Fig. 4.

$\mu$-band power is related to the response and condition: In the top row (A, B), the response is plotted against the z-scored logarithm of the bandpower of the (A) left-hemispheric/(B) right-hemispheric signal (in hexbin plots). No strong correlation is visible (Pearson’s R$<0.3$ for both). In the bottom row (C, D), the distribution of z-scored log power is contrasted in the two conditions (red: high, blue: low). Clearly, high-condition trials tend to have higher $\mu$-power too. This naturally begs the question, whether the difference in MEP amplitude between the high vs. low conditions could not be better explained by the $\mu$-band power — motivating the model given in Eq. (4).