 |
 |
Hi David,
Adjusting a timecourse refers to regressing out the effects of any columns of no interest from your timeseries. You do this by creating an F-contrast with an identity matrix over the columns of interest. Anything not in that contrast will be removed. SPM enables you to extract an ROI per session, so you can’t adjust the timecourse for multiple sessions. You have a choice of running separate DCMs on each session, or concatenating your onsets to create a single session, prior to ROI extraction and DCM.
Best,
Peter
From: David Hofmann [mailto:[log in to unmask]]
Sent: 29 July 2014 21:42
To: Zeidman, Peter; rogiedodgie
I'm afraid there is another thing I don't understand.
How to adjust the timecourse for multiple sessions per subject? What does this option mean?
In this case every subject has 3 sessions.
2014-07-29 18:05 GMT+02:00 David Hofmann <[log in to unmask]>:
Thanks guys, worked fine!
greetings
David
2014-07-29 17:52 GMT+02:00 Zeidman, Peter <[log in to unmask]>:
Hi David,
A couple of options. One is to set p=1 at the single subject level. Then you’ll get all activity in the region. The other alternative is to look for the nearest peak (e.g. p < 0.001) in each subject within some radius of the group peak.
Best,
Peter.
From: David Hofmann [mailto:[log in to unmask]]
Sent: 29 July 2014 15:40
To: Zeidman, Peter
Cc: [log in to unmask]
Subject: Re: [SPM] DCM: Group comparison & defining ROIs
Thanks Peter,
what would be a liberal threshold in this case? Since there may be no activation in the ROI I chose at group level in a subject I can't extract the timecourse via the Eigenvariate-button. Does this mean I have to choose a significance level where there is activation everywhere?
Or is there another way to extract timecourse data by means of coordinates?
greetings
David
2014-07-29 15:41 GMT+02:00 Zeidman, Peter <[log in to unmask]>:
Hi David,
That’s fine – not every subject needs individual level activation (although this could indicate different subjects are using different cognitive processes?).
A note of caution in case you’re using the GUI to extract ROIs… When you click the ‘eigenvariate’ button, the crosshairs will jump to the nearest activation. To prevent this you’d have to set the significance threshold to a very liberal value.
Best,
Peter.
From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] On Behalf Of David Hofmann
Sent: 29 July 2014 13:41
To: [log in to unmask]
Subject: [SPM] DCM: Group comparison & defining ROIs
Hi @all,
when defining ROIs for a DCM group analysis, is it necessary that in every subject those ROIs need to be significantly activated (e.g. with F-test over all effects of interest at p = 0.001 uncorrected)?
With 2nd level analysis I can define some ROIs with significant activation, but those are not active in every single person.
greetings
David