fmripop - postprocessing of fmriprep data compatible with nilearn 0.6.2

The main changes included in this release are:

1. All boolean input argument have a consistent behavior: by default they are set to False. To enable the functionality controlled by a given flag, just pass it as an argument when you inovoke the script.
   --detrend
2. The documentation of each input argument has been improved.
3. Updated to work with nilearn 0.6.2 -- this version of nilearn fixed the bug that prevented having heterogeneous smoothing (ie, having a different FWHM value along each axis). With this change, it means this script does not longer work with versions of nilearn < 0.6.2
4. Visual debugging functionality has been added. The functionality plots time-series of the default mode network taken from the input data, and the output data.
5. The argument previously called --add_mean_img_back is now called --add_orig_mean_img. This new name better reflects what the function actually does. The previous name was confusing. Many processing functions such as filtering, detrending and confound removal have the effect of removing the nonzero mean value of each voxel, where the mean is calculated across time frames.
6. Updated readme with the typical use cases for resting-state and task frmi data. These cases are also at the top of the script.
7. Includes a .yml file to create a working environment with conda.