Longitudinal analysis

Hello, everyone,

I would like to post a statistical question rather on a Brainstorm-related question.

I am dealing with longitudinal data during 4 different visits of stroke patients during a motor task (T1,T2,T3,T4). T1 and T2 represent baseline measurements in different days and T3 and T4 represent measurements after an intervention at different times after the intervention.

My goal is to investigate the potential effect of my intervention in these patients on T3 and T4 evaluations.

After a preliminary data exploration, TopoplotTFR data are well reproducible with little variance. The topoplotTFR data during the T3 and T4 evaluations change with the intervention, but they vary with a different pattern for each subject (some subjects will increase their activity in an area, others will reduce it, some the area of activity will change, etc.... )

Cluster analysis does not reveal any obvious changes because there are different patterns within patients.

I tried to do a PCA that reveals 2 patterns of EEG activation and explains 70% of the variance. But I got stuck and I don't know how to use the results of the PCA to organize my group statistics.

So my question is, if anyone has a method or a different track in order to evaluate these longitudinal changes?

Thanks to the community

Kind regards,

Alexandre Chalard

@Sylvain @pantazis @John_Mosher ?