Philosophy, Psychology and Neuroscience
"The expressive face-unculus: Localizing and interpreting the dynamics of facial feature encoding from MEG data"
Bubbles is best described as a reverse correlation method that reveals the sensitivity of brain signals to specific visual information. Arguably the method might be seen as generating seductive images of the content of mental processing (given brain data, and an extremely constrained set of sensory inputs). I will explain the method, illustrated with bubbles images constructed from meg source data, while subjects classify emotional expressions. This (preliminary) data shows a network interaction of multiple nodes processing different information in parallel. The 'Face Processing Area' appears in this network; our data suggest that it performs some kind of integration of features that are represented in detail elsewhere in the network. I will argue for a model of visual processing where the brains inferential processes make use of reverberating activity and re-entrant connections to accumulate evidence over the course of decision making, and discuss why this model (and certain other features of the data) make it difficult to interpret the relationship of bubbles images to mental, rather than brain processes.