Abstract
The neural mechanisms that allow a percept of a single surface to be formed from noncontiguous regions of an image remain largely unknown. We used fMRI to measure the responses of visual areas to ambiguous patterns that could be perceived as either a coherent surface or as unconnected fragments. Subjects viewed a dynamic array of small adjoining elements that were randomly assigned as either surface or noise every 100 ms. On each trial the color of the surface elements was identical and the color of each noise element was randomly sampled. Prior to scanning, each subject's threshold was defined as the proportion of elements assigned to surface when the subject perceived a surface on half of the trials. Each subject in the fMRI experiment viewed stimuli at their individually measured threshold as well as values above and below this threshold. Neural responses to each condition were estimated as the linear kernels that best fit the event-related timecourse from each of many visual areas. Early visual areas (V1,V2) responded most strongly during trials with a high proportion of noise elements while later visual areas responded most strongly during trials with a low proportion. Furthermore, in later visual areas responses to threshold stimuli were 33% higher during trials in which the subject perceived a surface than during trials in which the subject did not. Early visual areas did not show this pattern. Later visual areas likely contain neurons whose activity correlates with coherent surface perception on a trial-by-trial basis.
Supported by NIH EY11862