A biophysically-based neuromorphic model of spike rate- and timing-dependent plasticity

Area of Science:

• Neuromorphic Engineering
• Computational Neuroscience
• Integrated Circuit Design

Background:

• Advances in neuromorphic engineering enable real-time emulation of neuronal dynamics using CMOS analog circuits.
• Growing interest exists in neuromorphic emulation of Hebbian learning rules like spike-timing-dependent plasticity (STDP).

Purpose of the Study:

• To propose a CMOS circuit implementation of a biophysically grounded iono-neuromorphic model for synaptic plasticity.
• To capture both spike-rate-dependent plasticity (SRDP) and STDP rules within a single model.

Main Methods:

• Developed a complementary metal-oxide-semiconductor (CMOS) circuit for iono-neuromorphic synaptic plasticity.
• Modeled NMDA receptor-dependent and intracellular calcium-mediated long-term potentiation/depression.
• Incorporated retrograde endocannabinoid signaling as a secondary coincidence detector.

Main Results:

• The iono-neuromorphic model successfully reproduces bidirectional synaptic changes (long-term potentiation/depression).
• Synaptic weight changes are stored in a nonvolatile digital format, mimicking receptor channel dynamics.
• The model captures both SRDP (BCM type) and STDP learning rules.

Conclusions:

• The proposed versatile Hebbian synapse device is suitable for neuroprosthetics, brain-machine interfaces, and neurorobotics.
• This iono-neuromorphic synapse is applicable to neuromimetic computation, machine learning, and adaptive control.
• The circuit offers a compact and power-efficient solution for implementing complex synaptic plasticity in neuromorphic systems.