(a–c) Trials from Monkeys S, C, G were re-labeled as low, mid-low, mid-high, and high according to the non-normalized beta power during time slice −0.8 to 0.0 s with respect to the end of the neurofeedback epoch. The movement onset times (MOTs) of the re-labeled trials were compared and the mean (s.e.m) are plotted in each subplot. Below titles, p-values are shown for Cuzick’s test. (d–f) The same as (a-c) except using normalized non-beta frequencies indicated at the top of plots. (g) Right shows MOTs for version of task where the X axis position of the cursor is controlled by normalized 1–10 Hz power in addition to the Y axis position controlled by normalized 25–40 Hz power. Note that Cuzick’s test for the right plot assumes ordering of targets is mid-low, mid-high, high, low (instead of low, mid-low, mid-high, high). Left, same MOT plot for Monkey G from Figure 3c for comparison. (h, i) Z scored PSD plots (same method as Figure 2d–f) for different beta targets in the standard beta neurofeedback task (h), and the neurofeedback task that incorporates 1-10 Hz power modulation to control X axis position (i). Dotted lines indicate that the ordering of targets in the beta range follows the movement onset time ordering in (g). (j) Percent of time points within the last 0.8 s of the neurofeedback epoch that are part of beta oscillatory episodes during the task requiring both beta and 1–10 Hz control. *p<0.05, **p<0.01, ***p<0.001, Cuzick’s two-tailed test for ordered grouping.