Question on statistical comparison between PSD of 5 subjects (ieeg data)

💬 Support & Discussions
1

I am trying to compare the 5 subjects using PSD values. My null hypothesis is there is no difference in the mean PSD values of the 5 samples. The steps are:

  1. Calculate PSD of each subject using *mne.time_frequency.psd_multitaper (no. of freq x no of channels) for the task=AUDI
  2. find mean PSD at each frequency (averaged across channels - no of freq x 1)*
  3. Take log10(mean PSD )
  4. Do the same for 5 subjects (s1,s1,s3,s4,s5)
  5. Use scipy.stats.kruskal(s1,s1,s3,s4,s5, axis=0) – samples are non-parametric and independent
  6. I get the p-value - 0.0000e+00

Is my approach correct? If not, then please suggest me some other methods to compare the 5 subjects.

Thanks