Betweenness centrality in multiplex brain network during mental task evaluation

In this paper we study the structural properties of a functional network of the human brain during the evaluation of mental tasks using the concept of betweenness centrality. We carry out the experiments involving the alternating trials of mental task evaluation with simultaneous registration of electroencephalographic (EEG) data. Using the wavelet phase coherence we reconstruct the functional multiplex network of the brain considering the different typical frequency bands of EEG activity as interconnected layers. We reveal that transition from a resting state to evaluation of a cognitive task leads to the strong outflow of shortest paths from low frequencies and strengthening of high-frequency connectivity in the brain. At the same time, we observe that mental activity shapes the shortest paths in a more uniform distribution across the brain, which implies the emergence of a more distributed functional network. Our results are in good agreement with recent studies of cognitive activity and can be implied in the design of brain-computer interfaces for the estimation of cognitive load or attention.

Figure 1

(a) The depiction of an example of the Schulte table. (b) Principal design of the experiment on Schulte table evaluation. (c)–(f) The procedure of the reconstruction of the functional multiplex network (see text for details).

Figure 2

The relative value of betweenness centrality, which share each frequency band in a multiplex network for passive (purple) and active (orange) phase of the experiment are shown via box plots.

Figure 3

The relative value of betweenness centrality for each channel in the (a) delta, (b) theta, (c) alpha, (d) beta-1, (e) beta-2, and (f) gamma range for the passive (purple) and active (orange) phase of the experiment are shown via box plots.

Figure 4

The box plot (a) and spatial distributions (b)–(g) of the $\varepsilon$ quantity (see text for definition) for the (b) delta, (c) theta, (d) alpha, (e) beta-1, (f) beta-2, and (g) gamma frequency ranges.

Figure 5

The relative value of betweenness centrality, which share each frequency band in a multiplex network during the motor task (blue), and the evaluation of the Schulte table (orange) are shown via box plots.

Figure 6

The spatial distributions of betweenness centrality in the delta and gamma range in transition from a resting state to task evaluation. Here FPN = frontoparietal network, OC = occipital cortex, PFC = prefrontal cortex.