Two dimensional neuron models
Reduction to two dimensions
General approach
- Origin model:
ΣIk=gNam3h(u−ENa)+gKn4(u−Ek)+gL(u−EL)
It has 4arguments
m
- In this chapter,we reduce it to 2 arguments:
u
,
ω
Cdudt=−gNa[m0(u)]3(b−ω)(u−ENa)−gk(ωa)4(u−Ek)−gL(u−EL)+I
ω=b−h=an
- What changes?
because ofm
changes fast:
m(u,t) -> m0(u)
because ofh
andn
seems to have linear relationship:
h ->b−ω
n ->ωa
Morris–Lecar model
It seems that this model just change exponents of arguments to get a linear equation.
This section also give a approximate equation for
m0(u)
and
ω0(u)
FitzHugh–Nagumo model
Phase plane analysis
On phase plane, point
(u(t),ω(t))
Δu=u˙Δt
Δω=ω˙Δt
Use vector
(u˙,ω˙)
plot a vector field on phase plane:
For each point on the plane, we draw an arrow, the lenght of the arrow is proportional to the length of the vector, and the direction is same as the vection
Nullclines
-
u
-nullcline:
-points withu˙=0 .
-The direction of flow on the u-nullcline is in direction of - (0,ω˙)
-vertical -
ω
-nullcline:
-points with ω˙=0
- (0,u˙)
-horizontal - fix-point:
-intersection of u -nullcline andω -nullcline
-on nullclines the direction of arrows change at fix points
Stability of fixed points
At fix point
ddtx⃗ =(FuGuFωGω)x⃗Set x(t)=e⃗ eλt ,we get eigenfunction:
λx⃗ =(FuGuFωGω)x⃗
λ1+λ2=Fu+Gω
λ1λ2=FuGω−FωGuSaddle point:
λ1>0
λ2<0Stable points:
λ1<0
λ2<0Unstable points:
λ1>0
λ2>0