Figure 6

SARAT performance. (A) NRT to valid and invalid cues. Even after controlling for simple reaction time, fXPCs were slower than HCs (p < 0.001 in both the four-cue and three-cue ANCOVA). As expected, the effect of validity in the three-cue ANCOVA was significant (p = 0.02). Both groups showed a pattern of increased NRT with more validly cued locations, although the effect of cue number was not significant (all p > 0.17). (B) NRT to valid cues, for high- and low-saliency targets. As expected, the participants were slower to respond to low contrast targets than to high-contrast targets (p < 0.001 in both the four-cue and three-cue ANCOVA). FXPCs were as slow to respond to the high-contrast targets as HCs were to respond to the low-contrast targets (t =-0.97, p = 0.34). The interaction between group and saliency was not significant (both p > 0.76). The interaction between group and age was significant (both p < 0.001), such that NRT to valid cues increased with age in HCs (r = 0.60, p < 0.001) but decreased with age in fXPCs (r =-0.24, p = 0.003). (C) Cue cost for each number of cued locations. As expected, the effect of cue cost was significant, such that as the number of cued locations decreased, and the spatial predictability of the target increased, cue cost also increased (p = 0.004). (D) Cue cost for each number of cued locations, for high- and low-saliency targets. The effect of salience was significant, such that cue cost was greater for low-saliency than high-saliency targets (p = 0.009). Error bars represent SEM but are not visible when they are smaller than the data points.