Blindsight is color-blind to S-cone isolating stimuli: An fMRI study

Abstract

Purpose: To investigate and compare the role of achromatic contrast and S-cone contrast in blindsight in hemispherectomized subjects. Methods: We designed an fMRI paradigm using sinewave checkerboard stimuli (0.5cpd, spatial envelope s=1.5cycle, temporal envelope s=250ms) that isolated either the achromatic pathway or the S-cone (blue/yellow) pathway. Stimuli were presented binocularly at 5.4° in the right, left, or in both visual fields. We tested three subjects, one subject with blindsight, one subject without blindsight, and one control subject. We investigated hemispherectomized subjects to exclude the possibility that blindsight is mediated by spared islands of visual cortex. Results: Achromatic and S-cone isolating stimulus presentation in the normal visual field of both hemispherectomized subjects and to both visual fields in the healthy subject activated contralateral visual areas (V1 & V2). The hemispherectomized subject with blindsight, however, showed cortical activation to achromatic stimuli presented to his blind visual field (FEF & V5) but had no significant cortical activation to S-cone isolating stimuli. The hemispherectomized subject without blindsight showed no cortical activation to any stimulus presented in his blind visual field. In the control subject, presentation of an additional stimulus increased the cortical activation relative to that associated to a single stimulus for both stimulus types. The hemispherectomized subject with blindsight demonstrated this enhancement with achromatic stimuli only, consistent with our hypothesis that S-cone isolating stimuli are not processed in the blind field. The hemispherectomized subject without blindsight did not show any enhancement to bilateral stimulus presentations. Conclusions: These results demonstrate that hemispherectomized subjects are unable to process S-cone isolating stimuli in their blind field. This is consistent with the mediation of blindsight by the superior colliculus, which has been shown to lack S-cone inputs in primates as well as in our present subjects, and supports our previous behavioural results [Leh et al., 2007].

Footnotes

 Support: CRIR & FRSQ doctoral grant to SEL, FRSQ postdoctoral grant to SEL, NSERC research grant RGPIN 37354-02 to AP, CIHR grant MOP-10819 to KTM.