The plots in the links below show the average firing rates for the 9 most selective neurons on match trials and nonmatch trials for both experiments pre and post training (i.e., these are the same plots as used for Figs. 3b and S4b and for the post-training results, the first three neurons are the same neurons are shown in those figures). Each neuron's firing rates are plotting using the bin size that gave rise to the smallest p-value on an ANOVA at any time point (see the methods section of the paper). As can be seen in these plots, prior to training the results look very noisy and there is no clear different between match and nonmatch trials; this confirms that the results shown after training are not merely due to selecting neurons that show small p-values, but rather there is a dramatic change in the response to match vs. nonmatch trials after training.