Acquiring fear intensity from EEG

My bachelor's thesis project is about the measurement of fear using EEG signals, similarly to my previous lab work. In this case though I made the participants to play a horror game instead just inducing fear by the means of audio stimuli. Also, I attempted to obtain the intensity of the emotion not just whether it is present or not - in addition to EEG signals, self-assessment, heart rate (HR) and skin conductance (GSR) data were gathered to verify the intensity. If you are interested in the whole thing in detail you can read it here (or at the end of this post). Team of the Synetiq neuromarketing research company supervised my thesis. They helped me a lot in putting together the measurement layout, the preprocessing of HR, EEG, and GSR signals and in the analysis of results.

Course of measurement

The experiment consisted of six main parts: 1) first, baseline self-assessment, 2) open-, and closed-eye baseline measurements, 3) baseline gameplay, 4) second self-assessment, 5) fear inducing gameplay, and 6) the last self-assessment. In total, the experiment took about 30 to 40 minutes with the time spent on the placement of the measurement tools excluded.

Parts of the 30-minute measurement as of time – width of the boxes are proportional to the length of the corresponding events.

Open-, and closed-eye measurements were used in the EEG preprocessing phase, baseline gameplay EEG, HR, and GSR values were compared with ones during the fear inducing gameplay. Self-assessment results were compared across situations to check the changes in the intensity of discrete emotions.

Mean scoring of the self-report measurements for every emotion questioned. Each line represents a questionnaire filled in different times during measurement. Significant changes can be seen in calmness, happiness, and fear, which demonstrates the effective fear eliciting property of the game. After death scores are irrelevant as just a small number of participants died.

The chosen game to induce fear is called “Slender, The eight pages” by Parsec Productions (http://slendergame.com/). Slender is a first person game taking place in a forest at night in pitch dark. Players are provided with a weak and finite flash light as their only tool. The goal of the player is to find 8 manuscripts scattered in a large territory, while Slender Man - a tall faceless figure, mythically known for kidnapping children keeps following the player preventing them to accomplish their objectives. The choice fell on Slender for a couple of reasons: 1) it has a simple story and concept, 2) it is not language heavy, 3) it can be controlled in a conventional and easily learnable way, 4) it is a robust elicitor of fear and anxiety emotional experiences, and 5) it has a so-called "day mode", which was suitable for the baseline measurements - it did not occur to the participants that they played a horror game until the fear inducing gameplay started.

Picture of a participant playing the fearful ("normal mode") part of the measurement.

Frontal Alpha Asymmetry

The EEG asymmetry feature was obtained as

where the power spectral density (PSD) via Welch periodogram was used to determine the alpha band powers for the EEG signals. The alpha waves activity was taken between 8 and 13 Hz. Logarithm was calculated for both left and right hemisphere alpha activity before subtracting the right from the left.

Channel locations of the Emotiv EPOC EEG device used in measurements. The electrodes, from which the asymmetry features were obtained are indicated in red.

Results

HR and GSR values werecompared across four research conditions (open-eye baseline, closed-eye baseline, day game, night game) using full within-subject ANOVA. The within-subject variables were time (in seconds) and the name of the condition. HR comparisons yielded significant, while GSR comparisons yielded marginally significant results as for the expected increase of both during the fear inducing gameplay. Frontal Alpha Asymmetry within-subject ANOVA yielded insignificant results, probably due to the low sample size and the fast changing nature of the asymmetry feature. What is quite sure and can be seen on the plots that the EEG asymmetry drops when skin conductance (precisely SCR, phasic part of GSR) jumps during the fear inducing gamplay, which is the complete opposite of the expected. This may suggest that EEG asymmetry values are more likely to be affected by the process of fear regulation than reactivity. Also this may question the effectiveness of so far applied fear elicitor tools like IAPS or short movie clips from the movie The Shining. I intent to publish the results after gathering more data and analyzing asymmetry values with a finer method than just looking at means. I hope this study sheds light on the potential of video-games as emotion elicitors as well as the flaws of both the EEG asymmetry hypothesis and previously applied elicitors.

Heart rate means over all subjects within situations. Increase in HR shows imply increase in arousal, which in this case indicates increase in fear intensity.

Skin conductance response means over all subjects within situations. SCR - the phasic part of GSR - reflects instant changes in arousal – particularly to measure response to stimulus.

Frontal asymmetry means over all subjects within situations. A drop can be seen in asymmetry at the night game situation, which indicates the opposite of the alpha asymmetry hypothesis held so far, though the difference in means is not an appropriate measure to reflect this asymmetry phenomena - as EEG asymmetry change in a fast pace.

SCR (blue) and frontal EEG asymmetry (green) values of Subject #9. Here the open-eyed, closed-eyed, daygame and nightgame situations are concatenated. SCR increases are accompanied by EEG asymmetry drops under the fear inducing gameplay (after 250 sec), which contradicts the so far held hypothesis.

Thesis