Information And Control: Insights from within the brain

TL;DR

This paper explores how the brain's neural networks learn, self-organize, and integrate multisensory information, providing insights for biologically inspired control in robotics.

Contribution

It offers new understanding of multisensory integration and self-organization in neural networks, with implications for control systems in robotics.

Findings

Neural networks learn statistical regularities through synaptic changes.

Multisensory integration occurs in the somatosensory cortex for complex behaviors.

Self-organization enhances learning and coherent representation of environments.

Abstract

The neural networks of the brain are capable of learning statistical input regularities on the basis of synaptic learning, functional integration into increasingly larger, interconnected neural assemblies, and self organization. This self organizing ability has implications for biologically inspired control structures in robotics. On the basis of signal input from vision, sound, smell, touch and proprioception, multisensory representations for action are generated on the basis of physically specified input from the environment. The somatosensory cortex is a brain hub that delivers a choice example of integration for multifunctional representation and control. All sensory information is in a first instance topologically represented in the biological brain, and thereafter integrated in somatosensory neural networks for multimodal and multifunctional control of complex behaviors. Multisignal input triggers interactions between visual, auditory, tactile, olfactive, and proprioceptive mechanisms, which cooperate or compete during learning, and contribute to the formation of integrated representations for action, reflection, and communication between the agent and the outside world. Interaction fuels complex behavioral strategy deployment and further learning for increasingly coherent representation of intrinsically ambiguous physical environments.