The Testing of EEG and HRV Parameters to Quantitatively Differentiate between the IGD and Healthy Group

Studies have been conducted to determine the status of the internet gaming disorder (IGD) using various bio-signals, including quantitative studies using electroencephalography (EEG) and heart rate variability (HRV) that have suggested discriminant models to identify IGD subjects. This study aimed to test the accuracy of the suggested models especially when the EEG and HRV parameters were used together to build up a discriminant equation. An experiment was designed based on previous studies. The subjects consisted of 25 college students with an average age of 22.7 (±2.5) and were classified into the IGD group (n = 13) and the healthy group (n = 12) by using Young's Internet Addiction Test (IAT) and Compulsive Internet Use Scale (CIUS). The subjects played the League of Legends game for 30–40 min and collected EEG (16 channel) and ECG data throughout the game. The 240 EEG parameters (16ch. * 15) and 14 HRV parameters were used to extract the most effective sets of parameters. The t-test was conducted to sort out the parameters differentiating between the IGD group and the healthy group. Factor analysis was used to select parameters with the eigen value greater than 0.8. To remove multicollinearity, Pearson correlation was employed in this analysis. Finally, six sets of parameters were selected for logistic regression to differentiate two groups. As a result, the highest accuracy of the model was found to be Model 4. The HRV parameter has been dropped during the process of parameter elimination. The observed accuracy ranged from 63.3 to 71.4% whereas the existing accuracy ranged from 63.5 to 73.1%. In this study no synergy was observed when using both EEG and HRV parameters. In future study, a refined model can be further investigated focusing on EEG signals. Otherwise, bio-signals at particular events during game play can be explored to find a statistical model with a high and robust accuracy value.