Title :
Model synthesis identification a Hodgkin-Huxley-type neuron model
Author :
Csercsik, David ; Szederkenyi, Gabor ; Hangos, Katalin M. ; Farkas, Imre
Author_Institution :
Syst. & Control Lab., Comput. & Autom. Res. Inst., Budapest, Hungary
Abstract :
GnRH neurons are key elements of the reproductive neuroendocrine system and play important central regulating role in the dynamics of the hormonal cycle. A conductance-based Hodgkin-Huxley model structure is proposed in this paper in the form of nonlinear ordinary differential equations that is able to take into account up-to-date biological literature data related to ion channels. Measurement data were available for parameter estimation, which are originated from laboratory experiments done in the Institute of Experimental Medicine of the Hungarian Academy of Sciences in the form of whole cell patch-clamp recordings. The proposed neuron model is highly nonlinear in parameters and the evaluation of the objective function is computationally expensive, therefore the asynchronous parallel pattern search (APPS) procedure has been used for identification which is a gradient-free optimization method that can handle linear equality and inequality constraints and has advantageous convergence properties. The model with high number of estimated parameters provides a qualitatively good fit of both voltage clamp and current clamp traces.
Keywords :
constraint handling; neural nets; neurophysiology; nonlinear differential equations; optimisation; parameter estimation; search problems; APPS procedure; Hodgkin-Huxley-type neuron model; Hungarian Academy of Sciences; Institute of Experimental Medicine; asynchronous parallel pattern search procedure; biological literature data; cell patch-clamp recordings; conductance-based Hodgkin-Huxley model structure; convergence properties; current clamp traces; gradient-free optimization method; hormonal cycle; laboratory experiments; linear equality constraint handling; linear inequality constraint handling; measurement data; model synthesis identification; neuron model; nonlinear ordinary differential equations; parameter estimation; reproductive neuroendocrine system; voltage clamp traces; Biochemistry; Clamps; Current measurement; Electric potential; Mathematical model; Neurons; Voltage measurement;
Conference_Titel :
Control Conference (ECC), 2009 European
Conference_Location :
Budapest
Print_ISBN :
978-3-9524173-9-3