Examining the Bridge Between Neuroscience and Artificial Intelligence

A new analysis covering approximately 458 million works in the OpenAlex database has revealed that the connection between neuroscience and artificial intelligence is real and increasingly strengthening. However, it is noted that this relationship remains narrow in scope and tool-oriented.

Artificial intelligence and neuroscience have a long history, with many of the founding ideas of AI attempting to mimic the brain's functioning and reverse-engineer its mechanisms. Over the past decade, prominent figures and institutions—including Geoffrey Hinton, John Hopfield, and Demis Hassabis, who won the 2024 Nobel Prize for his work in AI—have argued that the two fields should nourish one another.

Under the umbrella of NeuroAI, it is suggested that neuroscience can provide principles, architectures, and algorithms to AI, while AI can contribute robust foundations and tools. An examination of approximately 458 million works in the OpenAlex data confirms the existence and momentum of the bridge between the two fields, yet it also shows that this interaction is mostly limited to specific institutions and research groups, and that AI is leveraging neuroscience concepts acquired decades ago.