Recurrent Spiking Neural Networks for Continuous Motor Control: A Biologically Plausible Model of Motor Cortex
Motor cortex acquires skilled movements through recurrent spiking dynamics and experience-dependent synaptic plasticity, yet how such circuits learn continuous motor control using biologically local signals remains unclear. Here, we develop a recurrent spiking neural network model of motor cortex that addresses both the computational and biological requirements of motor learning. We first…
The article explores the development of a recurrent spiking neural network model to replicate the continuous motor control exhibited by the brain's motor cortex. These networks resemble biological processes of skilled movement acquisition and experience-driven synaptic changes. The authors first derive the recurrent firing-rate state from synaptic input and train the network using implicit differentiation, bypassing backpropagation through extended spike trajectories.
State-Carry Initialization is implemented to facilitate closed-loop motor control by perpetually carrying the recurrent state across multiple actions. A learning rule called Feedback-Spike Nudging (FSN) is then devised, wherein teaching signals are transmitted as spikes via a feedback pathway. The synaptic alterations in this model are contingent solely on presynaptic activity and the feedback-induced shift in postsynaptic firing, mirroring biological mechanisms of synaptic plasticity within motor cortex.
Subsequent experimentation reveals that these networks can learn continuous motor control through supervised and reinforcement learning methods. These learning methods can be mapped onto the interactions of motor cortex with the cerebellum and basal ganglia, respectively. Furthermore, the Feedback-Spike Nudging learning rule aligns with experimental findings on structural spine plasticity in mouse motor cortex.
Ultimately, the study establishes a foundation for recurrent spiking models of motor cortex that are both functionally and anatomically grounded in biological reality.
Written by urgent.news from bioRxiv's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
