Emotion Recognition from EEG during Self-Paced Emotional Imagery

Abstract

Here we present an analysis of a 12-subject electroencephalographic (EEG) data set in which participants were asked to engage in prolonged, self-paced episodes of guided emotion imagination with eyes closed. Our goal is to correctly predict, given a short EEG segment, whether the participant was imagining a positive respectively negative-valence emotional scenario during the given segment using a predictive model learned via machine learning. The challenge lies in generalizing to novel (i.e., previously unseen) emotion episodes from a wide variety of scenarios including love, awe, frustration, anger, etc. based purely on spontaneous oscillatory EEG activity without stimulus event-locked responses. Using a variant of the FilterBank Common Spatial Pattern algorithm, we achieve an average accuracy of 71.3% correct classification of binary valence rating across 12 different emotional imagery scenarios under rigorous block-wise cross-validation.

A survey of affective brain computer interfaces: principles, state-of-the-art, and challenges

“…[121] It is thus important that aBCI researchers consider improving their de-noising methods, since the mentioned artifacts carry relevant emotional information that can artificially improve the performance of patternrecognition algorithms. This was exemplified by Kothe et al [109], who showed, using dense electrode recordings, that several discriminative EEG sources were originating from muscles. Since the goal is to achieve reliable emotion recognition, those artifacts can be considered as valuable for this task.…”

“…We are aware of only two studies which used an interactive situation where the participants were playing games to induce and measure emotions. [27,103] Similarly, two studies [109,110] relied on mental imagery, which is expected to elicit emotional patterns similar to everyday interactions. In all these cases, the experiments were carried out in the lab, and there is thus a strong need to evaluate the performance of aBCI in ecological contexts.…”

Human Emotion Estimation from EEG and Face Using Statistical Features and SVM

“…In [4] the achieved accuracy of the emotional valence is about 71%. The technique relies on changes in the power spectrum of short-time stationary oscillatory EEG processes within the standard EEG frequency bands.…”

“…Average accuracy for arousal and valence, % In [5] 96.2 In [8] 36.5 In [2] 66.5 In [4] 71.3 In this paper 76.4…”

Affective brain-computer interfaces as enabling technology for responsive psychiatric stimulation

“…ECOG signals offer excellent temporal resolution and may be able to leverage decoders originally developed for electroencephalography (EEG). Non-invasive EEG has been a very successful approach in affective BCI, with some real-time decoding of emotional information [16,17]. Uncertainty arises because all successful EEG affective decoding has been demonstrated in healthy volunteers.…”