thank you for your helpful reply!
When would the prime and target be presented to close in time? In my experiment the onset of the target word is 1 second apart from the offset of the priming word.
Von: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] Im Auftrag von Weissman, Daniel
Gesendet: Freitag, 18. März 2011 14:38
An: [log in to unmask]
Betreff: Re: [SPM] Parametric Modulation: 2 questions
You can certainly incorporate RT as a parametric modulator for the target word regressor in your model.
This "RT regressor" will capture variance in the BOLD signal that varies linearly with RT, which is not explained by the average BOLD response to the target word (or the combined response to the prime and target if these cannot be separated because they are presented very close in time with no jitter). If the prime and target are presented close in time, they may produce a combined BOLD response that deviates from the canonical shape. In that case, you may wish to consider using a finite impulse response model (e.g., Weissman et al., 2006)
The polynomial expansion that is most appropriate for your study depends on the scientific hypothesis being investigated. That said, in a number of event-related fMRI studies (e.g., Chee et al., 2008; Prado, Carp, & Weissman, 2011; Prado & Weissman, 2011), we have found that the BOLD signal varies in a predominantly linear fashion with RT. However, there are occasional quadratic effects as well (e.g., Chee et al., 2008). So, you might focus on the linear effect, but check to see whether higher-order effects are present as well.
Best and hope this helps!
Daniel Weissman, Ph.D.
Department of Psychology
University of Michigan
1012 East Hall
530 Church Street
Ann Arbor, MI 48109
From: SPM (Statistical Parametric Mapping) [[log in to unmask]] On Behalf Of Christian Baeuchl [[log in to unmask]]
Sent: Friday, March 18, 2011 5:29 AM
To: [log in to unmask]
Subject: [SPM] Parametric Modulation: 2 questions
I repost two questions which I posted 2 days ago which haven't received any responses.
I know these are simple questions but I would really appreciate any kind of help:
Dear SPM experts,
I have two questions regarding the usage of a parametric modulator in a semantic priming experiment:
1. Does it make sense to use reaction time as a parametric modulator of a target word which follows a priming word and of which subjects have to decide as quickly as possible if the target word is related to the priming word or not?
In the book on SPM by Karl Friston et al. I read in chapter 15:
..."So the choice of model depends on the assumptions about the duration of neuronal activity in the particular region of interest. If this is unknown, trials whose durations vary over a few seconds (as with typical RTs) are probably best modelled with two regressors: one modelling events, and a second modelling a parametric modulation of the response, by the RT on each trial."
In my case the length of the events (presentation of target word) is always the same and there-fore does not depend upon the RTs. Does it still make sense to use RT as a parametric modu-lator of the target word?
2. If it does make sense than the second question is: Which order of polynomial expansion should I use? I would've guessed that RT's require a second order polynomial expansion but I'm quite unsure about this!
I would appreciate any kind of help!