Synaptic variability in a cortical neuromorphic circuit

Area of Science:

• Neuromorphic Engineering
• Computational Neuroscience
• Materials Science

Background:

• Biological neurons exhibit variable behavior, offering potential for neuromorphic circuits.
• Carbon nanotube (CNT) transistors are explored for scalable neuromorphic applications.
• Sources of neural variability may be stochastic or chaotic.

Purpose of the Study:

• To present a neuromorphic cortical neuron with synaptic neurotransmitter-release variability.
• To investigate the impact of stochastic and chaotic variability on neural circuit behavior.
• To embed chaotic behavior into neuromorphic circuits using CNT transistors.

Main Methods:

• Designed and simulated a neuromorphic neuron using carbon nanotube (CNT) transistors.
• Embedded Gaussian noise or a chaotic signal into neuromorphic/synaptic circuits.
• Developed and simulated a chaotic signal generator circuit using CNT transistors and switched-current circuits.

Main Results:

• Neurotransmitter-release variability beneficially impacts spike generation reliability.
• Spike timing precision in CNT circuit simulations depends on postsynaptic potential transients.
• High neurotransmitter release variability yields reproducible spike train timing.

Conclusions:

• Neurotransmitter release variability is a key factor in reliable spike generation for neuromorphic systems.
• The developed CNT-based chaotic circuit can be integrated into electronic neural circuits.
• This research contributes to the Biomimetic Real-Time Cortex project by advancing neuromorphic hardware.