Confound removal and normalization in practice: a neuroimaging based sex prediction case study