Parameter estimation in Hodgkin-Huxley model with adaptive method

Neuron models are highly nonlinear and involve many electrophysiological variables and parameters, only some of the variables are easily measured experimentally, while other parameters are difficult to experimentally determine. However, successful estimation of the unknown parameters often cannot be guaranteed. In this paper, we proposed a model reference adaptive method to estimate different parameters of a neuron model (the H-H model as an example) simultaneously. Simulation result of two parameters is accurate as the values we set in the H-H model. However, overshoot occurs when we estimate three parameters with different orders of magnitude. By adjusting the coefficients to control the learning rates of different parameters, a good simulation result has been acquired efficiently. This method can also be applied to estimate parameters in other nonlinear systems.