mne.stats.permutation_cluster_1samp_test

Support & Discussions
2

Hi @CarinaFo,
if you want to compare two conditions, you want to use mne.stats.permutations_cluster_test, which as the first argument takes a list of conditions, where each condition is an array of (participants, ...) shape.

To compare your tfrs, I would:

  • first crop the tfrs to the desired time range (in your case - prestimulus) and frequency range:
     tfr_cond1_crop = [tfr.copy().crop(tmin=..., tmax=..., fmin=..., fmax=...) for tfr in tfr_cond1]
  • then stack all tfr arrays for each condition:
    tfr_array_cond1 = np.stack([tfr.data for tfr in tfr_cond1_crop], axis=0)
  • your X argument is now [tfr_array_cond1, tfr_array_cond2] (if you are going to use a t test the order of the conditions will affect the sign of the t values)
  • now you are almost ready to conduct the cluster permutation test, you just need to specify channels adjacency and then build the whole adjacency for the channels-time-frequency space using mne.stats.combine_adjacency. You also need to choose the correct stat_fun (that depends on the test you want to perform - I guess it is within subjects t test in your case?)