FIG. 1.

Plots of reward- vs. rank-selective indices, before and after Bayesian correction. Top: uncorrected indices. The x axis represents the index value for the the serial order saccade task. This is obtained through I_rank = (f₃ − f₁)/(f₃ + f₁), where f₁ and f₃ were the mean firing rates measured at the times of the 1st and 3rd saccades, respectively. The y axis indicates the index of selectivity for the size of the anticipated reward: I_reward = (f_b − f_s)/(f_b + f_z) where f_b and f_s were the firing rates during the postcue delay period on big-reward and small-reward trials, respectively. Bottom: plot of posterior means representing values after correction for noise.