Reverse Inference (RI) is an imaging-based type of inference from brain states to mental states, which has become highly widespread in neuroscience, most especially in neuroeconomics. Recent critical studies of RI may be taken to show that, if cautiously used, RI can help achieve research goals that may be difficult to achieve by way of behavior-based procedures alone. But can RI exceed the limits of these procedures and achieve research goals that are impossible for them to achieve alone? By way of answering this question we show that a conception of the mind-type identity-under which the answer is in the positive, is untenable for reasons that strongly support another conception of the mind-functionalism-under which the answer must be in the negative. On this basis we then conclude that RI cannot exceed the limits of behavior-based procedures in cognitive psychology.