Authors:

Christoff K. & OWEN, A.M.

Reference:

Year of publication:

2006

CBU number:

6399

Abstract:

In a recent TICS article, Poldrack [1] offers a highly informative analysis of the use and misuse of ‘reverse inference’ in neuroimaging, a common practice by which the engagement of a particular cognitive process is inferred from the activation of a particular brain region. Using a formal Bayesian analysis framework, Poldrack shows that the usefulness of reverse inference depends on the selectivity of activation in the region of interest (the ratio of process-specific activation to the overall likelihood of activation in that region across all tasks). Here we argue that the usefulness of reverse inference also depends on whether the relevant task characteristics for the region of interest are taken into account.