A01Imaging learning rule from neural and molecular activity

 Artificial Intelligence (AI) has been developed to solve a specific problem. In contrast, our brain can solve a wide range of problems. For the AI to expand its capacity toward “generalized AI”, uncovering the learning mechanism of the brain would be highly beneficial. For this purpose, we focus on the interplay between molecular and electrical activity of neurons. For the last few decades, the development of molecular techniques identified several key molecules important for learning and memory. For example, transcription factor CREB (cAMP responsive element binding protein) plays an important role in long-term potentiation of synaptic inputs; a genetically engineered mouse with activated form of CREB has enhances long term memory. These molecular techniques can alter the functionality of a certain gene and observe the changes in animal behavior. However, to provide the mechanistic understanding of how a certain molecular activity influences the brain function, it would be necessary to understand how molecular activity influences electrical activity of neurons in vivo. We are working on a simultaneous imaging of molecular activity and Calcium imaging in vivo by using a two-photon microscope. We are envisaging that this approach could provide an insight into how the brain, and possibly also AI, can efficiently learn a wide range of tasks in the real world.