Item

Abstract

The aim of this study was to investigate the effect of learning in the integration of local elements into coherent visual shapes. Human early (V1, V2, VP, V4) and higher visual areas known to be involved in the analysis of shape information (Lateral Occipital Complex-LOC) have been implicated in the perceptual integration of global shapes. We used event-related fMRI to investigate the effect of learning in the neural representation of shapes across the human visual cortex. The stimuli consisted of symmetrical and asymmetrical closed contours rendered by aligned gabor elements and embedded in random gabor background fields. Two stimuli were presented simultaneously in each trial to the left and right of the fixation point and the subjects were instructed to report which stimulus contained the symmetrical contour. Behavioral and fMRI responses were recorded while observers performed the 2-AFC shape discrimination task in two different sessions, one before and one after three days of training. Observers were shown different sets of novel stimuli in every session and a set of training stimuli rendered in novel backgrounds across trials. Prior to training, no differences were observed in the behavioral performance or the fMRI responses across visual areas for the novel vs. the training stimulus set. However, after training the observers showed significantly improved accuracy in the discrimination task for learned vs. novel shapes. Consistently, fMRI responses in the early and the higher visual areas were significantly stronger for learned than novel stimuli. These results provide evidence for behavioral and neuronal learning-based plasticity in the human visual areas involved in coherent shape perception.