Information Technology Reference
In-Depth Information
Fig. 2.
General overview of the LGN mode. The figure shows components and relations
involved in our model, stimuli, receptive field, Poisson generators, IAF neurons.
behaviors to those observed in relay cells and interneurons. The membrane po-
tential is computed according to the following equation:
dV
dt
w
+
I
A
E
s
)+
g
L
Δ
T
e
V
−
V
T
C
m
=
g
L
(
V
E
L
)+
g
s
(
t
)(
V
(5)
−
−
−
−
Δ
T
where
C
m
is the capacitance,
g
L
is the leak conductance,
E
L
is the resting
potential, and the exponential term creates a soft spiking threshold around
V
T
with softness determined by
Δ
T
,
g
s
,
I
e
.
The synaptic conductances are shaped by the alpha function (Rall 1967, Jack,
Noble and Tsien, 1975) with time constant
τ
s
:
g
s
(
t
)=
g
s
(
t
τ
s
)
e
−
t
(6)
τ
s
w is an adaptation current with time constant
τ
w
and sub-threshold adaptation
level set by a:
dw
dt
τ
w
=
a
(
V
E
L
)
w
−
−
(7)
A spike event is triggered when the membrane potential diverges due to the
exponential term. In practice, a spike is triggered when
V
reaches a suciently
large value such as
V
T
. When a spike occurs, the membrane potential is reset to
Vr and a spike adaptation
b
is added to the adaptation current w:
V
=
V
r
w
=
w
+
b
V
V
T
≥
(8)
4R su s
4.1
Impact of Lateral Inhibition in LGN Processing
In the following experiment we present a horizontal bar stimulus with and with-
out the lateral inhibition coming from the OFF channel (Figure 3). The following
results were obtained in our simulation for both situations.
Search WWH ::
Custom Search