Brain-Inspired Spiking Neural Network Controller for a Neurorobotic Whisker System

Area of Science:

• Neuroscience
• Robotics
• Computational Biology

Background:

• Animals use self-generated movements for active sensing, with rodents using whiskers for tactile exploration.
• The mouse whisker system is a key model for understanding active sensing and sensorimotor integration.
• Previous models have not fully captured the complexity of the whisker sensorimotor feedback loop.

Purpose of the Study:

• To develop a bioinspired spiking neural network model of the mouse whisker system.
• To integrate this model into a virtual robot for testing.
• To enable the model to exhibit active whisking with learning capabilities.

Main Methods:

• Developed a spiking neural network model of the peripheral whisker system, including trigeminal ganglion, nuclei, and central pattern generators.
• Embedded the neural network in a virtual mouse robot using the Human Brain Project's Neurorobotics Platform.
• Connected the whisker system to an adaptive cerebellar network controller.

Main Results:

• The developed system successfully drove active whisking in the virtual mouse robot.
• The model demonstrated learning capabilities in its whisking behavior.
• The simulated neural activity and behavior matched experimental recordings from mice.

Conclusions:

• The bioinspired spiking neural network effectively models the mouse whisker sensorimotor system.
• This approach enables the study of active sensing and sensorimotor integration in a neurorobotic context.
• The model provides a platform for understanding neural control of active behaviors.