Emotional brain network decoded by biological spiking neural network

Area of Science:

• Neuroscience
• Computational Neuroscience
• Biomedical Engineering

Background:

• Emotional disorders are linked to neurological and psychiatric diseases, necessitating better diagnostic and therapeutic tools.
• Bi-directional brain-computer interfaces (BCIs) show promise for aiding patients, but require clearer understanding of functional brain areas and biological markers for emotions.
• The dynamic connection mechanisms within emotional brain networks remain largely unknown.

Purpose of the Study:

• To identify emotional electroencephalography (EEG) brain networks and biological markers for different emotions.
• To explore the efficacy of spiking neural networks (SNNs) with binary coding for emotion recognition.
• To analyze dynamic connections and biological rhythms associated with fear, sadness, and happiness.

Main Methods:

• Collected EEG data from participants watching emotional videos (fear, sadness, happiness, neutrality).
• Utilized a spiking neural network algorithm with binary coding to analyze EEG data and identify emotional brain networks.
• Examined dynamic connections between electrodes and biological rhythms, focusing on the alpha (α) frequency band.

Main Results:

• Distinct brain network localizations were found: parietal lobe for fear/sadness, prefrontal-temporal-central areas for happiness.
• The alpha (α) frequency band served as a biological marker for negative emotions and happiness.
• High decoding accuracies were achieved: 86.36% for fear, 95.18% for sadness, and 89.09% for happiness, demonstrating effective emotional decoding.

Conclusions:

• The self-backpropagation mechanism significantly enhances SNN performance in emotion decoding.
• Specific EEG networks and alpha frequency band markers differentiate emotions, offering insights for BCI development.
• These findings provide crucial information for advancing BCIs to aid recovery from brain diseases associated with emotional disorders.