Blindsight is the term commonly used to describe visually guided behaviour elicited by a stimulus falling within the scotoma (blind area) caused by a lesion of the striate cortex. Such “vision” is normally held to be unconscious and to be mediated by subcortical pathways involving the superior colliculus. Blindsight is of considerable theoretical importance since it suggests that destriate man is more like destriate monkey than had been previously believed and also because it supports the classical notion of two visual systems. It is also of potential clinical importance, since it has been claimed recently that systematic practice in blindsight can lead to the recovery of normal visual function in patients with cortical lesions. From a review of the literature it is concluded that all of the phenomena of blindsight can be attributed either to light scatter into unimpaired parts of the visual field or to residual vision resulting from spared striate cortex. The possible contribution f other factors is also considered. It is concluded that blindsight studies have generally failed to control for such nonblindsight interpretations partly because of poor methodology and partly because of difficulties in defining the term “blindsight.”

Experiments were carried out to investigate the extent to which subjects can exhibit performance similar to blindsight when they are using scattered light as a cue. This was done both with hemianopic subjects (by manipulating the amount of scattered and direct light coming from a stimulus) and with normal subjects (by presenting targets within their blind spots). Good blindsight performance was observed when only scattered light was available as a cue to the subjects. It is therefore concluded that an adequate case for blindsight has not been made. It is probably impossible to demonstrate the existence of blindsight on purely behavioural grounds. What is required is the establishment of relationships between visual function and independent anatomical evidence.

Unilateral spatial neglect entails a failure to detect or respond to stimuli in the space opposite to a brain lesion. However, the contralesional hemispace can be determined by different frames of spatial coordinates, such as eyes-, head-, body-, or environment-centered coordinates. We observed 2 patients with a right hemisphere stroke whose left spatial neglect was modulated by distinct coordinates systems depending on the task. Four tasks were given in different conditions of central gaze and either the eyes or the head rotated 30° to the right or 30° to the left. While the 2 patients had a retinotopic defect in 1 visual field quadrant that remained the same irrespective of gaze direction (upper or lower quadrant in 1 case each), the other quadranopic field defect improved with eyes rotation to the right but not with head rotation, suggesting a head-centered spatiotopic deficit. Performance on line bisection was influenced both by eyes and head rotation, as well as by the position of the lines with respect to the trunk midline, suggesting the involvement of both head-centered and body-centered coordinates. Visual imagery and auditory extinction were not modified by changing the eyes or head position. These findings suggest that distinct spatial coordinates are brought into play depending on the tasks demands. (JINS, 1999, 5, 75–82.)