• Top page
  • Timetable
  • Per session
  • Per presentation
  • How to
  • Meeting Planner



Neural Network Modeling

開催日 2014/9/11
時間 16:00 - 17:00
会場 Poster / Exhibition(Event Hall B)

Parameter Estimation of Nav1.6 Channel Model by Differential Evolution Algorithm

  • P1-368
  • 北野 勝則 / Katsunori Kitano:1 山岸 勇人 / Hayato Yamagishi:2 
  • 1:立命大・情理工 / Dept Human & Comput Intel, Ritsumeikan Univ, Shiga, Japan 2:立命大院・情理工研 / Grad Sch Inform Sci & Eng, Ritsumeikan Univ, Shiga, Japan 

Subthamamic nucleus (STN) is one of the structures in the basal ganglia, which processes motor commands generated in the motor cortex and conveys them to the output nucleus of the basal ganglia, globus pallidus internus. Neurons in the STN exhibit distinctly different firing patterns between in the normal and the pathological condition occurred in patients of the Parkinson's disease; STN neurons show irregularly tonic firings in the normal condition whereas those evoked rhythmic bursts in the pathological condition. Because the activity patterns of STN neurons are highly relevant with the behavioral functions, it is of importance to elucidate functional roles of firing patterns of STN neurons. An STN neuron in vitro show an autonomous firing without an external input and can generate action potentials at a very high frequency over 100 spikes/s when an intensive external input is applied (Otsuka et al., 2004). Nav1.6 sodium channels are expressed on an STN neuron's membrane, which is one of the channels responsible for such a high frequency spiking activity (Do and Bean, 2003). The characteristic of the sodium current through Nav1.6 channels is the resurgent current that is the depolarizing current immediately after an action potential. To model the channels based on the Hodgkin-Huxley formalism, we require an additional gate variable except usual m and h gates (Ogata and Ohishi, 2002). Because there is no standard model of this type of channels, we developed the channel model by estimating parameters for its kinetics from voltage clamp data of the channel. To automate the parameter estimation, we formulated it as an optimization problem in which errors between the experimental and simulated voltage trace characteristics should be minimized. We applied the differential evolution, one of evolutional computations such as the genetic algorithm, to our multi-compartment model of a STN neuron that includes the Nav1.6 channels (Buhry et al., 2011). Our result showed that the obtained parameters for the Nav1.6 kinetics could generate the voltage traces similar to the experimental ones at most of holding potentials, suggesting that this formulation and method are quite useful to develop a computational model of a neuron.

Copyright © Neuroscience2014. All Right Reserved.