Biomedical Engineering Reference
In-Depth Information
and, by induction, the probability of a block ω m ,
n,n
m
D is given by:
n
P ω ( l ) ω l− 1
l−D P ω m + D− 1
.
P [ ω m ]=
(8.2)
m
l = m + D
Thus, knowing the probability of occurrence of the block ω m + D− m one can infer
the probability of forthcoming blocks by the mere multiplication of transition
probabilities.
Given the (joint) probability P [ ω m ] the (marginal) probability of sub-blocks
can be easily obtained, since for m
n 1
n 2
n ,
( n 1 ,n 2 )
P ω n 2
n 1 =
P [ ω m ] ,
(8.3)
m,n
where ( n 1 ,n 2 )
means that we sum up over all possible spiking patterns in the interval
m,n
{
(i.e., we sum up over all possible values of
ω ( m ) ,...,ω ( n 1 1) ( n 2 +1) ,...,ω ( n )).
As a consequence, from ( 8.2 )to( 8.3 ), the probability of the spike block ω n−D ,
of range D ,is:
m, n
}
excluding the interval
{
n 1 ,n 2 }
( n
D,n )
n
P ω ( l ) ω l− 1
P ω n−D =
l−D P ω m + D− 1
.
(8.4)
m
m,n
l = m + D
Knowing the probability of an initial block of range D (here ω m + D− m ) one infers
from this equation the probability of subsequent blocks of range D . Equation ( 8.4 )
can also be expressed in terms of vector-matrices multiplication, and the main
properties of the Markov chain can be deduced from linear algebra and matrices
spectra theorems [ 64 ]. For compactness we shall not use this possibility here though,
(see [ 74 ] for further details).
However, this equation shows that the “future” of the Markov chain (the proba-
bility of occurrence of blocks) depends on an initial condition (here P ω m + D− m ),
which is a priori undetermined. Moreover, there are a priori infinitely many possible
choices for the initial probability.
8.3.1.4
Asymptotic of the Markov Chain
Assume now that n
.
Practically, this limit corresponds to considering that the system began to exist in
a distant past (defined by the initial condition of the Markov chain) and that it has
evolved long enough, i.e., over a time larger than relaxation times in the system, so
m
+
in Eq. ( 8.4 ) and more precisely that m
→−∞
Search WWH ::




Custom Search