Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
6864890 | Neurocomputing | 2018 | 32 Pages |
Abstract
In this paper, a method of closed-loop interspike interval (ISI) clamp based on a computational model of single neuron is proposed. This method consists of a highly nonlinear filter-the Unscented Kalman Filter (UKF), on-line calculation of ISI response feature, and a feedback loop of controlling regularity of ISI. The regularity of ISI can be control by Proportional-Integral type iterative learning control (PI-type ILC) algorithm, and then the accuracy of calculating phase response curves (PRCs) is improved. Moreover, UKF can be suitably fitted into the proposed control frame to estimate hidden electrophysiological properties of a single neuron from membrane potentials and conduct electrophysiology analysis. The computational simulation is carried out on the Hodgkin-Huxley (HH) model in the presence of noise to verify the performance of closed-loop ISI clamp in controlling firing regularity of neurons. The Pinsky-Rinzel (PR) model is also used to conduct electrophysiological experiment and qualitatively analyze the roles of certain electrophysiological properties of a single neuron in shaping different firing patterns. Simulation results show the efficiency of closed-loop ISI clamp in controlling firing regularity of neurons and the potential application in electrophysiological analysis.
Related Topics
Physical Sciences and Engineering
Computer Science
Artificial Intelligence
Authors
Shanshan Li, Guoshan Zhang, Jiang Wang, Bin Deng,