Error-related negativity predicts failure in competitive dual-player video games

Abstract

Along with improvement in electroencephalogram (EEG)-measurement technology, limitations on the situations in which data can be recorded are gradually being overcome. EEG measurement in real environments has become increasingly important as a means to monitor brain activity in our daily lives, such as while playing consumer games in the living room. The present study measured brain EEG activity while two players engaged in a competitive consumer baseball game in conditions that closely resembled daily life. The recorded brain activity was thus likely related to natural mental reactions and cognitive function that occur in similar daily life activities. To measure the EEG from participants who freely moved while playing the game, we developed EEG devices that incorporated a wireless time synchronization system using Global Positioning Satellite (GPS) signals. These devices stamped the time obtained from the GPS signals onto each data sample, which was then used to synchronize the data that were recorded by different devices. When the batter in the game swung and missed, the error-related negativity component of the event-related EEG potential was strongly evoked in frontal electrodes of the participant controlling the batter. Furthermore, the error-related negativity was modulated according to who was winning and by how much. Thus, here we have demonstrated "real-world" brain activity using a competitive consumer game, which increases intrinsic participant motivation.

Fig 1. Wireless synchronization system among Electroencephalograms (EEGs), External devices (EXT), and the consumer game using a GPS signal.

This system can extract the onset time of game events from a composite video signal.

Fig 2. A schematic illustration of time adjustment.

All EEG and EXT signals are time-stamped with information obtained from the GPS signal. We adjusted the time information between EEG data and the recorded movies using the vertical synchronization signal (VSS) signal that was extracted from the composite signal.

Fig 3. A schematic illustration of the onset criteria for the three ball/strike events.

We visually identified the onset of game events from the recorded movies.

Fig 4. Grand-averaged ERPs for each event (ball, swinging strike, and called strike) for batters and pitchers at the five electrodes.

The yellow rectangles indicate the time window for statistical analysis. (A) Grand-averaged ERPs for batters. ERN can be seen before the onset of swinging strikes. P300 can be seen after the onset of all events. (B) Grand-averaged ERPs for pitchers. P300 can also be seen after the onset of all events.

Fig 5. Statistical results for ERN and P300 analysis.

(A) Statistical results for ERN. Error bars indicate the standard error. We observed significant differences in ERN amplitude of batters between each event type. (B) Statistical results for P300. Error bars indicate the standard error. We observed significant differences in the P300 amplitudes of batters and pitchers between swinging strikes and balls.

Fig 6. Correlation of ERN amplitude before swinging strikes with scores from the self-evaluation questionnaires (BIS11, BIS/BAS, and GEQ).

We found a trend for ERN amplitude to be smaller for higher values of ‘positive affect’.

Fig 7. Correlation between ERN amplitude and the average score difference for the three games.

Data in the yellow rectangle were used for fitting. ERN might be linearly modulated depending on who is winning the game and the relative score difference.