... values¹

Note one should test the image for local maximum to confirm inferentially the spatial description of activation

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... entropies²

The difference is always positive i.e. the entropy of a normal distribution $h(\phi_y)=-E(\ln(\phi_y))=1/2(p +p\ln(2\pi)+\ln(det(V)))$ has the largest entropy among p-dimensional distributions. The proof of the equivalent definitions comes from $E_{\phi_y}(\ln(\phi_y))=E_{f_y}(\ln(\phi_y))$ because $\ln(\phi_y)$ is a polynomial of degree two and $\phi_y$ and $f_y$ have same first and second moments [5]

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.