 |
 |
Dear all, I've heard a suggestion that never models any event as 0 seconds, as this will make a lot of signals go to the error term in the GLM and reduce the effect size. Is that true? If I did present my stimuli for 1 second, should I model the events as 1s instead of 0s? Thanks for your opinions! Best wishes, Meikei ________________________________________ From: SPM (Statistical Parametric Mapping) [[log in to unmask]] On Behalf Of Weissman, Daniel [[log in to unmask]] Sent: Wednesday, August 17, 2011 2:15 AM To: [log in to unmask] Subject: Re: [SPM] event duration 0 Dear Chaleece, If you specify a duration of 0 seconds, the entire hemodynamic response to a single event will be used to model the BOLD signal in each trial- not just the first second or two. If you specify a 4-second duration, then the hemodynamic response convolved with a 4-second block will be used to model the BOLD signal in each trial. If you want to estimate the average time course of activation for each of your trial types, then you could use a FIR model (e.g., Miezin et al., 2000; Ollinger et al., 2001a,b - NeuroImage). Best and hope this helps! Daniel Daniel Weissman, Ph.D. Assistant Professor Department of Psychology University of Michigan Room 4052 1012 East Hall 530 Church Street Ann Arbor, MI 48109 ________________________________________ From: SPM (Statistical Parametric Mapping) [[log in to unmask]] on behalf of Chaleece Sandberg [[log in to unmask]] Sent: Tuesday, August 16, 2011 2:08 PM To: [log in to unmask] Subject: [SPM] event duration 0 In the model specification, what is considered an instantaneous event? What is setting the durations to 0 doing with the hemodynamic response curve- is it taking the whole potential curve into account or only the first instantaneous response and binning the time according to the dt, or something else? For context, we are running an E-Prime experiment with a picture stimulus displayed for 4 or 5 seconds, followed by a fixation for 2 or 4 seconds, and we want to know what is the most effective way to model the response curve for viewing the picture stimulus.