Fig. 6: Reaction time (RT) variance is sensitive to the probability distribution of events.
a In all three sensory modalities, QQ-Plots reveal heavier right tails in the RT distributions in the PDFlip condition. b Mean of median RT in PDExp (‘E’) and PDFlip (‘F’) conditions. Average RT is not sensitive to the presented probability distribution. (planned contrasts, *P < 0.05, **P < 0.01, two-tailed Student’s t-test). c Mean of interquartile range of RT in PDExp and PDFlip condition. RT variance is higher in the PDFlip condition (planned contrasts, *P < 0.05, **P < 0.01, two-tailed Student’s t-test). d The Ex-Gaussian distribution is a convolution of an exponential and a Gaussian PDF. e The ex-Gaussian model fits RT histograms on the group level (e) and on the single-subject level (see Supplementary Fig. 15). f–h Ex-Gaussian fit parameters from single-subject fits. f Mean of Gaussian μ resembles mean of median RT (compare to b, planned contrasts, *P < 0.05, **P < 0.01, two-tailed Student’s t-test). g Mean of exponential τ resembles interquartile range of RT (compare to c, planned contrasts, *P < 0.05, **P < 0.01, two-tailed Student’s t-test). h Mean of Gaussian σ is not sensitive to the presented probability distribution (planned contrasts, *P < 0.05, **P < 0.01, two-tailed Student’s t-test). Error bars denote standard deviation. For ANOVAs, see Supplementary Tables 1–4.
