Fig. 4: Models based on probabilistically-blurred, reciprocal PDF capture reaction time (RT).
a Model fits to auditory (red), visual (blue), and somatosensory (violet) RT based on the temporally-blurred, reciprocal PDF (black line) in the PDExp condition. b As a but for PDFlip condition. Since temporal blurring increases with ‘go’ time, it has a stronger smoothing impact on the model at longer ‘go’ times. This feature is not reflected in the data (‘go’ time span 1.2 to 1.43 s highlighted). c Model fits to RT based on the probabilistically-blurred, reciprocal PDF in the PDExp condition, d As c but for PDFlip condition. e Comparison between two reciprocal PDF-based models (“temporally-blurred” and “probabilistically- blurred”), in each of four equally spaced bins across the ‘go’ time span. e In the PDExp condition, the quotient of squared residuals (“temporally-blurred”/“probabilistically-blurred”) is relatively close to 1 in all bins, indicating residuals of similar magnitude across the two models and thus a similar goodness-of-fit. f As e but for the PDFlip condition, residuals are bigger in the temporally-blurred model compared to the probabilistically-blurred model in all bins, especially in bin #4 (highlighted). See Supplementary Figs. 1 and 3 for a detailed summary of all other fitted PDF-based models.
