The Blue Brain Neurorobotics section investigates reconstructed brain and circuit models in closed sensory-motor loops.
Neural models in closed sensory-motor loops
BBP closed-loop studies use two neural models: first a simplified version of the neocortical microcircuit model, which is reduced to a point neuron-network, while preserving most of the connectivity, and the synaptic properties; second, a reconstructed whole-brain model of the mouse, based on the Allen Brain Atlas, also simulated at the point neuron level. Network simulations mainly use NEST (http://www.nest-simulator.org). Since even with NEST real-time simulation of large neural systems is still out of reach, the studies rely on simulated bodies, robots, and environments. The BBP’s latest body model is a virtual mouse, which is used to study the mapping between the central nervous system and the peripheral sensors and muscles.
Closed-loop models help to validate and improve BBP models of neural circuitry by integrating dynamic behavioural data, and can also be used to explore innovative neuro-inspired control strategies for physical robots.
The HBP Neurorobotics Platform
The BBP Neurorobotics team is a leading contributor to the Human Brain Project’s Neurorobotics Platform. The platform will provide a software and hardware infrastructure that allows researchers, like neuroscientists, who are not specialized in robotics to perform in silico experiments. Such experiments have the potential to provide valuable insights into the links between low level brain circuitry and high-level cognitive and behavioural functions and will facilitate the development of new types of robot controller derived from these insights.