 |
 |
Thanks so much, Donald! Very insightful as usual! Best Regards, Daniel From: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>> on behalf of "MCLAREN, Donald" <[log in to unmask]<mailto:[log in to unmask]>> Reply-To: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>> Date: Friday, September 25, 2015 at 11:17 AM To: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>> Subject: Re: [FSL] block designs and PPI analysis Yes. They can still be interpreted. Without any rest periods, you can simplify the model and treat it as only 1 condition and the other condition would be the implicit baseline. Best Regards, Donald McLaren, PhD On Fri, Sep 25, 2015 at 11:09 AM, Yang, Daniel <[log in to unmask]<mailto:[log in to unmask]>> wrote: Dear Donald, Thanks so much for your reply. Very helpful as always. I am wondering what if there is no rest period at all, that is, only ABABABAB… Could AxROI-BxROI, that is, (1) - (2), be meaningfully interpretable in this case? Many thanks! Daniel From: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>> on behalf of "MCLAREN, Donald" <[log in to unmask]<mailto:[log in to unmask]>> Reply-To: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>> Date: Friday, September 25, 2015 at 11:01 AM To: "[log in to unmask]<mailto:[log in to unmask]>" <[log in to unmask]<mailto:[log in to unmask]>> Subject: Re: [FSL] block designs and PPI analysis Both designs can be assessed. For both designs, as you have at least one period of r, you will want to use the generalized PPI approach (McLaren et al. 2012). The PPI design matrix will have 5 columns: (1) PPI for Task 1 (2) PPI for Task 2 (3) Task 1 Regressor (from task model) (4) Task 2 Regressor (from task model) (5) Seed time course Then you can compute the contrast for (1)-(2). Hope this helps. Best Regards, Donald McLaren, PhD On Fri, Sep 25, 2015 at 10:47 AM, Yang, Daniel <[log in to unmask]<mailto:[log in to unmask]>> wrote: Dear FSL experts, When running a PPI analysis with a block design, I am wondering if there is a restriction to whether a certain kind of block design may be more appropriate. Specifically, there are two kinds of block design: 1. rABABABAB… 2. rArBrArBrArBrArB… Can PPI be applied to both designs? My concern about 1. Is that the time series of a given ROI may confound the effect of both A and B, and may lead to results uninterpretable? Thanks, Daniel