Purchase this article with an account.
Jacqueline Snow, Scott Squires, Kevin Stubbs, Jody Culham; fMRI reveals different activation patterns for real objects vs. photographs of objects. Journal of Vision 2016;16(12):512. https://doi.org/10.1167/16.12.512.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Hundreds of functional magnetic resonance imaging (fMRI) experiments have revealed the neural substrates of object processing using photos of objects. Here we used univariate and multivariate pattern analysis (MVPA) of fMRI responses to determine whether photos are represented similarly to real objects in the human brain. The stimuli were everyday objects of comparable size and elongation. The photos were matched closely to the real objects for size and viewpoint. The stimuli were presented in rapid succession in the fMRI scanner using a custom-designed conveyor belt. We used a block design in which subjects viewed four exemplars (of different color, form, etc.) of one object type (e.g., whisks) in either real or photo format in each block. Univariate subtraction analysis revealed higher activation for real objects than photos in object-selective areas of the ventral visual stream (including the middle temporal gyrus, lateral occipital cortex, and fusiform gyrus) and dorsally in primary somatosensory cortex. Surprisingly, whole-brain searchlight representational similarity analysis showed that the lateral occipitotemporal cortex (LOTC) and anterior intraparietal sulcus (aIPS) were sensitive to the format in which stimuli were viewed, with stronger correlations for stimuli displayed in the same (i.e., Real-Real or Photo-Photo) versus different (i.e., Real-Photo) viewing dimensions. Finally, multidimensional scaling within independently-defined volumes of interest in left LOTC and aIPS revealed a global grouping that reflected a categorical distinction between real object and image displays, with more distinct representations for the real object exemplars than the images. Taken together, our results indicate that the human brain does not treat photographs as being equivalent to real objects. Real objects might elicit different brain-based responses to photos because they provide richer visual information, they have definite haptic qualities, and they afford genuine action.
Meeting abstract presented at VSS 2016
This PDF is available to Subscribers Only