move from numpy array of confounds to pandas df

SIMEXP / load_confounds

Load fMRIprep confounds in python

MIT License

36 stars 12 forks source link

move from numpy array of confounds to pandas df #147

Closed pbellec closed 3 years ago

pbellec commented 3 years ago

my understanding is that nilearn maskers are going to support pandas df. It is a much cleaner format to output confounds, and would be preferable over the numpy arrays. This requires minimal changes, as internally load_confounds manipulate the confound array using pandas already.

htwangtw commented 3 years ago

This will also simplify a lot of the code from what I am aware of. Shall we aim to merge #142 first?

pbellec commented 3 years ago

yes, I am going to look into #142 later today, then we can deal with the df.