Mediation analyses of N1, N2, P2, and gamma responses

First, mediation effects of N1, N2, P2, and gamma responses were assessed. To this end, the amplitudes of N1, N2, P2, and gamma responses were entered into separate two-path mediation analyses to examine whether they mediated between stimulus intensity and the different pain dimensions. Thus, four separate mediation analyses were calculated for each condition, resulting in a total of 12 mediation analyses. Additional mediation analyses with covariates were performed whenever more than one brain response was found to be a significant mediator to investigate whether these brain responses carried redundant or complementary information. Likewise, additional three-path mediation analyses were performed using those brain responses, which were found to be significant mediators in the two-path mediation models as first and second mediator. Peak latencies of brain responses (Table 1) determined the sequence in which the two mediators were entered into the model. To control for the plausibility of the results, the three-path mediation models were re-calculated with an inverted sequence of mediators.

Finally, we aimed to compare the pattern of mediation effects across conditions. To this end, for all conditions separately, we determined the relative mediation effect of each brain response in comparison to the total mediation effect of all brain responses combined. The relative mediation effect was calculated based on the following equation (shown here exemplarily for the relative mediation effect of the N2 response in the perceptual condition):

Importantly, the mediation effect for each brain response was determined based on a mediation model in which all other brain responses were included as covariates.