hi Asad,
The code does not recompute the matrix if it exists (and isn't smart enough to know you updated the paramers), you can set the matrix to None,
default_cortex.local_connectivity.matrix = None
and then call the compute_local_connectivity method. You should be aware that the toarray method on a sparse matrix returns a dense version, and the default local connectivity is relatively sparse. If you change the kernel, you may run into out of memory
errors. You should be able to save the sparse matrix directly to MATLAB format, just drop the toarray() call.
cheers,
Marmaduke
hi Asad
In fact, the coupling strength is a scaling factor applied to the local connectivity matrix during simulation, and it doesn't affect the local_connectivity.matrix, which is just hte kernel evaluated over geodesic distance.
In other words, you can evaluate local connectivity once and simply multiply its values by the scaling factor, and then continue whatever it is you are doing.
hope that helps,
Marmaduke
To view this discussion on the web visit https://groups.google.com/d/msgid/tvb-users/c55168bf-f9f3-41ea-8955-1829d10958c8%40googlegroups.com.