The documentation of canica does not show a way to fetch the mixing matrix. Does anyone have an idea how to obtain it?
Simply type canica.transform(func_filenames)
, assuming the notations of example
https://nilearn.github.io/auto_examples/03_connectivity/plot_compare_decomposition.html#sphx-glr-auto-examples-03-connectivity-plot-compare-decomposition-py
transform() gives you the time series associated with the ICs, which is presumably what you want, when you ask for the mixing matrix ?
HTH,
Bertrand