Neural computing model can shed light on the development of cognitive abilities

A new study presents a new neural computing model of the human brain that could shed light on how the brain develops complex cognitive abilities and advances neural artificial intelligence research. Published September 19, the study was carried out by an international group of researchers from Institut Pasteur and Sorbonne University in Paris, CHU Sainte-Justine, Mila – Quebec Institute of Artificial Intelligence, and the University of Montreal.

The model who made the magazine cover Proceedings of the National Academy of Sciences of the United States of America (PNAS), describes neural development at three hierarchical levels of information processing:

  • The first sensory level explores how the brain’s internal activity learns patterns from perception and relates them to action;
  • Cognitive level examines how the brain combines these patterns in terms of context;
  • Finally, the conscious level studies how the brain disconnects from the outside world and manipulates patterns acquired (via memory) that are no longer available for perception.

The team’s research provides clues to the underlying mechanisms underlying cognition thanks to the model’s focus on the interaction between two fundamental types of learning: hippy learning, which is associated with statistical regularity (i.e. repetition)—or, as neuropsychologist Donald Hebb put it, “neurons that fire together, wire together.” — and reinforcement learning, linked to reward and the neurotransmitter dopamine.

The model solves three tasks of increasing complexity across those levels, from visual recognition to cognitive manipulation of conscious perception. Each time, the team introduced a new core mechanism to enable it to advance.

The results highlight two basic mechanisms for the multi-level development of cognitive abilities in biological neural networks:

  • Synaptic epigenetics, with Hebbian learning at the local level and reinforcement learning at the global level;
  • and dynamics of self-regulation, through spontaneous activity and a balanced excitatory/inhibitory ratio of neurons.

Our model demonstrates how the convergence of AI and neural intelligence highlights biological mechanisms and cognitive structures that can fuel the development of the next generation of AI, and ultimately lead to AI awareness.”

Guillaume Dumas, team member, associate professor of computational psychiatry at UdeM, and principal investigator at the CHU Sainte-Justine Research Center

He added that reaching this achievement may require integrating the social dimension of cognition. Researchers are now looking at integrating biological and social dimensions into human cognition. The team has already pioneered the first simulation of whole brains in interaction.

The team believes that embedding future computational models into biological and social realities will not only continue to shed light on the underlying mechanisms underlying cognition, but will also help provide a unique bridge for AI toward the only system known to be advanced social consciousness: the human brain.


Journal reference:

Volzhinin, K., et al. (2022) Multilevel development of cognitive abilities in an artificial neural network. PNAS.