Effect of Preadolescents’ Obesity on Inhibitory Control During Stop-Signal Task: A Functional EEG Network Study

Figures

• 

  Figure 1.. The workflow of data analysis. The procedure consists of four parts: participant enrollment, experiment design (stop-signal task: “Go”, and “No-go” conditions are represented in a timeline), EEG preprocessing (ICA: Independent component analysis. The ICA was used to remove residual artifacts by removing artifact components), and EEG analysis. Specifically, the EEG analysis including the network construction using the weighted phase lag index, calculation of network properties, BMI scores prediction based on network features using the stepwise linear regression). Herein, the “Non-obese” represents the healthy preadolescents and the “Obese” represents the preadolescents with obesity.

• 

  Figure 2.. Comparison of the network topology within the beta band between obese and non-obese groups, as well as between “Go” and “No-go” conditions. (a) Edges with significant main effect of the Group factor in network topology (p < 0.0001, false discovery rate (FDR) corrected). (b) Differences in network topologies between “Go” and “No-go” conditions for obese preadolescents (p < 0.0001, FDR corrected). (c) Differences in network topologies between “Go” and “No-go” conditions for non-obese preadolescents (p < 0.0001, FDR corrected). (d) Edges with significant main effect of the Condition factor in network topology (p < 0.01, FDR corrected). (e) Differences in network topologies between obese and non-obese groups under “Go” condition (p < 0.01, FDR corrected). (f) Differences in network topologies between obese and non-obese groups under “No-go” condition (p < 0.01, FDR corrected). In each subfigure, the blue lines denote the stronger edges of the obese group than that of the non-obese group, as well as denote the stronger edges of the “No-go” condition than that of the “Go” condition.

• 

  Figure 3.. Network properties of obese and non-obese groups under the “Go” and “No-go” conditions. (a) Comparison of the network properties between obese and non-obese groups under the “Go” and “No-go” conditions within the beta band. Values are the mean and standard deviation, * represents p < 0.05, and *** represents p < 0.001. (b) Differences of the network properties between “Go” and “No-go” conditions for obese and non-obese preadolescents within the beta band. Values are the mean and standard deviation of differences, *** represents p < 0.001.

• 

  Figure 4.. Relationships between body mass index (BMI) scores and network features. (a) Relationships between BMI scores and network connections under the “Go” condition. (b) Relationships between BMI scores and network properties under the “Go” condition, (c) relationships between BMI scores and network connections under the “No-go” condition. (d) Relationships between BMI scores and network properties under the “No-go” condition. In the network topologies, the blue lines indicate the significant positive correlations between the BMI scores and the network connections (p < 0.05, FDR corrected), and in each scatter correlogram, the black line is the fitted linear trend between two variables, the region of the black dashed line is the confidence interval, the circles are the included participants, r is the correlation coefficient, and p represents the statistical significance level.

• 

  Figure 5.. The correlation between predicted and actual BMI scores of preadolescents. The p represents statistical significance, r represents the correlation coefficient, RMSE is the root mean square error and NRMSE is the normalized root mean square error of predicting the model. The orange-filled circles are the enrolled participants.