Title: Neuro-GPT: Towards A Foundation Model For EEG

• Ref: https://ieeexplore.ieee.org/abstract/document/10635453
• Code: https://github.com/wenhui0206/NeuroGPT

• Summary: Neuro-GPT is a foundation model to handle the scarcity and heterogeneity of EEG data for BCI tasks, consisting of an EEG encoder and a GPT model.

  • Pre-trained Model: Uses a large-scale dataset for a self-supervised task (reconstructing masked EEG segments).
  • Fine-tuning: For example, fine-tuning on a motor imagery classification task for validation.

🚀 Insights: Are emergent properties of EEG foundation models beginning to emerge?