Fig. 2.

Illustration of behavioral tasks, descriptive data, and model estimates. (a) Illustration of one trial of the cued flanker task that was used to assess attention. (b) Illustration of one trial of the EmpaToM task that was used to assess trial-wise experience of compassion and accuracy of mental state inference (i.e. ToM). Note that this simplified illustration omits fixation periods between the screens and the name of the speaker presented in the beginning of each trial. For details of the tasks, please refer to the methods section. (c-e) Descriptive plots of mean values for attention (difference of correct response proportions in the reorienting and conflict condition minus baseline condition), compassion (mean ratings on a scale from 0 to 6) and ToM (proportion of correct responses in ToM questions) per time point and group. Note that differences between groups at T0 were not significant (F-Test p-values all >.11). The mean values of each individual were used to calculate change scores for each available pair of two consecutive time points, which were used to estimate effects for retest and the three training modules shown in (f-h). In panels (g) and (h), to the right of the dashed lines, estimates are averaged across time intervals as was done to test the main hypotheses for these two measures. Error bars in (f-h) indicate 95% confidence intervals.