News Releases & Research Results Reinforcement of the pain control system in the brain through simultaneous learning by the brain and AI agents

The results of international collaborative research and development project conducted by the Brain Information Communication Research Laboratory Group of the Advanced Telecommunications Research Institute International (ATR), the Center for Information and Neural Networks (CiNet) of the National Institute of Information and Communications Technology, Cambridge University, and Oxford University.

• An interactive neurofeedback system that allows mutual training between humans and artificial intelligence (AI) agents in computers was successfully developed to reinforce the pain control system in the brain by learning both sides.
• Specifically, participants received training for controlling the brain activity patterns of pain to explore the feasibility of developing a novel pain treatment based on the “decoded neurofeedback (DecNef)*.” As a result, the activity patterns of the pain control system in the brain were successfully modified.
  *A method for inducing specific activity patterns in a target brain region by combining functional magnetic resonance imaging (fMRI) and artificial intelligence technology.
• The results of R&D should contribute to future pain treatments (e.g., application to the development of novel chronic pain treatments).

This program was supported by the Strategic Research Program for Brain Sciences and the Strategic International Brain Science Research Promotion Program by AMED.

The results of this R&D project were published online in Current Biology on August 14.

Suyi Zhang., et al. Pain control by co-adaptive learning in a brain-machine interface Current Biology
DOI: 10.1016/j.cub.2020.07.066