Adversarial AI reveals mechanisms and treatments for disorders of consciousness - PubMed
6 hours ago
- #neuroscience research
- #adversarial AI
- #disorders of consciousness
- An adversarial AI framework was developed to simulate conscious and comatose brain states using neuroelectrophysiology data from over 680,000 samples and validated on 565 subjects.
- The model recapitulated empirical neurophysiological features across humans, monkeys, rats, and bats, retrodicting known responses to brain stimulation in disorders of consciousness.
- Predictions validated include disruption of the basal ganglia indirect pathway (supported by diffusion MRI in 51 patients) and increased cortical inhibitory-to-inhibitory synaptic coupling (supported by RNA sequencing and rat models).
- High-frequency stimulation of the subthalamic nucleus was identified as a promising intervention, supported by electrophysiological data from human patients.
- This work presents an AI approach for causal inference and therapeutic discovery in consciousness research and complex systems.