Using fmriprep output for conn surface analysis

I would like to smooth and denoise resting state fmriprep output in conn. As a first step, I am trying to get conn to run surface based smoothing on the data. When I enter the T1 from my first subject’s freesurfer mri folder, I get the message “Freesurfer aseg.mgz files found. Do you want to import Grey/White/CSF masks from these files?”, suggesting freesurfer segmentations are being imported successfully. However, whether I try to use *space-T1w_desc-preproc_bold.nii.gz or *space-fsLR_den-91k_bold.dtseries.nii as the functional, when I try to run surface-based smoothing, I get the error:

Error using conn_surf_smooth (line 28)
Incorrect dimensions in file

I have also tried importing the fmriprep data, but this seems to only load volumetric data, and, as far as I can tell, doesn’t load freesurfer surfaces.

I found this thread: NITRC: CONN : functional connectivity toolbox: how to employ func.gii to perform surface-based FC analysis?

which may answer my question. Will report back.