Imaging of neural oscillations with embedded inferential and group prevalence statistics
Fig 5
Comparing statistical performance of iES with minimum-norm imaging for mapping induced oscillations.
a) Subspace scanning results: (left) significant average subcorr map (p < .05, see text for procedure). Note that results were obtained from the subgroup of participants that presented the hypothesized effect (n = 8, see Fig 4). (right) histogram from observed data and permutation tests to derive a subcorr threshold corresponding to p < .05. b) Minimum-norm imaging results: average maps of log-transformed power ratios (stimulus/baseline, p < .05). Note that the distinction between positive and negative effects is not possible. The results were derived from the same subgroup (n = 8) to allow comparison with a), the results obtained with the full group (n = 17) are shown as an outline. Contrary to iES, no increase in beta power could be detected over the right post-central gyrus region, with the same subgroup of subjects. Unthresholded maps are shown in the supplementary material. (right) histograms of observed data and permutation tests to determine significance of minimum-norm maps at p < .05. Note how the strong negative effects inflated the permutation distribution and prevented the detection of the smaller positive effects. As shown using iES, positive and negative effects could be evaluated separately and specifically.