Cannot understand 'val_matrix' values

[DarioSimonato]

Hi everyone,
I'm not able to completely understand how the the result dictionary of PCMCI and the plotting function work. In particular, by having a look at the val_matrix and the graph parameters, I cannot understand why in val_matrix there are non-zero coefficients in position i,j even if variable i is not parent of variable j or viceversa. I would have expected that such matrix had non-zero values for elements in each position only if the element in the same position in graph was non-zero, therefore encoding the coefficient of a structural causal model.

In the attached screenshot we have a network with 2 links, but the val_matrix is almost non-zero everywhere.

Thanks