Hemispheric asymmetries in global⧹local processing are modulated by perceptual salience

Abstract

It has been claimed that a left hemisphere bias toward local and right hemisphere bias toward global visual processing can be explained in terms of specialization for relatively high and low spatial frequencies, respectively. Using non-representational figures, we tested this hypothesis in experiment 1 using positron emission tomography (PET) measures of cerebral activity in 10 normal volunteers. In each block of trials subjects viewed either a relatively high or a relatively low spatial frequency grating. The orientation (vertical or horizontal) of the grating changed from trial to trial. In a directed attention task, subjects reported the orientation of either the whole stimulus (globally directed attention) or the orientation of the component parts thereof (locally directed attention). A significant interaction between hierarchical processing level (global or local) and stimulus level (relatively high or relatively low spatial frequency within the absolute low spatial frequency range) was found. Globally directed attention led to significantly increased cerebral activity in the right hemisphere when relatively high spatial frequency stimuli were used but not when relatively low spatial frequency stimuli were used. Likewise, locally directed attention increased cerebral activity when low but not high spatial frequency stimuli were used. On the basis of these results we suggest that perceptual salience of the global or local form modulates hemispheric processing asymmetries in early visual cortex. In experiment 2, the perceptual salience of global form in relatively high spatial frequency stimuli and of local form in relatively low spatial frequency stimuli was confirmed in a reaction time (RT) study. In combination, the results of the two experiments suggest that perceptual salience takes precedence over spatial frequency (within the range studied here) in determining the cerebral organization of global⧹local processing.