Dear FSL experts,
we are seeking to reconsrtruct individual component maps from a group
melodic result (time concatenation mode), via the melodic_mix file that
contains the individual time courses.
What approach is recommended for this? The simpliest solution seems
regression of the filtered_func_data.* on the time course of the respective
component and subject of interest.
From a number of explorative attempts we though that maybe additional
nuisance regressors (motion?) could be included, but it is of course
debatable. The goal is to derive maps for 2nd level voxelwise comparison in
a GLM.
We'd appreciate any comments on how to best reconstruct individual components,
thanks a lot in advance,
Philipp
|