Reverse inference is not a fallacy per se: Cognitive processes can be inferred from functional imaging data

When inferring the presence of a specific cognitive process from observed brain activation a kind of reasoning is applied that is called reverse inference. Poldrack (2006) rightly criticized the careless use of reverse inference. As a consequence, reverse inference is assumed as intrinsically weak by many and its validity has been increasingly regarded as limited. Although it is undisputed that the careless use of reverse inference is a problematic practice, the current view of reverse inference is to the author's opinion overly pessimistic. The present manuscript provides a revised formulation of reverse inference that includes an additional conditional constraint that has been previously acknowledged, but so far not implemented: the task-setting. This revised formulation I.) reveals that reverse inference can have high predictive power (as demonstrated by an example estimation) and II.) allows an estimation of reverse inference on the basis of meta-analyses instead of large-scale databases. It is concluded that reverse inference cannot be disregarded as a fallacy per se. Rather, the predictive power of reverse inference can even be "decisive"-dependent on the cognitive process of interest, the specific brain region activated, and the task-setting used.