By this point of time, everyone agrees that there is a distinct pattern of response for each stimulus category with the object vision pathway in the human brain. Point of debate is the nature of distinctiveness.
- distributed patterns of activation throughout the temporal cortex, not just in areas that respond maximally, can predict category membership of visually-presented objects.
- contrasts with modular theories where specific brain regions are thought to be responsible for recognizing objects from a given category (e.g., the “face” area identified by Kanwisher & colleagues).
- Take away in short The entire response pattern rather than the most activated voxels has to be considered for understanding the potential mechanism of object representation with in human brain.
As per the paper,
Distinctiveness of the response to a given category was not due simply to the modular regions that responded maximally to that category. Based on the experimental fact from the paper that,
- The category being viewed also could be identified on the basis of the pattern of response when those regions were excluded from the analysis.
- Patterns of response that discriminated among all categories were found even within cortical regions that responded maximally to only one category.
Usage of a new correlation-based functional magnetic resonance imaging (fMRI) data analysis technique.
Paper Demonstrate that the response profile across category-selective regions of cortex (region with maximal response) contains some regions that are responsible to process some information not related a particular category stimulus.
Inversely when considering a specific category based stimulus, there lies region outside the maximal responding region that process some information about specific categories under consideration.
But paper fails to provide any specific details on which if the specific area is used for face perception. Now this can just be an argument. But even now we don’t have enough of technological or methodological advancement that would allow us to compare or map fMRI data onto baseline scale or map for comparison of neuroimaging data across subjects.
what might go wrong with current experimental procedure
Are the data being representative in nature for even and odd scan: stimuli from a given category were presented in blocks, and even scans were used to predict category membership from odd scans. Thus, any kind of temporal trend in the signal within a block could contribute.
The correlational data shown in Fig 4 shows that excluding the maximally-responding areas for stimuli that are thought to have their own “module” (e.g., faces) reduces correlations, whereas excluding maximal areas for non-modular stimuli does not.
Above fact suggests that the “modular” areas are making a differentially important contribution. So concluding that it is a little of both: specialized areas but also widely distributed representations.
 Haxby, J. V., Gobbini, M. I., Furey, M. L., Ishai, A., Schouten, J. L., & Pietrini, P. (2001). Distributed and overlapping representations of faces and objects in ventral temporal cortex. Science, 293(5539), 2425-2430.