 |
 |
Hi Todd, > Why not just use a Finite Impulse Response (FIR) model and let the data tell you what the shape of the hrf is instead of including temporal derivatives to pick up variance missed by the traditional hrf shape? I would say because 2nd level modelling is quite difficult with FIR - you still have to use priors either making up a time course using weights at the 1st level, or picking up specifics bins at the second level. Derivatives capture quite well a range of hdr and can be combined easily (using the tool mentioned earlier) to have a descent estimate of the response amplitude per condition and subject... at least that's the way I see it .. but FIR is better as exploratory tool for sure cyril Hi Joelle and Cyril, Why not just use a Finite Impulse Response (FIR) model and let the data tell you what the shape of the hrf is instead of including temporal derivatives to pick up variance missed by the traditional hrf shape? Best, Todd On Tue, Sep 22, 2015 at 2:38 AM, cyril pernet <[log in to unmask]<mailto:[log in to unmask]>> wrote: Joelle, it is likely that there are many factors that can influence the the timing and shape of the hrf - difference in the vasculature across regions, but also people (think aging) - difference in the neuro-vascular coupling across regions - difference in the neural response most of the time people just use the hrf because it is always hard to use properly derivatives at the group level to make inference, but I think it is worthwhile .. cyril Thanks Cyril, would you say that it's more standard to do it this way or the default (without time and dispersion)? I understand that the canonical HRF response function was developed based on the visual areas, so that maybe if looking at other areas should allow for some changes in the function? On Mon, Sep 21, 2015 at 3:56 PM, PERNET Cyril <[log in to unmask]<mailto:[log in to unmask]>> wrote: that's right Joelle, temporal derivatives weights allows earlier/later fit whilst dispersion derivative allows a bit of change in the shape (how wide) of the hrf - the model is hrf + 1st derivative + 2nd derivative and the combination of the 3 regressors gives you the modeled data more info on how this works here http://journal.frontiersin.org/article/10.3389/fnins.2014.00001/full Cyril [http://www.frontiersin.org/files/MyHome+Article+Library/58014/58014_Thumb_60.jpg]<http://journal.frontiersin.org/article/10.3389/fnins.2014.00001/full> Frontiers | Misconceptions in the use of the General ... Misconceptions in the use of the General Linear Model applied to functional MRI: a tutorial for junior neuro-imagers Read more...<http://journal.frontiersin.org/article/10.3389/fnins.2014.00001/full> ________________________________ From: SPM (Statistical Parametric Mapping) <<mailto:[log in to unmask]>[log in to unmask]<mailto:[log in to unmask]>> on behalf of Joelle Zimmermann <<mailto:[log in to unmask]>[log in to unmask]<mailto:[log in to unmask]>> Sent: 21 September 2015 20:41 To: [log in to unmask]<mailto:[log in to unmask]> Subject: [SPM] time and dispersion derivatives - Canonical HRF's Hi SPM'ers, Setting up a first-level model, Im faced with the decision if to include time and dispersion parameters under the Canonical HRF model. Based off of the help, as well as the manual, I understand that these parameters give the HRF function some leeway in it's 'preciseness' - ie that it can happen a few second earlier or later and has a larger spread (dispersion). Is this correct? Thanks, Joelle The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. -- Dr Cyril Pernet, Centre for Clinical Brain Sciences (CCBS) Neuroimaging Sciences The University of Edinburgh Chancellor's Building Room GU426D 49 Little France Crescent Edinburgh EH16 4SB [log in to unmask]<mailto:[log in to unmask]> tel: 0131 465 9530 http://www.sbirc.ed.ac.uk/cyril http://www.ed.ac.uk/edinburgh-imaging The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. -- Todd S. Woodward, Ph.D. Cognitive Neuroscience of Schizophrenia Lab (CNoS) http://www.bcmhari.ca/faculty/woodward-todd.php http://www.cnoslab.com/donate.html 604-875-2000 x4724 ***** Associate Professor Department of Psychiatry University of British Columbia BC Mental Health and Addiction Research Institute Vancouver, Canada