For the orientation adaptation to stimulus of angle 
,
substituting equation (10) into equation (17), it is
not difficult to derive that the network response to stimulus of
angle 0 (vertical) is changed to
in which 
is the feedforward tuning width chosen to be 
and 
is the parameter of the strength of decorrelation feedback.
The theoretical curve of perceived orientation 
is
derived by assuming the maximum likelihood of the the neural population,
i.e., the perceived angle 
is the angle at which 
is
maximized. It is shown in figure 5 (right). The solid line is
the theoretical curve and the experimental data come
from [3] (they did not give the errors, the error bars are
of our estimation 
). The parameter obtained through
fit is the strength of decorrelation feedback: 
.
Figure 5:
Quantitative comparison of the theoretical predictions with the
experimental data of orientation contrast (left) and orientation
adaptation (right).
It is very interesting that we can derive a relationship which is
independent of the parameter of the strength of decorrelation feedback
,
in which 
is the adaptation angle at which the tilt
aftereffect is most significant and 
is the perceived angle.
