Title :
Parameter estimation methods for neural models
Author :
Durnd, D.M. ; Tawfik, Bassel ; Lin, J.C.
Author_Institution :
Dept. of Biomed. Eng., Case Western Reserve Univ., Cleveland, OH, USA
Abstract :
Two methods are presented for the estimation of the electrotonic parameters of the somatic shunt cable model. In one method, a matrix is built using known properties of the parameter space and the parameters are estimated by interpolation. This method has proved to be successful at avoiding local minima, provided “good” estimates in the presence of noise and missing data points. A second method which combines the advantages of both iterative and associative method is also presented. These methods are applied to the somatic shunt cable model of neurons and the accuracy of the estimated parameters compared
Keywords :
bioelectric phenomena; interpolation; inverse problems; neurophysiology; parameter estimation; physiological models; electrotonic parameters; estimated parameters accuracy; matrix; missing data points; neural models; noise; parameter estimation methods; parameter space properties; somatic shunt cable model; Associative memory; Biomedical engineering; Gradient methods; Interpolation; Inverse problems; Iterative methods; Neurons; Parameter estimation; Recursive estimation; Vectors;
Conference_Titel :
[Engineering in Medicine and Biology, 1999. 21st Annual Conference and the 1999 Annual Fall Meetring of the Biomedical Engineering Society] BMES/EMBS Conference, 1999. Proceedings of the First Joint
Conference_Location :
Atlanta, GA
Print_ISBN :
0-7803-5674-8
DOI :
10.1109/IEMBS.1999.802473
