ROI analysis
As the cluster-based permutation statistics indicated an interaction of
pre-cue condition by experimental phase in source clusters covering
various ROIs as defined by the co-registered brain atlas (see above),
mean relative ssVEF power across sources that intersected with
significant source clusters and ROI regions were analyzed with a
repeated measures ANOVA with within-subject factors of pre-cue baseline,
experimental phase and ROI collapsed across hemispheres for homologous
left and right source clusters if both ROIs indicated an interaction in
the previous cluster based permutation statistics. For ROIs that only
indicated an interaction in the right hemisphere, only right hemisphere
relative ssVEF power values were considered. Greenhouse-Geisser
corrections were applied when necessary.
As this analysis revealed a significant pre-cue baseline by experimental
phase by ROI interaction, a repeated measure ANOVA with within-subject
factors of pre-cue baseline and experimental phase was fitted to the
data for each ROI, separately. In order to detangle the observed
interactions at each ROI, modulation differences between conditions were
compared using paired t-tests (testing the difference of condition
differences). Finally, mean relative ssVEF power within each ROI were
contrasted between the two pre-cue baseline conditions (attend central
fixation cross vs. attend both peripheral rings) using paired t-tests.