Events
Online Parameter Estimation using GPU super-computing
Friday 3 October 13:00 until 14:00
Pevensey I, 1A6.
Speaker: Thomas Nowotny
Part of the series: Work In Progress Seminars: School of Engineering and Informatics
Abstract
The standard method for characterising ion channels in neurons is voltage clamp. However, in the classical procedure, measurements are performed with constant voltage steps and chemical channel blockers are used to isolate individual ion channel types. Because chemical blockers can be irreversible, different ion channels have to be measured in different individual cells, sometimes even in preparations from different individual animals.
Using this paradigm it has been observed in identified cells of invertebrates, previously believed to be prototypical across individual animals, that the parameters of channels, in particular the expression levels (maximal conductances of channels), can vary largely. This raises the question how the nervous system can function reliably in the face of the apparent large variability of its components. One of the main reasons for this gap in our knowledge is the inability of measuring all ionic conductances together in one individual neuron.
Here I present work in progress that holds a promise to go beyond classical voltage to isolate the effect of different ion channels without chemical blockers. Furthermore, I propose to use closed-loop online parameter estimation methods to build a model of all ionic currents in an individual neuron simultaneously. I will present preliminary results obtained in simulations that such an approach is feasible in principle.
By: Luke Scott
Last updated: Wednesday, 1 October 2014