If we do planned comparisons on these data, the difference between the two partially incongruent conditions is significant in the colour

task [t(6) = −3.32, p = .01; colour incongruent > shape incongruent], and a trend in the shape task [t(6) = 2.04, p = .08; shape incongruent > colour incongruent 2], with this pattern also evident in all synaesthetes individually. The identical analysis on control data from these conditions show no reliable difference in the colour task [t(6) = −.97, p = .36] and a reliable difference in the shape task [t(6) = 2.39, p = .05; shape incongruent > colour incongruent]. In Supplementary Materials, we report an alternative exploratory analysis, ERK inhibitor research buy which treats each feature as an individual congruency

factor, to test how task-related attentional set modulates the respective impact of synaesthetic colour and shape. The results are consistent with the planned comparisons, such that, for synaesthetes only, the impact of synaesthetic colour is more powerful PI3K inhibitor in the colour than in the shape task and, conversely, the impact of synaesthetic shape is stronger in the shape than in the colour task. The same analyses on the error rate of each condition reveal a significant main effect of congruency [F(2, 24) = 4.15, p = .02, η2 = .25], with no post-hoc tests being significant (all ps > .10). No other statistics reached significance (all ps > .12). Errors (2.5%) and outliers (.2%) were excluded from further analyses. Fig. 6 shows the mean correct RT and repeated-measures SE of each condition for

synaesthetes and controls. The mean error rate of each condition is reported in Table 2. Note that in Experiment 2 we used different image sets in the colour and shape task to control for the effects of the third feature (shape or colour in different tasks). The displayed shape was always congruent with the synaesthetic shape in the colour task and vice versa for the colour in the shape heptaminol task, while the other feature and location were manipulated. Therefore, we conducted separate analyses for the colour and shape tasks. All other aspects of the analyses matched Experiment 1. For the colour task, we carried out a mixed design ANOVA with a between-participant factor of group (synaesthetes vs controls) and a within-participant factor of congruency (both features congruent, location incongruent, colour incongruent, and both features incongruent). Consistent with the pattern we found in Experiment 1, synaesthetes showed effects of synaesthetic congruency that were not present in controls. The ANOVA revealed no significant main effect of group (F < 1.0, n.s.), a significant main effect of congruency [F(1.57, 18.92) = 10.10, p = .002, η2 = .45], and a significant group × congruency interaction [F(3, 36) = 5.47, p = .003, η2 = .31; see Fig. 6a]. Post-hoc tests (the Bonferroni corrected α-level: .