ModelDB is moving. Check out our new site at https://modeldb.science. The corresponding page is https://modeldb.science/249705.

Glutamate mediated dendritic and somatic plateau potentials in cortical L5 pyr cells (Gao et al '20)

 Download zip file 
Help downloading and running models
Accession:249705
Our model was built on a reconstructed Layer 5 pyramidal neuron of the rat medial prefrontal cortex, and constrained by 4 sets of experimental data: (i) voltage waveforms obtained at the site of the glutamatergic input in distal basal dendrite, including initial sodium spikelet, fast rise, plateau phase and abrupt collapse of the plateau; (ii) a family of voltage traces describing dendritic membrane responses to gradually increasing intensity of glutamatergic stimulation; (iii) voltage waveforms of backpropagating action potentials in basal dendrites (Antic, 2003); and (iv) the change of backpropagating action potential amplitude in response to drugs that block Na+ or K+ channels (Acker and Antic, 2009). Both, synaptic AMPA/NMDA and extrasynaptic NMDA inputs were placed on basal dendrites to model the induction of local regenerative potentials termed "glutamate-mediated dendritic plateau potentials". The active properties of the cell were tuned to match the voltage waveform, amplitude and duration of experimentally observed plateau potentials. The effects of input location, receptor conductance, channel properties and membrane time constant during plateau were explored. The new model predicted that during dendritic plateau potential the somatic membrane time constant is reduced. This and other model predictions were then tested in real neurons. Overall, the results support our theoretical framework that dendritic plateau potentials bring neuronal cell body into a depolarized state ("UP state"), which lasts 200 - 500 ms, or more. Plateau potentials profoundly change neuronal state -- a plateau potential triggered in one basal dendrite depolarizes the soma and shortens membrane time constant, making the cell more susceptible to action potential firing triggered by other afferent inputs. Plateau potentials may allow cortical pyramidal neurons to tune into ongoing network activity and potentially enable synchronized firing, to form active neural ensembles.
Reference:
1 . Gao PP, Graham JW, Zhou WL, Jang J, Angulo SL, Dura-Bernal S, Hines ML, Lytton W, Antic SD (2020) Local Glutamate-Mediated Dendritic Plateau Potentials Change the State of the Cortical Pyramidal Neuron. J Neurophysiol [PubMed]
Model Information (Click on a link to find other models with that property)
Model Type: Dendrite; Neuron or other electrically excitable cell;
Brain Region(s)/Organism: Prefrontal cortex (PFC); Neocortex;
Cell Type(s): Neocortex L5/6 pyramidal GLU cell;
Channel(s): I A; I K; I h; I K,Ca;
Gap Junctions:
Receptor(s): Glutamate; NMDA;
Gene(s):
Transmitter(s): Glutamate;
Simulation Environment: NEURON; Python;
Model Concept(s): Action Potentials; Active Dendrites; Calcium dynamics; Axonal Action Potentials; Dendritic Bistability; Detailed Neuronal Models; Membrane Properties; Synaptic Integration;
Implementer(s): Antic, Srdjan [antic at neuron.uchc.edu]; Gao, Peng [peng at uchc.edu];
Search NeuronDB for information about:  Neocortex L5/6 pyramidal GLU cell; NMDA; Glutamate; I A; I K; I h; I K,Ca; Glutamate;
/
plateau-potentials
mod
x86_64
ampa.mod *
ca.mod *
Ca_HVA.mod *
Ca_LVAst.mod *
Cad.mod *
CaDynamics_E2.mod *
CaT.mod *
epsp.mod *
gabaa.mod *
gabab.mod *
glutamate.mod *
h_kole.mod *
h_migliore.mod *
Ih.mod *
IL.mod *
Im.mod *
K_Pst.mod *
K_Tst.mod *
kadist.mod *
kaprox.mod *
kBK.mod *
kv.mod *
na.mod
Nap_Et2.mod *
NaTa_t.mod *
NaTs2_t.mod *
NMDA.mod *
NMDAeee.mod
NMDAmajor.mod
PlateauConductance.mod *
SK_E2.mod *
SKv3_1.mod *
vecstim.mod *
vmax.mod *
ghk.inc *
                            
:nmda.mod G. Major 20/2/95 23.i.99 synaptic conductance with nmda v-dependence

COMMENT

g_max has voltage dependence of nmda receptor
g_max not same as maximum gNMDA
looks like model from Brodin et al. J. Neurophysiol.
66(2) 1991: 473-84
with some parameters (alpha and beta) from Ascher and Novak 1988
J. Physiol. 399 247-266
need to change Mg to 1
kinetics of 3-exponential envelope based on
D'Angelo Rossi and Taglietti
Eur. J. Neurosci 6:640-645 (1994)
p_vspread=12.484 mV  p_vhalf=-19.906 mv Brodin et al 91, Ascher Novak 88 Mg 1.8
ENDCOMMENT

UNITS {
    (molar) = (/liter)
    (mM) = (millimolar)
    (nA) = (nanoamp)
    (mV) = (millivolt)
    (umho) = (micromho)
    (uS) = (micromho)
}

NEURON {
    SUFFIX nmda
    POINT_PROCESS nmda
    NONSPECIFIC_CURRENT i
    RANGE onset, gmax, e, i, g, genv, q,
          Mg, taup, pinf,
          alf, alfA, alfslope,
          bet, betA, betslope,
          tau_on, tau_on0, tau_onslope,
          tau_off1, tau_off1_0, tau_off1slope,
          tau_off2, tau_off2_0, tau_off2slope,
          f_fast, f0, fslope,
          q10, Mg_time_factor
}

PARAMETER {
    celsius= 32 : 22 (degC)
    q10 = 3 ()
    Mg_time_factor = 1 ()
    dt (ms)
    onset=0 (ms)
    e=0	(mV)
    v	(mV)
    Mg= 1.8
    gmax = 0 (umho)  : uS
    alfslope = 47 (mV)
    alfA = 5.4 (/ms)
    betslope = 17 (mV)
    betA = 0.61(/mM-ms)
    tau_on0 = 2.915 (ms)
    tau_onslope = -0.004125 (ms/mV)
    tau_off1_0 = 61.5 (ms)
    tau_off1slope = 0.5625 (ms/mV)
    tau_off2_0 = 352.5 (ms)
    tau_off2slope = 5.7375 (ms/mV)
    f0 = 0.515  : should be unitless? was (mV) until 5/Mar/08
    fslope = -0.003125 (/mV)
}

ASSIGNED {
i (nA)  g (uS) genv (uS) pinf alf bet
taup (ms) tau_on (ms) tau_off1 (ms) tau_off2 (ms) f_fast
}

STATE {
 p q (nanocoulombs)
}

UNITSOFF

INITIAL {
   nmda_rates(v)
   nmda_taus(v,1.1*onset)
   p = pinf
   q = 0  : charge
}

BREAKPOINT {
    LOCAL trel
    SOLVE nmda_states METHOD cnexp
    if (t>onset) {
        trel= t - onset
    	genv = gmax*( -            exp(-trel/tau_on)
                   +     f_fast*exp(-trel/tau_off1)
                   + (1-f_fast)*exp(-trel/tau_off2)
                 ) : time dependent "maximum possible" or envelope conductance
        g=genv*p  : voltage dependency
    } else {
        genv=0
        g=0
    }
    i = g*(v - e)
}

DERIVATIVE nmda_states {
      nmda_rates(v) : compute p at this v and dt
      nmda_taus(v,t)
      p' = (pinf - p)/taup
      q' = i*(1e-3)  : dt  is in ms: want nC/ms not nC/s (nA)? i is in nA=10^-9 C/s, q' is in nC/ms = 10^-9/10^-3 C/s =10^-6 C/s
}

PROCEDURE nmda_taus(v,t) { : call once from HOC to initialise for each nmda syn
        LOCAL temp_factor
        temp_factor = q10^((celsius - 28.50)/10)
        if (t>onset) {
           f_fast= f0 + fslope*v
           if (f_fast>1) {
             f_fast=1
           }
           if (f_fast<0) {
             f_fast = 0
           }
           tau_on  =(tau_on0 + tau_onslope*v)/temp_factor
           tau_off1=(tau_off1_0 + tau_off1slope*v)/temp_factor
           tau_off2=(tau_off2_0 + tau_off2slope*v)/temp_factor
           if (tau_off1<tau_off1_0/10) {
              tau_off1=tau_off1_0/10
           }
           if (tau_off2<tau_off1) {
              tau_off2=tau_off1
           }
        }
}

PROCEDURE nmda_rates(v) { : call once from HOC to initialise
        LOCAL  temperature_factor :
        TABLE pinf, taup, alf, bet
          DEPEND q10, celsius,
                 alfA, betA, Mg, alfslope, betslope
          FROM -100 TO 100 WITH 200
        temperature_factor = q10^((celsius - 20)/10)
        alf = temperature_factor*alfA*exp(v/alfslope)
        bet = temperature_factor*betA*Mg*exp(-v/betslope)
        : Mg_time_factor = : is 1 in standard formula
	taup = Mg_time_factor/(alf + bet)
        :printf("\n NMDA taup %g  ms  (from time_factor %g) ", taup, Mg_time_factor)
        pinf = alf/(alf + bet)
}

UNITSON

Loading data, please wait...