Figure 5

Response times and Accuracy as a function of confidence. (a) Response times, (b) Accuracy. For both panels: each sub-panel represents a different participant. Dots are experimental data with 95% bootstrapped confidence interval as error bars. Lines are averages over 20 simulations of the attractor neural network model (calibrated as explained in the Methods Section). The shaded area represents the 95% bootstrapped confidence interval on the mean.