Unsupervised learning of brain state dynamics during emotion imagination using high-density EEG

Area of Science:

• Neuroscience
• Cognitive Science
• Signal Processing

Background:

• Understanding the neural basis of emotion is crucial for mental health.
• Electroencephalography (EEG) offers high temporal resolution for studying brain dynamics.
• Independent Component Analysis (ICA) is a powerful tool for dissecting complex EEG signals.

Purpose of the Study:

• To apply Adaptive Mixture Independent Component Analysis (AMICA) to high-density EEG data during imagined emotions.
• To identify spatiotemporal EEG patterns and their transitions associated with specific emotional states.
• To explore the neural correlates of emotion imagination and their implications for emotion decoding.

Main Methods:

• Utilized 20-model AMICA decomposition on long-duration (1-2h), 128-channel EEG data.
• Analyzed EEG recorded during guided imagination of 15 distinct emotions.
• Examined model probability transitions and spatial distributions of independent component processes (ICs).

Main Results:

• AMICA models identified distinct spatiotemporal EEG states within emotion imagination periods.
• Transitions between EEG states varied across emotions, with 'grief' and 'happiness' showing abrupt changes.
• Specific brain regions, including prefrontal cortex and insula, showed differences in IC spatial distributions during emotion imagination versus relaxation.

Conclusions:

• AMICA modeling of EEG dynamics provides data-driven insights into brain activity during emotional experiences.
• Findings suggest potential for improved EEG-based emotion decoding and a deeper understanding of emotion.
• The study highlights the utility of advanced signal processing techniques for neuroscientific research.