Hi,
A quick thought - would ASL-fMRI not be an option to circumvent most of the
issues, especially the global scaling? If you are interested in such
'long-term' effects, ASL should be just fine...
Helmut
----- Original Message -----
From: "cyril pernet" <[log in to unmask]>
Sent: Wednesday, January 31, 2007 4:39 PM
Subject: Re: scan pauses
> Hi Markus
>> dear spmer.
>> a few weeks ago there was a question about a paradigm with scan pauses. i
>> could not find any reply to this (neither could i find something in the
>> archives). i have to ask almost the same question due to a paradigm,
>> which needs scan pauses (due to the noise of the scanner).
>> we explored the following paradigm:
>> 60 scans baseline - 5 min pause with sensoric testing- 108 scans
>> stimulation - 15 min pause with sensoric testing - 60 scans - 15 min
>> pause with sensoric testing - 60 scans.
>> we want to answer the BOLD-signal changes after the stimulation over a
>> tine period of 50 minutes.
>> now we have to face the problem of the interruption of the time series.
>> here my questions:
>>
> ok I'll try - no warranties ...
>> 1. is it absolutely impossible to conduct a paradigm with scanning pause
>> (and why)?
>>
> I don't think so ... :-\
>> 2. is there a time limit of the pause which would be tolerable?
>>
> well the time your subjects can stay in
>> 3. is there any method, how we can test, whether the pause is really a
>> problem or not (and not just theoretical)?
>>
> can help on this one
>> 4. can we do something to solve this problem (like correction for global
>> changes)?
>>
> yep - the global scaling should do the trick - or you can correct for
> globals using another method see e.g. NeuroImage 22 p360 the paper by
> Macey et al.
>> 5. if we have to regard the 4 blocks as individual sessions, is it
>> possible to use the first session (baseline) as the baseline condition
>> and contrast all the other sessions with it?
>> i would appreciate any commend about this very much (and even more, if it
>> is not to technical).
>>
> well once it is corrected for globals it should be ok. Note that a nice a
> way to look at the effect of learning is to use time as parametric
> regressor .. I don't remember exactly how but for sure you'll find details
> in paper from Henson et al.
>
> Hope this helps (a bit)
> cyril
>
|