Using Decoding toolbox with preprocessed images created via fmriprep

Dear all,
I am learning the decoding toolbox to apply searchlight on the data I have preprocessed with fmriprep. However, in the original article of the toolbox, it mentions not to have spatial normalization. Considering I have used fmriprep command with a default script (rather than excluding or including extra pipeline steps), is there a way that I can use these images for the decoding toolbox, or must I preprocess my data all over again without normalization?

Kind regards,