TitleGraph analysis of sensorimotor cortex functional networks - comparison of alpha vs beta rhythm in motor imagery and execution
Publication TypeConference Paper
Year of Publication2014
AuthorsAthanasiou, Alkinoos, Foroglou Nicolas, Polyzoidis Konstantinos, and Bamidis Panagiotis
Conference NameSociety of Applied Neuroscience (SAN) & National Initiative Brain & Cognition (NIHC) Meeting 2014
Conference LocationUtrecht, the Netherlands

Background: Alpha (8-12Hz) and beta (12-30Hz) rhythms constitute the main electroencephalographic (EEG) bandwidth associated with sensorimotor processes. While they feature distinct characteristics, it is suggested that they are phase-coupled.
Objective: By calculating Functional Connectivity (FC) of the sensorimotor cortex during motor tasks we studied the structural properties of the networks that each rhythm forms.
Methodology: Seven healthy, right-handed subjects performed four motor tasks (execution and imagery of hand and foot) under EEG recording. We defined a cortical source model with 16 regions of interest over the sensorimotor cortex and calculated FC by means of Directed Transfer Function for alpha and beta rhythm networks. We calculated four graph properties for each network: characteristic path length (L), clustering coefficient (C), density (D) and small-worldness (SW). Analysis of variance was used to determine statistical significance of observed differences.
Results: In beta rhythm networks, for all motor tasks, L was found 53.59% to 66.16% less (p<0.05), C was 28.17% to 39.76% less (p<0.05) and D was 36.14% to 40.65% greater (p<0.001) than in alpha rhythm networks. SW presented no statistically significant difference.
Conclusions: Alpha rhythm sensorimotor networks appear to form more segregated nodes, while beta rhythm sensorimotor networks form more effective connections despite being denser.
Discussion: We propose a model regarding the functional role of both rhythms, in which alpha carries elaborate information on the neurophysiological process, using greater wiring costs, while beta carries coordinative information among clusters of nodes. This model needs to be validated with appropriate experiments and different cortical source models.