Hi,
I am performing fMRI analysis of preliminary data collected for spontaneous human behaviour, recorded using skin conductance measures. I scan each participants for two runs of 20 minutes. These behaviours can apear randomly any time during the run. But they are usually clustured in time. I was wondering if someone could point me towards the right direction to analyse these data by answering the following questions:
1) Is it valid to epoch fMRI time-series from time of start to the end of appearance of those events and perform analysis only based on that fMRI time series? This approach will also be usefull if there is a sudden motion prior to or after the appearance of those clustered events.
2) If we cannot epoch fMRI-time series as mentioned above, how do we account for variance introduced by more than 70 % of the total fMRI time-series that may not contain any events of interest.
3) If some subjects show these behaviours after 10 minutes and some subjects after 15 minutes, do we need to account for this time-difference in occurance of those events.
4) How should I model for haemodyanamic activities that occur prior to occurance of these events and that may potentially cause these events.
Regards,
Peter T
