Back to PLS HelpER-fMRI time window
Posted on 02/02/07 12:42:55
Number of posts: 100
Author: Agatha (---.csbmb.Princeton.EDU)
Date: 01-22-07 10:47
hello again ~
i have two questions:
1. is there any way (manually or via gui) that the time window for analysis can be modified AFTER creation of the datamat, much like is done for conditions? (i.e., i have not found any easily solution apart from re-creating the datamats - but this is very timely) i realize that the ref scan can not be changed, but one might, for instance, want to analyze a shorter interval vs a longer interval if there is some overlap between events (and thus HRF responses)
2. statistically and theoretically - do you know of the effects of time window selection on the results? i.e., how much weight does each TR in the time window have on the final result?
thanks very much for any insight you might have on this ~
agatha
Date: 01-22-07 10:47
hello again ~
i have two questions:
1. is there any way (manually or via gui) that the time window for analysis can be modified AFTER creation of the datamat, much like is done for conditions? (i.e., i have not found any easily solution apart from re-creating the datamats - but this is very timely) i realize that the ref scan can not be changed, but one might, for instance, want to analyze a shorter interval vs a longer interval if there is some overlap between events (and thus HRF responses)
2. statistically and theoretically - do you know of the effects of time window selection on the results? i.e., how much weight does each TR in the time window have on the final result?
thanks very much for any insight you might have on this ~
agatha
Replies:
Re: ER-fMRI time window
Posted on 02/02/07 12:43:12
Number of posts: 100
Author: Randy (---.rotman-baycrest.on.ca)
Date: 01-22-07 12:12
Agatha wrote:
> 1. is there any way (manually or via gui) that the time window
> for analysis can be modified AFTER creation of the datamat,
> much like is done for conditions? (i.e., i have not found any
> easily solution apart from re-creating the datamats - but this
> is very timely) i realize that the ref scan can not be changed,
> but one might, for instance, want to analyze a shorter interval
> vs a longer interval if there is some overlap between events
> (and thus HRF responses)
currently there is no easy way to do this without recreating the datamats. Since the datamats are flattened 4D data, its not an easy task to edit the matlab files directly. Probably the easiest thing to do is use the batch_plsgui procedure to shorten the pointing & clicking to redo datamats. Examples of the batch files are included in the PLSGUI zip files.
> 2. statistically and theoretically - do you know of the effects
> of time window selection on the results? i.e., how much weight
> does each TR in the time window have on the final result?
TRs each equal weighting, but if there is a strong effect at a latter TR that is eliminated with a shorter analysis window, there is a chance for a change in the outcome of the results. In most "well-behaved" data sets, the fMRI signal changes extend across 2 or more TRs, so the window length has less of an effect.
Date: 01-22-07 12:12
Agatha wrote:
> 1. is there any way (manually or via gui) that the time window
> for analysis can be modified AFTER creation of the datamat,
> much like is done for conditions? (i.e., i have not found any
> easily solution apart from re-creating the datamats - but this
> is very timely) i realize that the ref scan can not be changed,
> but one might, for instance, want to analyze a shorter interval
> vs a longer interval if there is some overlap between events
> (and thus HRF responses)
currently there is no easy way to do this without recreating the datamats. Since the datamats are flattened 4D data, its not an easy task to edit the matlab files directly. Probably the easiest thing to do is use the batch_plsgui procedure to shorten the pointing & clicking to redo datamats. Examples of the batch files are included in the PLSGUI zip files.
> 2. statistically and theoretically - do you know of the effects
> of time window selection on the results? i.e., how much weight
> does each TR in the time window have on the final result?
TRs each equal weighting, but if there is a strong effect at a latter TR that is eliminated with a shorter analysis window, there is a chance for a change in the outcome of the results. In most "well-behaved" data sets, the fMRI signal changes extend across 2 or more TRs, so the window length has less of an effect.
Login to reply to this topic.
