www.neuron.yale.edu

Posted: **Mon Jun 18, 2007 8:38 am**

Hi All,

I'm trying to build the "Naundorf model" for a single compartment with the following kinetics equations

BREAKPOINT  {
    SOLVE scheme METHOD sparse
    gHHO = gHHObar*O
    ina = gHHO*(v-ena)
}
KINETIC scheme{
    ~C <-> O (alpha_a(v+K*J*O),beta_a(v+K*J*O))
    ~I <-> C (alpha_ci(v),beta_ci(v))
    ~O <-> I (1/tau_i,0)
    CONSERVE C + O + I = 1
}

Is it syntactically correct?
Is the use of state var O in the rate function (alpha/beta) permited? Does it raise integration method issues?
Does the following use of TMP help or make it worse?

Code: Select all

BREAKPOINT  {
    TMP = O
    SOLVE scheme METHOD sparse
    gHHO = gHHObar*O
    ina = gHHO*(v-ena)
}
KINETIC scheme{
    ~C <-> O (alpha_a(v+K*J*TMP),beta_a(v+K*J*TMP))
    ~I <-> C (alpha_ci(v),beta_ci(v))
    ~O <-> I (1/tau_i,0)
    CONSERVE C + O + I = 1
}

Thanks in advance,
Albert.

Posted: **Tue Jun 19, 2007 9:10 am**

agidon20 wrote:Is it syntactically correct?
Is the use of state var O in the rate function (alpha/beta) permited? Does it raise integration method issues?
Does the following use of TMP help or make it worse?

In the same sequence:
Looks OK.
I don't see why not, as long as you declare it to be a STATE; what happened when you tried it?
Unless there is something peculiar about the rate functions, I don't see why it should.
Does TMP help what, or make what worse? I don't see any useful purpose for TMP.

Posted: **Tue Jun 19, 2007 2:38 pm**

Thanks Ted,

Sometimes I get (I run a fitting algorithm) :

Code: Select all

Convergence not achieved in maximum number of iterations
at line 71 in file Naundorf.mod:

When TMP is used, I don't get errors, but this made me doubt the results.

Posted: **Tue Jun 19, 2007 8:01 pm**

Using STATE variables in the rate terms of kinetic equations can cause
the system to be numerically unstable. With TMP the rates are
independent of the states and so there can be no such problem.
However, the solution of course is for only the voltage dependent
portion of the rate. Can you write a kinetic scheme in which
the concentration dependent term moves out of the rate and into
a state transition? The alternative is to do the whole thing in
a DERIVATIVE block and use
SOLVE scheme METHOD derivimplicit

www.neuron.yale.edu

Naundorf kinetic scheme

Naundorf kinetic scheme

Re: Naundorf kinetic scheme

Re: Naundorf kinetic scheme