Hi, I am trying to implement a published single compartmentmodel in Neuron or Matlab that is not available on Modeldb:
CLAYTON T. DICKSON, JACOPO MAGISTRETTI, MARK H. SHALINSKY, ERIK FRANSEN, MICHAEL E. HASSELMO, AND ANGEL ALONSO1Properties and Role of Ih in the Pacing of Subthreshold Oscillations in Entorhinal Cortex Layer II Neurons. J. Neurophysiol. 83: 2562–2579, 2000.
I cannot get the model to oscillate using what I presume to be the HHtype equations the authors used (I had to adjust some of the rate equations for sign and magnitude, but the shape appears to be generally correct). The resting potential and input resistance are close to what the authors report in their paper. My main goal in trying to implement the model is to get a general model of deterministic oscillations due to interactions between the persistent sodium current (Nap) and the HCNgated current (Ih) that I can modify for my specific purposes rather than trying to implement this exact model. By deterministic I mean a model that does not rely upon on a noise source or a stochastic mechanism to oscillate.
1) Can someone refer me to a model with code available of Nap/Ih intrinsic oscillations (as opposed to those due to network mechanisms, stochastic ion channel gating, synaptic noise)?
2) Has anyone had any luck implementing the Dickson et al. 2000 model? Here's the relevant equations that I'm using:
For NaP (note activation is modeled as instantaneous in the Dickson paper):
mInf = 1.0/(1+exp((v vhalfa)/ka))
hInf = 1.0/(1+exp((v vhalfh)/kh))
hAlpha = (0.001*(2.88e3*v  4.9e2))/(1  exp((v + 4.9e2/2.88e3)/4.63))
hBeta = (0.001*(6.94e3*v + 0.447))/(1  exp((v + 0.447/6.94e3)/2.63))
hTau = 1/(hAlpha + hBeta)
For Ih:
Singlecompartment deterministic oscillating cell model?
Moderator: tom_morse
Re: Singlecompartment deterministic oscillating cell model?
Sorry, here's the Ih gating parameters, which I had to tweak a bit from the published equations to get them to resemble the fits in Fig. 11C:
mInfFast = 1.0/(1 + exp((v + 67.4)/12.66))
mInfSlow = 1.0/(1 + exp((v + 57.92)/9.26))
mAlphaFast = (0.3*(2.89e3*v 0.445))/(1  exp((v + (0.445/(2.89e3)))/24.02))
mBetaFast = (2.71e2*v  1.024)/(1  exp((v  (1.024/(2.71e2)))/17.4))
mTauFast = 0.8/(mAlphaFast + mBetaFast)
mAlphaSlow = (0.333*(3.18e3*v 0.695))/(1  exp((v + (0.695/(3.18e3)))/26.72))
mBetaSlow = (1.5*(2.16e2*v 1.065))/(1  exp((v  (1.065/(2.16e2)))/14.25))
mTauSlow = 0.8/(mAlphaSlow + mBetaSlow)
mInfFast = 1.0/(1 + exp((v + 67.4)/12.66))
mInfSlow = 1.0/(1 + exp((v + 57.92)/9.26))
mAlphaFast = (0.3*(2.89e3*v 0.445))/(1  exp((v + (0.445/(2.89e3)))/24.02))
mBetaFast = (2.71e2*v  1.024)/(1  exp((v  (1.024/(2.71e2)))/17.4))
mTauFast = 0.8/(mAlphaFast + mBetaFast)
mAlphaSlow = (0.333*(3.18e3*v 0.695))/(1  exp((v + (0.695/(3.18e3)))/26.72))
mBetaSlow = (1.5*(2.16e2*v 1.065))/(1  exp((v  (1.065/(2.16e2)))/14.25))
mTauSlow = 0.8/(mAlphaSlow + mBetaSlow)

 Site Admin
 Posts: 5919
 Joined: Wed May 18, 2005 4:50 pm
 Location: Yale University School of Medicine
 Contact:
Re: Singlecompartment deterministic oscillating cell model?
Good luck, but the success rate for attempts to reproduce a computational neuroscience model from the contents of a single article is very low. A PubMed search for
dickson hasselmo
turns up two other articles by these authors; you might check those for any mention of availability of source code, or perhaps a restatement of the model specification that would help you get this going. You might also check the authors' personal or lab web sites to see if they have made the source code available. Finally, you might contact them by email to ask if they can find and send you the source code. Fransen and Hasselmo might be the best ones to ask.
dickson hasselmo
turns up two other articles by these authors; you might check those for any mention of availability of source code, or perhaps a restatement of the model specification that would help you get this going. You might also check the authors' personal or lab web sites to see if they have made the source code available. Finally, you might contact them by email to ask if they can find and send you the source code. Fransen and Hasselmo might be the best ones to ask.