Full decomposition of time-series fit

[kendrickkay]

Some users are interested in getting all components of the final GLM fit, e.g., the component due to the single trial beta amplitudes, the polynomial fit, the correlated noise component found, and the residuals.

Currently, this is not supported by the API outputs and is difficult to return without substantive architectural rewrites. On the other hand, with some legwork it may be possible to post-hoc derive these measures from the outputs...

An example script that is sort of along these lines is here: https://github.com/cvnlab/GLMsingle/blob/main/matlab/examples/example7_makepredictedtimeseries.m
But it doesn't quite do everything one wants...