Emotion Recognition with Portable EEG in Immersive 360-Degree Environment.

Area of Science:

• Neuroscience
• Human-Computer Interaction
• Affective Computing

Background:

• Traditional emotion recognition studies often use artificial stimuli presentation.
• Portable electroencephalography (EEG) offers a promising avenue for real-world emotion detection.
• Immersive 360-degree environments enhance the naturalism of sensory experiences.

Purpose of the Study:

• To assess the feasibility of using portable single-channel dry electrode EEG headbands for emotion identification.
• To distinguish human emotions elicited by multimodal stimuli in a 360-degree immersive environment.
• To compare the effectiveness of different machine learning models and differential entropy features for emotion classification.

Main Methods:

• Development of a multimodal stimulation paradigm within a 360-degree immersive environment.
• Recording of electroencephalography (EEG) data using portable single-channel dry electrode headbands.
• Application of differential entropy (DE) features for EEG signal analysis and machine learning (ML) classification.
• Subjective emotional state assessment via self-rating scales.

Main Results:

• Single-channel EEG signals, after artifact removal and DE feature application, effectively distinguished various emotional states.
• Machine learning models demonstrated significant performance in classifying emotions elicited in the immersive environment.
• Differential entropy proved to be a robust feature for capturing EEG dynamics related to emotional changes.

Conclusions:

• Single-channel EEG in a 360-degree immersive setting is a feasible approach for naturalistic emotion recognition.
• This methodology combines the benefits of controlled experimental conditions with ecological validity.
• The portability and convenience of EEG headbands support potential multi-user applications in affective computing.