 |
 |
The flexible factorial model in SPM with 3 factors (group, time, and subject) should work fine for assessing differences in learning between young/old groups. The interaction tests whether the change between pre- and post-test is different between young and old subjects. The one caveat is that this does not adjust or control for learning differences if learning is related to pre- learning connectivity. For that, you'd need a different test as follows: (1) Compute the difference in pre- and post- connectivity scores; (2) Create a 2 sample t-test with a covariate and have the covariate interact with factor 1; (3) You can test whether the difference in connectivity is different between groups and/or if the connectivity change with learning is related to baseline connectivity and/or if the relationship of the connectivity change to baseline connectivity varies as a function of group. If the latter is the case, then you should not report the findings of group differences as the group difference will be dependent on baseline connectivity. I'd recommend reading the following webpage: http://mumford.fmripower.org/mean_centering/ Additionally, as you will have an imaging covariate, you will need to use Biological Parametric Mapping (http://fmri.wfubmc.edu/software/Bpm); however, the website only lists compatibility with SPM8. There seems to be another toolbox on the nitrc website, that may be useful (I haven't downloaded it or tested it): http://www.nitrc.org/projects/rbpm/ One possibility is to pick one connection to use as the baseline connectivity, this would eliminate the imaging covariate from the model and you could use SPM without BPM. Computing a 2 sample ttest for the pre-learning and then for post-learning will show you group differences for each part of your study; however, there is no statistical way to say that the effect of learning is different between groups based on those two ttests. Likewise, 2 paired ttests for each group wouldn't allow you to say the groups are different. You need to be able to test the interaction within a single model to be able to say learning has differential effect between groups. Hope this helps. Best Regards, Donald McLaren ================= D.G. McLaren, Ph.D. Research Fellow, Department of Neurology, Massachusetts General Hospital and Harvard Medical School Postdoctoral Research Fellow, GRECC, Bedford VA Website: http://www.martinos.org/~mclaren Office: (773) 406-2464 ===================== This e-mail contains CONFIDENTIAL INFORMATION which may contain PROTECTED HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and which is intended only for the use of the individual or entity named above. If the reader of the e-mail is not the intended recipient or the employee or agent responsible for delivering it to the intended recipient, you are hereby notified that you are in possession of confidential and privileged information. Any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited and may be unlawful. If you have received this e-mail unintentionally, please immediately notify the sender via telephone at (773) 406-2464 or email. On Tue, Mar 17, 2015 at 9:32 AM, Elisabeth Kaminski <[log in to unmask]> wrote: > Dear colleagues, > > I have problems in finding the best design for a learning study comparing > old and young participants. > We acquired pre and post resting state connectivity and now want to do a > seed-based correlation analysis and compare the pre and post connectivity > among the 2 subgroups. > What do you think might be the best design to answer this question? > Especially because there are definite baseline differences between young > and old subjects regarding their connectivity. Is there a way to somehow > control for these baseline differences to only have the pure learning > difference? > I already tried to set up a flexible factorial design and look for > interactions but often there are none. If I compute 2sample ttests pre and > post, you can see differences between pre and post significant areas but I > guess this is not the appropriate way to approach this question. > Any ideas would be highly appreciated. > Thanks a lot in advance. > Elisabeth >