 |
 |
yes indeed, Donald is right,I assumed event related . If you search the email database a post from Jesper Andersson (a long time ago) explains that this is actually not a good idea to add derivatives for blocks cyril <http://www.ed.ac.uk/edinburgh-imaging> ________________________________ From: SPM (Statistical Parametric Mapping) <[log in to unmask]> on behalf of MCLAREN, Donald <[log in to unmask]> Sent: 14 July 2016 01:57:41 To: [log in to unmask] Subject: Re: [SPM] Directional contrast with temporal derivatives at individual subject level If its a block design, then the HR latency will have a minimal effect as the block represents a stable state that will drive the estimates. I would try without the temporal derivative. Best Regards, Donald McLaren, PhD On Sat, Jul 9, 2016 at 5:15 PM, Ekaterina V. Pechenkova <[log in to unmask]<mailto:[log in to unmask]>> wrote: Dear members of SPM list, is someone aware of how one can perform directional contrasts on individual subject level after including temporal derivatives into the model? For the group level, I'm using Calhoun et al. (2004) / Steffener et al. (2010) approach of 'boosting' con images with the temporal derivatives. But besides group analysis I also need individual activation maps for each subject. I cannot sacrifice the information from the temporal derivatives at the individual level because the subjects are from a special population and there are somewhat pronounced diffrerences in the HR latencies. At the same time I'm not satisfied with the results from the F-test incorporating both the canonical HRF and the temporal derivaitve, since it is bi-directional and introduces the certain amount of 'deactivation' that we're not interested in. The design is very simple, it's just a block-design motor task. I even think of masking the 'deactivation' out, but hope there is still a better way. Can someone suggest a solution, please? Sincerely, Ekaterina.