 AINCE Metanet is a novel framework for classifying brain states under anesthetic conditions. It consists of two main components, CNNs to extract power spectral features from the EEG signal and LSTMs to capture temporal dependencies. A meta-learning framework is then employed to handle large cross-subject variability. This framework is trained in multiple stages, with each stage focusing on a specific aspect of the problem. Visualizations of the high-level feature maps show that the proposed method outperforms existing approaches. This article was authored by Qi Hong-Wang, Feng Liu, Gu Wei Hong-Wang, and others.