Effects of category learning on visual representations

Vanderbilt University, Nashville TN

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): How do visual representations of objects change with experience? While it is now widely acknowledged that visual representations change with experience, questions abound concerning the time-course, flexibility, and degree to which experience can change object representations. Objects can be represented psychologically as points in the space of dimensions on which they vary (e.g., color, shape, manner of motion). The central hypothesis of this proposal is that some dimensions of objects can be enhanced by experience categorizing those objects. We will consider under the same conceptual umbrella a number of phenomena and mechanisms that represent cases of such dimensional modulation, so that they can be compared and integrated in a unified framework. The first aim is to seek evidence for dimensional modulation in visual cortex with category learning. There is some controversy over whether categorization experience influences visual cortical representations. Resolving this controversy is essential to fully understand experience- dependent neural plasticity in visual cortex. The second aim is to study the neural basis of both flexible and stable dimensional modulation. Flexible dimensional modulation is often characterized as a top-down influence on visual representations through executive control, while stable dimensional modulation is instead characterized as a bottom-up change in the underlying perceptual representations. This work aims to clarify how top-down and bottom-up influences shape experience-dependent neural plasticity in visual cortex. We vary task demands, the amount of learning, types of stimulus dimensions, and category structures to explore the multiple facets of dimensional modulation caused by category learning, both behaviorally and using fMRI. Our fMRI work uses a combination of two complementary techniques for measuring differences within overlapping neural networks, release from repetition suppression and multivoxel pattern analysis. Our work will provide an opportunity to use converging evidence from the two techniques, comparing their strengths and weaknesses within the same study. Ultimately, an understanding of the plasticity of the neural code in visual cortex has important implications for understand deficits in fine perceptual skills seen in dyslexia, autism, and visual agnosia, and for creating future neural prosthetics that interface directly the human brain by interpreting patterns of neural firing.