 |
 |
Hi Phillipe The answer to your question will depend on the details of your FIR model and on what you mean by percent signal change. I have taken the liberty copying your question and my reply to the SPM list - it's a slightly tricky issue and someone will hopefully correct me if I have it wrong. The units of SPM parameter estimates in general represent 'response per event', and this is true of FIR parameter estimates in SPM2 as long as the original event has duration = 0 (i.e. a standard event). [Note however that if you are modelling events of non-zero original duration using spm_graph PSTH plotting (or any of the spm_graph hacks) then you will get a response per second's worth of event - see http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind04&L=spm&P=R499765&D=0&I=-1 for details] The units of this response per event for each FIR timebin depend on whether you have used grand mean scaling to 100 - the SPM2 default. If you have, then they are percent *global* signal change, and if you want local signal change you need to normalise by voxel values, for example as suggested by Will - http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind04&L=spm&D=0&I=-1&P=344672. [The situation is slightly different for non-FIR basis functions such as the canonical HRF, as explained by Karl - http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind03&L=spm&P=R131393&I=-1&m=6166] Perhaps reassuringly, though, it doesn't necessarily matter if one's plots are in arbitrary units, as we don't really know what the *real* physiological units of interest are, as Geraint Rees points out in an earlier post (http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind01&L=spm&D=0&I=-1&P=103670)!. I hope this helps Alexa On Apr 15 2005, Philippe Tobler wrote: > Dear Alexa, > > may I bother you with another FIR-related question? I was wondering > what steps I would need to do in order to transform the dimensionless > "effect size" that I get out of an FIR into % signal change and > whether it would be possible at all? > > All the best > Philippe > -- Dr. Alexa Morcom Department of Psychiatry University of Cambridge Box 189 Addenbrookes Hospital Cambridge CB2 2QQ Tel: 01223 336598 http://www-bmu.psychiatry.cam.ac.uk/people/amm96/