Cross-Subject EEG Emotion Recognition Based on Temporal Asynchronous Alignment Contrastive Learning
Researchers have developed a new framework called Temporal Asynchronous Alignment-based Contrastive Learning (TA2CL) to improve cross-subject electroencephalography (EEG) emotion recognition. This method addresses the challenge of temporal misalignment in EEG signals between different individuals by employing a fine-grained local matching mechanism, inspired by NLP techniques. The TA2CL framework adaptively aligns segments of EEG data, effectively reducing the impact of inter-subject differences and temporal delays. Experiments on public datasets like FACED, SEED, and SEED-V show significant performance gains, with accuracies reaching up to 86.4% on the SEED dataset. AI
IMPACT Introduces a novel contrastive learning approach for EEG emotion recognition, potentially improving human-computer interaction systems.