Abstract
Abstract
Objective:
In recent years, immersive videogame technologies such as virtual reality have been shown to affect psychological welfare in such way that they can be applied to clinical psychology treatments. However, the effects of videogaming with other immersive gaming apparatuses such as commercial electroencephalography (EEG)-based brain–computer interfaces (BCIs) on psychological welfare have not been extensively researched. Thus, we aimed at providing early insights into some of these effects by looking at how videogaming with a commercial EEG-based BCI would impact mood and physiological arousal.
Materials and Methods:
A total of 26 participants were sampled. Participants were randomly assigned to either a BCI condition or a traditional condition wherein they played an action videogame with a commercial EEG-based BCI or a standard keyboard and mouse interface for 20 minutes. In both conditions, participants filled out the profile of mood states to assess mood and the perceived stress scale to control for stress. We also measured heart rate, heart rate variability as measured by the root mean square of successive differences, and galvanic skin response (GSR) amplitude differences.
Results:
Participants in the BCI condition overall reported a significantly higher total mood disturbance (P < 0.05), tension (P < 0.05), confusion (P < 0.05), and significantly less vigor (P < 0.05). We also found that participants in the BCI condition had significantly lower GSR amplitude differences between gaming and baseline (P < 0.05).
Conclusion:
The results suggest that the use of commercial EEG-based BCIs for playing with videogames can induce greater frustration and negative moods than playing with a traditional keyboard and mouse interface, possibly limiting their use in clinical psychology settings.
Introduction
There are documented beneficial psychological effects of videogaming. These effects range from the improvement of reaction times in target acquisition tasks 1 to therapeutic effects through the alteration of psychological processing. 2 However, there are limitations associated with studying the effects of videogames without considering videogaming interfaces. Moreover, it is possible that novel gaming interfaces could modulate the aforementioned effects of videogaming. One example of this is virtual reality (VR), which has been shown to work well in patients who are resistant to traditional exposure therapies 3 and it is a promising avenue for the treatment of psychosis. 4
Brain–computer interfaces (BCIs) are another type of novel gaming interface. They are rarely studied in the context of videogaming, but there is an increasing interest in using them to play games since they could greatly enhance immersion. 5 In particular, self-paced BCIs, which allow users to send mental commands in real time, 6 along with the availability of commercial BCI headsets, make it possible to study how BCIs affect psychological welfare when used in real-life gaming sessions.
Thus, the aim of this study was to provide insights into how playing videogames with a commercial electroencephalography (EEG)-based BCI affected physiological arousal and mood compared with playing with a traditional keyboard and mouse interface. We hypothesized that playing with the BCI would induce higher total mood disturbance (TMD) and would induce more physiological arousal. Specifically, we expected to see higher galvanic skin response (GSR) amplitude differences between gaming and baseline, lower root mean square of successive differences (RMSSD), and higher average heart rate (HR).
Materials and Methods
Participants
Twenty-six participants were recruited from Nova Southeastern University (NSU) (mean age = 18.19, standard deviation = 0.63; nine males). The ethical aspects of the study were approved by NSU's Institutional Review Board (2017-161).
Materials
For this study, we used the five-channel EMOTIV Insight BCI (San Francisco, CA). We also used the EMOTIV Control Panel and Emokey software to handle mental commands. Participants played the videogame Call of Duty: Ghosts (Activision, Supplementary Table S1) on an MSI GS70 laptop. We also administered the profile of mood states (POMS) to measure mood given its traditional usage in mood reserach,7,8 and the perceived stress scale (PSS) as a self-reported measure of stress. 9 Lastly, we used the Empatica E4 (Boston, MA) wristband to measure GSR, average HR, and heart rate variability (HRV).
Procedure
After obtaining written informed consent, participants were randomly assigned to either the traditional condition where they played with the keyboard and mouse interface or with the BCI condition. In the BCI condition, participants were trained to operate the BCI using EMOTIV Emokey and Control Panel. First, we trained a neutral state in which the participant was idle. Then, we trained the following actions until participants acquired expertise, and assigned them to the following keys and videogame actions:
“Surprise” Facial Expression—Left Mouse Click (shoot) “Frown” Facial Expression—Right Mouse Click (aim) “Push” Mental Command—W Keyboard Key (move forward).
To allow the participants to look around in the game, the BCI's gyroscope was used. In the traditional condition, participants were told to use the W key as well as the left and right mouse clicks only. They were also instructed to move the mouse sideways to look around in the game.
All participants completed the PSS to control for prior stress before playing the videogame and the POMS after playing the game to assess their differences in mood. In both conditions, participants had to wear the Empatica E4. GSR was recorded in two epochs: the 5-minute baseline epoch before videogaming and the 20-minute gaming epoch wherein participants played Call of Duty: Ghosts at the “regular” difficulty with the assigned interface. HR and HRV were recorded continuously across the gameplay epoch.
GSR data were smoothed using a simple moving average algorithm. We calculated the amplitude by subtracting the peak conductance during gameplay from the average conductance during baseline to account for individual variability. RMSSD was calculated using the Kubios HRV standard edition software.
Results
To analyze our data, we conducted independent groups t-tests. We found that participants in the BCI condition reported significantly higher levels of TMD (t(24) = 2.24, P < 0.05), tension (t(24) = 1.79, P < 0.05), and confusion (t(24) = 2.04, P < 0.05). They also reported significantly lower levels of vigor (t(24) = 1.84, P < 0.05). See Table 1 and Figure 1 for more details. Participants in the BCI condition also showed significantly less GSR amplitudes (t(24) = 2.45, P < 0.05; Fig. 2). We did not observe differences in PSS scores, average HR, or RMSSD (P > 0.05).
Changes in mood categories by experimental condition. Error bars are ±SEM. BCI, brain–computer interface; POMS, profile of mood states. *Statistically significant difference at P < 0.05.
Average GSR amplitude difference in micro Siemens by experimental condition. Error bars are ±SEM. GSR, galvanic skin response. *Statistically significant difference at P < 0.05.
Descriptive Statistics of All Measures
P < 0.05.
BCI, brain–computer interface; GSR, galvanic skin response; HR, heart rate; POMS, profile of mood states; PSS, perceived stress scale; RMSSD, root mean square of successive differences; SEM, standard error of the mean; TMD, total mood disturbance.
Discussion
The goal of this study was to provide early insights into how playing a videogame with a commercial EEG-based BCI would impact a participant's mood and physiological arousal compared with a traditional videogaming setup. We hypothesized that playing with a BCI would lead to more disturbed moods and more physiological arousal. In terms of mood, our results aligned with our hypothesis since participants on the BCI condition had significantly higher TMD after gaming. However, we also found that the participants in the BCI condition had significantly lower GSR amplitudes. This may seem counterintuitive, but there is evidence that GSR may be negatively correlated with frustration and positively correlated with challenge. 10 Therefore, it is possible that participants felt frustrated due to difficulties operating the BCI.
Other studies have shown that playing action videogames with a BCI could be an unpleasant experience based on self-reported measures. Complex videogames can take long to understand for short exposures, which can lead to users not feeling in control. 11 Unwanted command detections have also been found to be more frustrating than failures to detect a command by the BCI. 11 We also encountered this issue; however, it is an inherent limitation of EEG-based BCIs given the high sensitivity to noise of EEG. 12
Future research could involve a longitudinal study looking at the same variables, but with more structured training to allow participants to gain expertise operating the game and the BCI. In addition, a comparison with VR setups would provide a holistic view of how immersive interfaces affect mood and physiological arousal.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
