Robot-Embodied Neuronal Networks as an Interactive Model of Learning

Area of Science:

• Neuroscience
• Bioengineering
• Robotics

Background:

• Neuronal cell cultures offer a reductionist approach to studying synaptic signaling and neuronal development.
• Existing ex vivo neuronal networks lack sensory input capabilities, limiting their resemblance to in vivo systems.
• Multi-electrode arrays facilitate recording and stimulation of cultured neuronal networks.

Purpose of the Study:

• To embody ex vivo neuronal networks within a closed-loop cybernetic system.
• To enable neuronal networks to receive sensory input and generate motor output.
• To explore applications in brain-computer interfaces and disease modeling.

Main Methods:

• Integration of ex vivo murine cortical neuronal networks into a cybernetic system.
• Utilized digitized video signals as sensory input for the neuronal network.
• Employed a robot arm as the motor output controlled by the neuronal network.

Main Results:

• The embodied neuronal network functioned as a rudimentary central nervous system.
• The network demonstrated the ability to track a target in a simulated environment.
• Confirmed that ex vivo neuronal networks can respond to sensory stimuli and direct motor actions.

Conclusions:

• Findings support the development of neuronal-computer interfaces for prosthetic applications.
• Ex vivo networks show altered signaling patterns upon exposure to amyloid-beta.
• Embodied networks provide a model for studying neurodevelopment and neurotoxicity.