Neural Coding of Image Blur Assessed by fMRI
Loading...
Authors
Mussell, Katherine E.
Issue Date
2015
Type
Thesis
Language
Keywords
blur , fmri , norm-based coding , spatial frequency
Alternative Title
Abstract
Blur is a fundamental perceptual attribute of images, but the way in which the visual system encodes this attribute remains poorly understood. We examined the neural correlates of image blur by measuring the fMRI BOLD response to images that varied from focused to either too blurred or too sharpened. Observers viewed grayscale images of natural scenes, filtered by varying the slope of the log amplitude spectra from -1 (strongly blurred) to +1 (strongly sharpened), with RMS contrast equated to the original after filtering. In primary visual cortex (V1) there was higher activation for the in-focus images than for the sharpened or blurred images. Peak responses were similar for focused and sharpened images in foveal V1, while both blurred and sharpened images resulted in lower activity in more peripheral retinotopic regions. Similar patterns were observed in extra-striate areas (V2 to V3), though the differential response to focused images was greatest in V1. These results suggest that focused images provide the strongest neural activity in V1, and run counter to expectations from norm-based or predictive coding in which focus is encoded implicitly as an absence or expected attribute.
Description
Citation
Publisher
License
In Copyright(All Rights Reserved)