Effectiveness of different task paradigms in revealing blindsight.

Four patients with hemianopia from posterior cerebral artery infarction were tested for residual unconscious vision ('blindsight') in their anopic hemifield. One task tested for spatial summation between the normal and the abnormal hemifields. A second task tested for temporal interactions between the two hemifields. A third task required a manual localization of the site of a brief flash in the hemianopic side. The last test was administered either in a condition of free gaze during pointing or in a condition in which the gaze was kept fixed throughout all phases of the task. The first two tasks differed from the third not only in the type of visual function, that is, detection vs spatial localization, but also in the response criterion. A forced-choice procedure was adopted in the localization task while in the spatial and temporal interfield interaction tasks a reaction time paradigm that did not require guessing was used. One patient showed unequivocal evidence of blindsight on all tasks, thus excluding the possibility that his blindsight depended on the nature of the response. An interesting dissociation was found in the localization task where 1 patient showed evidence of blindsight when tested under conditions of unrestrained gaze during pointing but not when tested under fixed gaze conditions.