Study On Functional Connectivity Of Multi-channel EEGs In Healthy Subjects Based On Different Difficulty Mental Arithmetic

Objective Mental Arithmetic is a skill in daily life and it is closely related to cognition. Mental arithmetic tasks require the interaction of multiple brain regions. The correlations of key brain regions under mental arithmetic task are changeable when task difficulties are different. From the perspective of the frequency domain, this study designed three mental arithmetic tasks with different difficulties, and applied Granger causality analysis method to calculate EEGs causal connection. The information flow gain was calculated based on the connection matrix and the frequency band information flow gain between different brain regions under these three mental arithmetic tasks was compared. This may provide supports for further exploration of mental arithmetic information processing mechanisms.Methods 1. Experimental data: 19-channel EEGs were recorded from 10 healthy subjects under three types of mental arithmetic tasks. After applying the signal preprocessing of the baseline correction and 50 Hz reduction, 500 target stimuli and 500 non-target stimuli data segment were manually selected in each task. 2. Spatial distribution of power: Fourier transform was used to analyze 19 channels target and non-target stimulus EEGs. The topographies of power in the delta, theta, alpha, beta and gamma bands were calculated and compared. 3. DTFij calculation: Applied the Granger causality analysis to analyze the EEGs. For each subject, the directed transform function of the 5 band components under the three mental arithmetic tasks was calculated. 4. Information flow gain calculation: By using DTF matrix, the flow gain topographic maps were calculated. Then the information flow gain of frontal, central, temporal, and parietal region were calculated. 5. Connectivity strengths calculation: Based on the DTFij, the functional connectivity strengths of whole brain network were calculated within and among the brain regions(frontal, central, temporal and parietal regions). 6. Statistical analysis: One-way ANOVA analysis was used to evaluate the statistical differences among three mental arithmetic tasks(500 target segment from 10 subjects for each task). The group statistical differences were evaluated using the Student Newman Keuls(S-N-K) test.Results1. EEG spatial distribution of power in 5 frequency bands under the three mental arithmetic tasks The delta, theta and alpha power were mainly showed in the frontal region and central region; and the complexity of the task increased with increasing power. The beta and gamma power mainly showed in the frontal region; with the increase of task difficulty, the power in the left temporal region became concentrative and the power in the frontal region increased. 2. The topographic of information flow in 5 frequency bands under the three mental arithmetic tasks For the three mental arithmetic task, the DTF inflows information(DTFin) was distributed in the peripheral region; the outflow information(DTFout) was most active in the parietal region and the information flow gain was relatively active in the central-parietal region. 3. Information flow gain of the beta band in the four brain regions The comparison of the information flow gain among the frontal, central, temporal and parietal regions under the three mental arithmetic tasks showed that higher frequencies demonstrated higher statistical differences, especially the beta band. With the increase of task difficulty, the information flow gain in the beta band increased at the frontal and temporal regions(p<0.01), decreased at the parietal region(p<0.001), decreased first and then increased at the central region(p<0.05). Pair-wise comparison indicated that the flow gain of Chinese numerals mental arithmetic task was significantly higher than the counting mental task within the frontal and temporal region(p <0.01), and it was significantly lower than the other two tasks within the parietal region(p <0.001). 4. Functional connectivity strengths of EEGs under the three mental arithmetic tasks Functional connectivity strengths of the whole brain network for the three tasks in full band showed that the DTFglobal value of single-digit addition task was significantly lower than the other two tasks(p <0.001). Results of frequency division showed that the DTFglobal value of single-digit addition task was significantly LOWer than the other two tasks in the theta(p <0.01), alpha(p <0.001), and beta band(p <0.001)(S-N-K test, p <0.01). Functional connectivity strengths within and across brain regions showed that only the strengths from the frontal to the parietal regions(DTFPF) had significant differences among five bands(delta, theta, alpha: p <0.01; beta: p <0.001; gamma: p<0.05) and the DTF strengths of the count task were significantly smaller than the Chinese numerals mental arithmetic task. The others showed statistical differences in the higher frequency band.Conclusion 1. The results of information flow gain of the three mental arithmetic tasks indicated that the frontal and parietal regions were specific brain areas associated with mental arithmetic difficulty. In the higher frequency band of the given three tasks, especially the beta band, flow gain characteristics could be used to distinguish simple counting arithmetic task from Chinese numerals mental arithmetic task. Furthermore, in the beta band, the information flow activity within the right frontal region increased while within the parietal region decreased when the difficulty of mental arithmetic task augmented. 2. For the DTF within and across brain regions under the three mental arithmetic tasks, the functional connectivity showed a higher statistical difference in the beta band. With the increase of task difficulty, the frontal-parietal network connection strength enhanced.