Ephaptic Coupling in Ultralow-Power Ion-Gel Nanofiber Artificial Synapses for Enhanced Working Memory

Area of Science:

• Materials Science
• Neuroscience
• Computer Science

Background:

• Neuromorphic devices aim to replicate biological neural networks' energy efficiency.
• Existing devices often neglect ion transport, limiting biomimicry and complex connectivity like ephaptic coupling.

Purpose of the Study:

• To develop an ionic biomimetic synaptic device focusing on ion transport mechanisms.
• To enable functionalities such as ephaptic coupling and synaptic memory effects.

Main Methods:

• Fabrication of a flexible ion-gel nanofiber network for ionic synaptic devices.
• Characterization of ion transport behavior and energy consumption (6 fJ).
• Integration into reservoir computing systems for MNIST dataset classification and edge learning.

Main Results:

• The device exhibits synaptic-like memory effects due to hysteretic ion transport.
• Achieved high efficiency in MNIST handwritten digit classification and edge learning.
• Device arrays demonstrated global oscillatory behavior mimicking biological ephaptic coupling.

Conclusions:

• The proposed ionic synaptic device successfully replicates key biological neural network features, including ion transport and ephaptic coupling.
• This technology enhances reservoir computing performance and enables working memory tasks.
• Paves the way for complex, brain-like computing systems with vast connectivity.