An ion-electronic hybrid artificial neuron with a widely tunable frequency

Area of Science:

• Materials Science
• Neuroscience
• Electrical Engineering

Background:

• Biological nervous systems utilize diverse spiking frequencies for information processing.
• Replicating this broad frequency range in artificial neurons is crucial for emulating biosignal diversity but remains a significant challenge.
• Existing artificial neurons often lack the adaptability and sensing capabilities of biological counterparts.

Purpose of the Study:

• To develop a novel ion-electronic hybrid artificial neuron with a wide, tunable spiking frequency.
• To integrate inherent sensing capabilities for environmental stimuli.
• To explore applications in e-skin, neurointerfaces, and pattern recognition.

Main Methods:

• Compact integration of a nonlinear electrochemical element with a solid-state memristor to create a minimalist hybrid neuron architecture.
• Utilizing stimuli-dependent ion fluxes for inherent afferent sensing.
• Connecting the hybrid neuron array to biomotor nerves for muscle actuation and testing pattern recognition capabilities.

Main Results:

• The hybrid artificial neuron demonstrated a tunable spiking frequency spanning five orders of magnitude, exceeding the capabilities of purely electronic devices.
• Inherent sensing of liquid flow, temperature, and chemical constituents was achieved through ion fluxes, negating the need for external sensors.
• Successful muscle actuation via biomotor nerve connection and recognition of handwritten patterns using a hybrid neuron array were demonstrated.

Conclusions:

• The developed ion-electronic hybrid artificial neuron offers a promising platform for advanced neuro-mimetic technologies.
• The device's broad frequency tunability and integrated sensing capabilities represent a significant advancement over traditional artificial neurons.
• This design paves the way for next-generation electronic skin (e-skin) and neurointerface applications.