Explicit and implicit perception of illusory contours in unilateral spatial neglect: behavioural and anatomical correlates of preattentive grouping mechanisms.

Studies of hemispatial neglect suggest that some perceptual processes still operate on contralesional stimuli independent from spatial attention or awareness. Here we examined whether preattentive processing in extrastriate areas may group unconnected elements inducing illusory contours despite neglect. While it has been debated whether illusory contours arise from preattentive grouping or higher cognitive processes, neurophysiological studies show that neurones in secondary visual cortex (V2) can code for illusory contours. Twelve patients with right hemisphere damage and left neglect were tested for implicit and explicit detection of illusory contours using, respectively: (1) a bisection task where patients were not explicitly required to attend to lateral elements and judged the midpoint of Kanizsa illusory stimuli, as well as other physically connected or unconnected stimuli of the same length; (2) a matching task where patients had to overtly attend to lateral elements and made same/different judgements on pairs of illusory stimuli with identical or different inducers on the right or left side. In some patients, bisection judgements were consistently similar on Kanizsa stimuli with illusory contours and connected stimuli with real contours but different on unconnected gap figures, regardless of their length, suggesting implicit grouping of inducing elements prior to processing stages where a spatial attentional bias arose. Their lesions centred on the inferior parietal cortex or thalamus. Other patients did not show a systematic bisection pattern and had lesions extending posteriorly in the lateral occipital cortex. However, both groups of patients failed to detect left-side inducers in explicit matching judgements, even though errors often revealed unconscious processing, and they showed similar neglect severity on other standard tests. These findings suggest that grouping by illusory contours can occur preattentively and influence bisection independently from the ability to detect contralateral inducers explicitly, severity of inattention, and other forms of unconscious processing. Implicit grouping may depend on the sparing of lateral occipital areas involved in figure-ground segmentation at early stages of visual processing.