Fig. 5: Integration of the results of the four onset-tracking methods.

a Rank distributions (i.e., ignoring the numerical value of onset) combined for all four methods from Fig. 4b–e, with each dot representing one bootstrap (N = 86) from one method (N = 4). All areas shown here achieved significant conjunction. Areas are ordered by their average rank; b average onset time of activation (color coded), relative to onset of area BA4, for all areas with significant consistency between methods. Each dot is located at the center of mass of all recording leads that contributed and the size of the circle represents the number of recording leads; c schematic of information flow suggested by the results in Figs. 3, and a and b: stimulus information and rule information all converge onto a shared decision process, consisting of an early (dlPFC, PFt, and PMd) and a late phase (PFcm, OP1, and OP3). See “Discussion” section for a detailed description.