Moving Forward With fMRI Data

The use of fMRI to inform cognitive theory depends upon the assumption that there is an isomorphic relationship between functional states and brain states. Even if this assumption is accepted, employing fMRI to make cognitive inferences is not straightforward. We discuss these inferential difficulties and describe an example in which fMRI data has had a significant impact on a cognitive theory despite them. The embodied cognition framework states that accessing a conceptual feature engages the same processes that are active when the feature is directly experienced. fMRI studies have consistently found that accessing a feature activates cortical regions slightly anterior to the regions activated when the feature is experienced. We suggest that this neural dissociation between accessing and perceiving a feature is inconsistent with the strong form of embodied cognition theory and that this inconsistency would likely not have been identified without fMRI. We describe how the observation of this forward shift has led to new ways of thinking about perceptual and conceptual representations and the relation between them. Finally, we argue that despite the strong assumptions that are needed to make functional inferences from fMRI, this does not detract from its value as a source of convergent evidence.