 |
 |
Hi Sharlene, You might wish to thoroughly familiarize yourself with this paper: http://www.ncbi.nlm.nih.gov/m/pubmed/24650594/ And the software package by Tom Nichols's group: http://www2.warwick.ac.uk/fac/sci/statistics/staff/academic-research/nichols/software/swe Best RCW ----------------------------- Robert C Welsh, PHD Associate Professor Department of Radiology Department of Psychiatry University of Michigan ------------------------------ typos due to iPhone 4S > On May 25, 2015, at 13:00, Sharlene Newman <[log in to unmask]> wrote: > > SPM users: > > We have recently conducted a longitudinal fMRI study in which we collected three time points. There was an uneven number of participants per time point, such that we went from 35 to 25 to 12 subjects (which means we have 12 subjects that came back for all 3 sessions, but several others attrited). > > We have two questions: > 1) Is there an analysis technique that we can use to compare activation using all of the subjects (including the attrited ones)? Currently we have analyzed the data with just the 12 subjects using paired t-tests between the time points. > > 2) More importantly, we are interested in correlating individual differences measures with activation across time. For instance, we collected Stroop data. We want to see how the correlations with their Stroop performance changed over time, including all subjects and not just the 12 that came back all three time points. > > Any advice here will be greatly appreciated. Thanks. > > Sharlene