 |
 |
Hi, Probably the easiest and most accurate way to do this all is just to run a second-level FE cross-run within-subject analysis and include the contrasts (eg A vs C) at that level - this will then generate the optimal COPEs and VARCOPEs for you, easily extracted with Featquery. Cheers, Steve. On 12 Dec 2007, at 22:50, Michelle Voss wrote: > Hello all, > > I'm trying to determine whether it's valid to create an effect size > of sorts that can compare %signal change within an ROI for separate > conditions of separate block designs. For example, say I have an > ROI and I want to examine the degree to which it responds greater > for a specific condition compared to another condition, but the > conditions are in separate runs (but same session). E.g., you for > one fMRI session a blocked run with blocks for stim types A and B, > and another for stim types C and D, and post hoc you want to compare > response within an ROI for stims A and C. > > I've run featquery for copeA and copeC to extract the mean percent > signal change for these conditions compared to baseline within the > ROI. If I want to normalize this difference by a variance term, > could the correct variance term be given by sqrt(average of varcopeA > and varcopeC). doing the latter by running featquery just as before > but 1) not converting to %signal change and 2) extracting the mean > varcope within the ROI instead of mean cope. from this i have the > mean varcopeA and mean varcope C within the ROI, from which I can > take the average, and the sqrt of this average to have a pooled > standard deviation as a variance term to be used something like: > > within an ROI: > > (mean %signal change stim A - mean %signal change stim C)/pooled > standard deviation > > but, is this the right variance term for what i want to do here? > the reason i'm interested in the variance term is because i'm > comparing two groups with potentially different variances, so i want > to normalize the difference in response by a variance term.. > > > best, > Michelle > > > > --------------------------------------------------------------------------- Stephen M. Smith, Professor of Biomedical Engineering Associate Director, Oxford University FMRIB Centre FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK +44 (0) 1865 222726 (fax 222717) [log in to unmask] http://www.fmrib.ox.ac.uk/~steve ---------------------------------------------------------------------------