Digital twin brain simulator for real-time consciousness monitoring and virtual intervention using primate electrocorticogram data.
NPJ digital medicine February 10, 2025 Yuta Takahashi, Hayato Idei, Misako Komatsu et al. 21 citations
A real-time electrocorticogram (ECoG) simulator based on a digital twin brain concept was developed. Using a Variational Bayesian Recurrent Neural Network with hierarchical latent units, the model dynamically predicted ECoG signals by assimilating data from macaque monkeys in awake and anesthetized conditions. The model updated its latent states in real-time, improving simulation precision. Self-organization of latent states reflected brain states and individuality, enabling simulation of virtual drug administration and revealing functional networks underlying anesthesia-induced changes. The simulator achieves high-accuracy real-time brain signal simulation and helps uncover underlying information processing dynamics.