Fig. S5.

Analogous to Fig. 5A, but with trials excluded based on ISI: A Only on those trials with $\mathrm{ISI}<4\mathrm{s}$ are included. B Only on those trials with $\mathrm{ISI}>2.35\mathrm{s}$ are included. C Only on those trials with $\mathrm{ISI}\in [2.35\mathrm{s},6\mathrm{s}]$ are included (wider range to retain a comparable amount of trials, and not loose too much statistical power). The exact estimates differ, yet the interaction of Condition and ISI remains similar.