retain information. We started our investigation with calibrations of Pearson correlation in EEG analysis and then derived two indices, link strength and node connectivity, to make comparisons between different states. Finally, we constructed and studied trans-state brain connectivity variation topography. Comparing memory and control states with quiet state, we found that the beta topography highlights links between T5/T6 and O1/O2, which represents the visual ventral stream, and the gamma topography conveys strengthening of inter-hemisphere links and weakening of intra-hemisphere frontal-posterior links, implying parallel inter-hemisphere coordination combined with sequential intra-hemisphere coordination when subjects are confronted with visual stimuli and a motor task. For comparison between memory and control states, we also found that the node connectivity of T6 stands out in gamma topography, which provides strong proof from scalp EEG for the information binding or relational processing function of the temporal lobe in memory formation. To our knowledge, this is the first time for any method to effectively capture brain connectivity variation associated with working memory from a relatively large scale both in time (from a second to a minute) and in space (from the scalp). The method can track brain activity continuously with minimal manual interruptions; therefore, it has promising potential in applications such as brain computer interfaces and brain training.