<div>Hi everyone,</div>
<div> </div>
<div>I have had some success processing the subject-level differences between two conditions for phase-based measures such as coherence and plv but I was wondering what the approach would be for the directed or mvar-based measures such as dtf, granger, or pdc? Since there are no trials to shuffle as you normally would for the phase-based connectivity measures, what would be the optimal statistical approach? I thought about shuffling the trials and recomputing the mvar over and over again for each condition as a simulation but this doesn't seem practical as it would be too computationally intensive.</div>
<div> </div>
<div>Thank you,</div>
<div>-Tony</div>