Global and event-specific changes in neural timescales. We estimated the effect size of the separation between areas’ change from baseline by calculating the rank-biserial correlation for each statistical test in Fig. 4 . We use the average effect size at each time point as an index of the differentiation between areas, with low values reflecting a global change in neural timescales (i.e., most areas displayed the same level of change), and high values reflecting area-specific changes (i.e., most areas were statistically different from each other). Y-axis: the average rank-biserial correlation ± s.e.m. across multiple comparisons (i.e., 28 comparisons per time point). X-axis: time point before and after the lever presses, with the colors indicating the associated statistical table in Fig. 4 . Circles: mean across statistical tests ± s.e.m ( N = 28).
