 |
 |
Regarding alternatives, I'll defer to others that have had success in getting "case study" rs-fcMRI studies published. Good luck, -MH On Tue, 2011-12-13 at 14:54 +0000, David Soto wrote: > Hi, thanks for that; I know about the independence issue from the > statistical perspective > > > I thought about this becos it is a common trick used in behaviour > neuropsychological studies of single patients, where > dependent measures are not independent but treated as such.... > > > what would be the alternative option then? > On Tue, Dec 13, 2011 at 2:47 PM, Michael Harms > <[log in to unmask]> wrote: > Hi David, > 4-5 scans of the same subject are not going to be > "independent" scans. > Statistics resulting from treating them as independent are not > likely to > be valid. > > cheers, > -MH > > On Tue, 2011-12-13 at 09:03 +0000, David Soto wrote: > > Apologies for re-posting this in a short time period. I am a > bit on a hurry on this because the patients > > are coming for a scan this thursday and would like to know > how best to use the time > > > > The thing is that we have a couple of focal lesion patients > (with a rather interesting behavioural impairment) > > and we want to assess resting state functional connectivity > with regard a group of age match controls. using concat ICA > followed by dual regression. > > > > What I thought to increase power (given only 2 patients), > is to take 4 or 5 resting state scans for each of the 2 > patients-- this way can have 8-10 samples which I could > 'treat' as the patient group into randomise of dual > regression, so each resting state scan would be treated as an > 'independent' subject > > > > what you do think? any advise appreciated > > > > Thanks, David > >