Abstract
An inexpensive method to characterize visual field defects due to neural damages was developed. Stimuli are presented on a large computer screen at a number of pre-selected locations in a randomized order. The sensitivities at these locations are measured independently with a number of simultaneously running threshold estimation algorithms. The stimuli are M-scaled, which means their size is increased with increasing eccentricity in inverse proportion to the cortical magnification factor. The purpose of M-scaling was to make the stimulus representation homogenous on the primary visual cortex so that the sensitivity values measured at different visual field locations would be comparable. The task was either the detection of flickering or static disks or recognition of flickering or static letters. Two groups of subject participated in the experiments - normal sighted and hemianopic patients. There was a general tendency that sensitivities increased toward peripheral vision to some extent in the intact visual field. This effect was slightly more pronounced for letter recognition than disk detection. In disk detection flicker more than doubled the sensitivity in comparison to static stimulus detection. However, in letter recognition flicker only had a negligible effect. In the hemianopic visual field, the sensitivity was in general considerably lower than in the intact visual field. However, completely blind spots were rare, and all patients exhibited residual vision of the defected visual field. Hemianopia produced a reduction of relative sensitivity that was not significantly dependent of stimulus type. The results also showed that the new method is sensitive to cortical visual field defects and it is reasonably fast.