Since its inception over 30 years ago, fMRI brain imaging has had remarkable influence on research and society. However, since fMRI images require complex statistical analyses, interpretational and statistical errors are common, leading to stagnation and consternation into the field. In this seminar it is demonstrated how dropping common assumptions can allow task-based fMRI to provide extremely reliable temporal and spatial information that can lead to detailed cognitive theories and generation of hypotheses for future studies, with strong potential for clinical applications.