Parameter Estimation Of Electrotonic Properties Of Granule Cells

Abstract

An optimization procedure is outlined to estimate the electrotonic parameters of a class of neurons that can be modelled by equivalent cylinders. The five neuronal parameters to be estimated are the electrotonic length L, the dendritic and somatic membrane time constants, the input resistance of the dendrites and the soma. A model of the electrode was also implemented adding two parameters (electrcde resistance and capacitance) to the estimation procedure. The voltage resulting from the injection into the cell of a small current pulse is first calculated in the frequency domain. An inverse FFT is then applied and the objective function calculated in the time domain. This procedure has several advantages over previous techniques and is successful at estimating the electrotonic parameters of granule cells. dendrites have approximately the same length. This model has 5 parameters: somatic resistance (Rs), somatic time constant (Ts), dendritic membrane time constant (Tm), electrotonic length and input resistance of the equivalent cylinder (L, Rd). This model was tested but was only partially successful because of the presence of a large stimulation artifact generated by the impedance of the electrode. We have developed a modification of the model (Fig. 1) which takes into account the presence of an artifact at the onset of the voltage.