Question about fmriprep reorientation of images

Hello experts,

I am fairly new to fmriprep and have a question about if/how functional images are reoriented during preprocessing. I typically have used SPM for preprocessing and later analyses of fMRI data. When I load an unpreprocessed functional nifti file into SPM12 using its display function, the voxel size dimensions read -3x3x4. My understanding is that the -3 occurs because SPM12 reads the orientation information from the header and applies a reflection between the coordinate system of the voxels and the mm coordinate system. This allow me to assume what is right and left as reflected in the x-axis coordinates in the crosshairs.

When I load a functional nifti preprocessed using fmriprep into SPM12 using its display function, the voxel size dimensions read 2x2x2. 2-mm isotropic makes perfect sense because the data were resampled. I am curious why it doesn’t read -2x2x2. Is the switch from negative to positive because fmriprep reorients images to RAS+? I have no errors when running fmriprep on these data.

Thanks so much!