Re: [Freesurfer] precoloring - adjusted errors

2012-06-15 Thread Douglas N Greve
It won't be so simple. If you can represent your filter as a matrix, then you can set the flac.TFmtx, but you'll need to do this in multiple places. You can do a search for TFmtx in the matlab code to see where On 06/15/2012 03:23 PM, Caspar M. Schwiedrzik wrote: > if i come up with the code for

Re: [Freesurfer] precoloring - adjusted errors

2012-06-15 Thread Caspar M. Schwiedrzik
if i come up with the code for the newey white standard error, where in selxavg3-sess would i need to plug that in? caspar 2012/6/15 Douglas N Greve > I don't know anything about it. Looks interesting. > doug > > > On 06/15/2012 02:14 PM, Caspar M. Schwiedrzik wrote: > >> Hi Doug, >> if I filte

Re: [Freesurfer] precoloring - adjusted errors

2012-06-15 Thread Douglas N Greve
I don't know anything about it. Looks interesting. doug On 06/15/2012 02:14 PM, Caspar M. Schwiedrzik wrote: > Hi Doug, > if I filter, an alternative might be the Newey-West estimator > (http://en.wikipedia.org/wiki/Newey-West_estimator), I guess? > Caspar > > > 2012/6/15 Douglas N Greve

Re: [Freesurfer] precoloring - adjusted errors

2012-06-15 Thread Caspar M. Schwiedrzik
Hi Doug, if I filter, an alternative might be the Newey-West estimator ( http://en.wikipedia.org/wiki/Newey-West_estimator), I guess? Caspar 2012/6/15 Douglas N Greve > Hi Caspar, it does not make sense to low pass filter and prewhiten. > Unfortunately, I don't have the precoloring implemented

Re: [Freesurfer] precoloring - adjusted errors

2012-06-15 Thread Douglas N Greve
Hi Caspar, it does not make sense to low pass filter and prewhiten. Unfortunately, I don't have the precoloring implemented in FSFAST. doug On 06/15/2012 01:36 PM, Caspar M. Schwiedrzik wrote: > Hi! > I am processing a resting state data set, which I would like to low > pass filter to get rid of

[Freesurfer] precoloring - adjusted errors

2012-06-15 Thread Caspar M. Schwiedrzik
Hi! I am processing a resting state data set, which I would like to low pass filter to get rid of some artifacts. The sample will be very small, so I won't run a RFX GLM, but an FFX GLM and/or single subject analyses. Since the low pass filter introduces high autocorrelations, it does not make sens