## Time constant modeling

The basics of how to develop, test, and use models.
yvp

### Time constant modeling

Hi,

I have a very basic question on how time constants are modeled for voltage gated channels. I have a tau vs. voltage bell shaped curve from voltage clamp data that I am trying to model using either the direct HH form with double exponentials or the Borg Graham method as in the modelDB examples. I tried running a 5 and 6 parameter fit on MRF using my data but the error was infinite. Also, the Borg Graham form asks for parameters like effective valence and gating asymmetry that I am not sure how to get from the data. Can you please explain how the time constant curves are usually fitted in the models?

Thanks
ted
Posts: 5819
Joined: Wed May 18, 2005 4:50 pm
Location: Yale University School of Medicine
Contact:

### Re: Time constant modeling

yvp wrote:I tried running a 5 and 6 parameter fit on MRF using my data but the error was infinite.
Many steps are involved in setting up and using the MRF, and it is easy to make one or more minor errors that can have disastrous results. You might find it helpful to begin with an example that has a known solution, and use this to verify that the MRF has been properly configured. Always try to write equations and formulas in such a way that optimizable parameters will be positive definite (i.e. always > 0); this will allow the use of logarithmic scaling with helps accelerate convergence to a (possibly local) minimum.

Code: Select all

``Also, the Borg Graham form asks for parameters like effective valence and gating asymmetry that I am not sure how to get from the data. Can you please explain how the time constant curves are usually fitted in the models?``
Do your data really require this extra complexity? The more parameters that must be tuned, the more difficult the optimization becomes. Start with the simplest case first: plain vanilla hh-style formulas (if you're looking for parameterized equations, check out the ones embedded in the CellBuilder).