Electrophysiological representations of multivariate human emotion experience

Area of Science:

• Neuroscience
• Cognitive Science
• Psychology

Background:

• Traditional neuroscience studies often use a univariate approach to understand emotion, focusing on single emotion categories.
• This overlooks the natural co-occurrence of multiple emotions in human experience.
• A multivariate perspective is needed for a comprehensive understanding of emotion's neural basis.

Purpose of the Study:

• To investigate the neural representations of multivariate emotion experience.
• To explore how multiple emotions are simultaneously represented in the brain.
• To apply inter-situation representational similarity analysis (RSA) to electroencephalography (EEG) data.

Main Methods:

• Utilized an EEG dataset from 78 participants viewing 28 video clips.
• Participants rated their experience across eight emotion categories.
• Extracted EEG-based electrophysiological representations using power spectral density (PSD) features in five frequency bands.

Main Results:

• Inter-situation RSA identified significant correlations between multivariate emotion experience ratings and PSD features in Alpha and Beta bands.
• These neural representations were primarily observed in frontal and parietal-occipital brain regions.
• Ablation analyses confirmed the reliability, stability, and specificity of these EEG representations.

Conclusions:

• A multivariate approach is crucial for accurately understanding the neural representation of human emotion experience.
• EEG, analyzed with inter-situation RSA, can reliably capture the neural correlates of co-occurring emotions.
• Findings underscore the complexity of emotion processing in the brain.