Visuospatial neglect in action

Abstract

It is well established that patients with hemispatial neglect present with severe visuospatial impairments, but studies that have investigated visuomotor control directly have revealed diverging results, with some investigations finding impairments mirroring the perceptual difficulties of these patients, while others have shown that such neglect patients perform relatively better in action tasks.In this review we attempt to reconcile these diverging findings, addressing differences in the type of visuomotor tasks studied but also highlighting the diverging neuroanatomy that seems to be driving the differences in performance.We argue that there are different types of actions and that these in turn depend on different cortical networks (Goodale, Westwood, & Milner, 2004; Milner & Goodale, 2006). Patients with visuospatial neglect, in contrast to patients with optic ataxia, are relatively unimpaired at performing target-directed tasks even towards stimuli located in their ‘neglected’ field. We relate these findings to the view that for the on-line guidance of action, spatial information is coded in egocentric coordinates and depends on the visuomotor networks of the visual dorsal stream. Furthermore, based on recent lesion-symptom mapping studies, we postulate that deficits in on-line actions that are observed after right-brain damage are associated with damage to the visuomotor control network, in particular with damage to the basal ganglia, frontal and parieto-occipital regions. On the other hand, clear neglect-specific deficits emerge when the action is off-line and not directly target-driven, thus requiring relational metrics or scene-based coordinates (as is the case for example in delayed and mirrored (anti-pointing) reaches). We review recent studies that support our argument that such deficits in off-line actions are associated with damage to occipito-temporal and parahippocampal cortex, perhaps as part of the ventral visual stream or areas where information from the two visual streams is combined.