EEG-Based Evaluation of the Effect of Emotion on Relaxation Management

Emotion is a vital part of our lives in our interactions with the environment. With more hectic lifestyles, relaxation is critical to re-energising our mind. Emotional adjustment and relaxation bring benefits such as reduced anxiety and improved task performance. Studies have shown the relationship between emotions and relaxation in applying relaxation techniques to reduce negative emotional states, such as stress. In contrast, our study investigated, using EEG features, how different emotional states, fear, anxiety and happiness, affect one’s ability to relax using EEG features. We collected EEG from 15 participants according to the experiment protocol comprising baseline, emotion-arousing “non-relaxed” tasks (happiness, fear and anxiety), each followed by a “relaxed task”. We used six band-power features extracted from EEG signals applying statistical analysis methods to test research hypotheses. We applied band-power with statistical analysis methods to test our research hypotheses. Using paired samples t-test, individual band-power features (alpha, theta, beta) of “non-relaxed” tasks compared with that of baseline tasks showed a highly significant difference (p < 0.001). We then used supervised machine learning to test binary classification accuracy of relaxation state among different task pairs. The results showed lower 9.61% mean accuracy of modified tasks compared with baseline task. Baseline task pair achieved 82.98% while only obtained 74.54%, 73.52% and 72.06% accuracy for anxiety, fear and happiness non-relaxed and relaxed task pairs, respectively. Our results showed that emotions affect one’s ability to relax and to varying degrees. These findings could allow a better understanding of what it is required to neutralise emotion effects and enhance relaxation.