A decade of spectacular progress in functional magnetic resonance imaging (fMRI) technology and systems neuroscience research has so far yielded few changes in our daily lives. The dearth of clinical applications of this prolific and academically promising research tool began raising the eyebrows of the public and the research funding agencies. This may be one of the reasons for the enthusiasm and interest paid to the growing body of literature suggesting that blood oxygenation level-dependent (BOLD) fMRI of the brain could be sensitive to the differences between lie and truth. The word ‘differences’ is critical here since it refers to the often-ignored core concept of BOLD fMRI: it is only sensitive to differences between two brain states. Thus, available studies report using fMRI to discriminate between lie and truth or some other comparative state rather than to positively identify deception. This nuance is an example of the extent to which applied neuroscience research does not lend itself to the type of over-simplification that has plagued the interpretation of fMRI-based lie detection by the popular press and the increasingly vocal academic critics. As an early contributor to the modest stream of data on fMRI-based lie detection, I was asked by Dr Aldert Vrij to write a piece in favour of fMRI-based lie detection, to be contrasted with a piece by Dr Sean Spence presenting an opposite point of view (Spence, 2008). This seemingly straightforward task presented two hurdles: having to respond to the popular as well as scientific view of what lie detection with fMRI is and present a wholly positive view of evolving experimental data.