Identifying natural images from human brain activity

Top Cited Papers
Publisher Website
Google Scholar
Abstract
Recent functional magnetic resonance imaging (fMRI) studies have shown that, based on patterns of activity evoked by different categories of visual images, it is possible to deduce simple features in the visual scene, or to which category it belongs. Kay et al. take this approach a tantalizing step further. Their newly developed decoding method, based on quantitative receptive field models that characterize the relationship between visual stimuli and fMRI activity in early visual areas, can identify with high accuracy which specific natural image an observer saw, even for an image chosen at random from 1,000 distinct images. This prompts the thought that it may soon be possible to decode subjective perceptual experiences such as visual imagery and dreams, an idea previously restricted to the realm of science fiction. Recent functional magnetic resonance imaging (fMRI) studies have shown that it is possible to deduce simple features in the visual scene or to which category it belongs. A decoding method based on quantitative receptive field models that characterize the relationship between visual stimuli and fMRI activity in early visual areas has now been developed. These models make it possible to identify, out of a large set of completely novel complex images, which specific image was seen by an observer. A challenging goal in neuroscience is to be able to read out, or decode, mental content from brain activity. Recent functional magnetic resonance imaging (fMRI) studies have decoded orientation1,2, position3 and object category4,5 from activity in visual cortex. However, these studies typically used relatively simple stimuli (for example, gratings) or images drawn from fixed categories (for example, faces, houses), and decoding was based on previous measurements of brain activity evoked by those same stimuli or categories. To overcome these limitations, here we develop a decoding method based on quantitative receptive-field models that characterize the relationship between visual stimuli and fMRI activity in early visual areas. These models describe the tuning of individual voxels for space, orientation and spatial frequency, and are estimated directly from responses evoked by natural images. We show that these receptive-field models make it possible to identify, from a large set of completely novel natural images, which specific image was seen by an observer. Identification is not a mere consequence of the retinotopic organization of visual areas; simpler receptive-field models that describe only spatial tuning yield much poorer identification performance. Our results suggest that it may soon be possible to reconstruct a picture of a person’s visual experience from measurements of brain activity alone.