Purchase this article with an account.
Jiedong Zhang, Yaoda Xu; The representation of object parts in the human brain. Journal of Vision 2013;13(9):130. doi: 10.1167/13.9.130.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The human brain is an efficient information processing system that can distinguish and recognize visual objects consisting of a great variety of parts. Although parts contribute significantly to object representation and recognition, our understanding is still incomplete regarding what kinds of parts are extracted and represented in the human brain during visual object processing. Behavioral research has shown that human vision automatically segments objects into parts using contour concavities as boundaries between parts. Using fMRI, here we tested whether such parts are indeed represented in human visual cortex during object processing. We used the image of a mug as our stimulus and generated two kinds of parts. The "natural parts" were generated at shape contour concavity and consisted of the handle of a mug as one part and the body of the mug as the other part; and the "unnatural parts" were generated by placing a cut at the mug body and consisted of the handle plus half of the mug body as one part and the rest of the mug body as the other part. We presented these parts either alone, or together in an intact or a scrambled mug configuration. If natural parts are automatically extracted and represented during visual object processing in the brain, then the neural response pattern of the intact object would be better predicted by combining those from the natural parts than from the unnatural parts. By examining fMRI response patterns and using multi voxel pattern analysis, we indeed obtained this result in lateral occipital complex, a brain region previously shown to process object shapes. This result demonstrates that natural parts are computed and represented spontaneously in the human brain during visual object processing.
Meeting abstract presented at VSS 2013
This PDF is available to Subscribers Only