 |
 |
No. I would only put the non-neural (e.g. motion, noise) components in the null space. Do the extraction once, then work on separating it into the two task components. Doing this twice will not get you orthogonal time courses because both extractions would share the residual variance of the model. The point of adjustment is to remove the non-neural/neurovascular coupling sources of variance from the signal. Trying to use adjustment to remove the effect of one task won't work because the effect of the task is not 100% represented by the HRF, there is variation between trials as well as within trials that the HRF doesn't capture. If I were doing SEM, I'd take a look at beta-series correlation, gPPI, and DCM to get an idea of how to build an appropriate model and how to separate the two tasks properly. Best Regards, Donald McLaren ================= D.G. McLaren, Ph.D. Research Fellow, Department of Neurology, Massachusetts General Hospital and Harvard Medical School Postdoctoral Research Fellow, GRECC, Bedford VA Website: http://www.martinos.org/~mclaren Office: (773) 406-2464 ===================== This e-mail contains CONFIDENTIAL INFORMATION which may contain PROTECTED HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and which is intended only for the use of the individual or entity named above. If the reader of the e-mail is not the intended recipient or the employee or agent responsible for delivering it to the intended recipient, you are hereby notified that you are in possession of confidential and privileged information. Any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited and may be unlawful. If you have received this e-mail unintentionally, please immediately notify the sender via telephone at (773) 406-2464 or email. On Mon, Oct 28, 2013 at 9:19 AM, Sevel,Landrew S <[log in to unmask]>wrote: > Donald, > > Thanks for your response. From this then, would it be sensible to say > that you are only interested in the effects of one condition on the > regional response and therefore place variance associated with the other > condition in the null space? > > Would this work as a way to sort of orthogonalize the extracted time > course? > > -Drew > > From: <MCLAREN>, Donald <[log in to unmask]> > Date: Friday, October 25, 2013 4:47 PM > To: Landrew Sevel <[log in to unmask]> > Cc: SPM <[log in to unmask]> > Subject: Re: [SPM] Eigenvariate--Adjusting for a contrast > > Drew, > > The point of adjusting is to remove the effect of the null-space effects > from the data (e.g the effects of motion and constant term). If you are > going to adjust, you want to use the same adjustment for all conditions. > You want to use an F-contrast that tests for the effect of any condition as > the contrast to adjust for. > > Best Regards, Donald McLaren > ================= > D.G. McLaren, Ph.D. > Research Fellow, Department of Neurology, Massachusetts General Hospital > and > Harvard Medical School > Postdoctoral Research Fellow, GRECC, Bedford VA > Website: http://www.martinos.org/~mclaren > Office: (773) 406-2464 > ===================== > This e-mail contains CONFIDENTIAL INFORMATION which may contain PROTECTED > HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and which is > intended only for the use of the individual or entity named above. If the > reader of the e-mail is not the intended recipient or the employee or agent > responsible for delivering it to the intended recipient, you are hereby > notified that you are in possession of confidential and privileged > information. Any unauthorized use, disclosure, copying or the taking of any > action in reliance on the contents of this information is strictly > prohibited and may be unlawful. If you have received this e-mail > unintentionally, please immediately notify the sender via telephone at > (773) > 406-2464 or email. > > > On Thu, Oct 17, 2013 at 6:31 PM, Sevel,Landrew S <[log in to unmask]>wrote: > >> SPM'ers, >> >> What are the specific effects of adjusting for an F contrast on the >> eigenvariate values that are produced? >> >> Specifically, we are looking to apply SEM to time series from a >> protocol with two conditions with the goal of comparing models generated >> between conditions. Would it be appropriate to create an F contrast for >> condition 1 and extract a time series from region 1 after adjusting for >> contrast one and then extracting an eigenvariate from the same region but >> adjusting for an F contrast of condition two? >> >> I do realize the potential utility of using DCM to address this >> question but nonetheless would like to know if this would be an appropriate >> approach for SEM. >> >> Many thanks, >> >> Drew >> > >