John,
Thank you for your answer. Indeed, looking back at my data 2x seems
like a more accurate number, and your explanation meets my initial
guesses. Still, it's good to know that other researchers are getting
the same results.
Best,
Gus
On Aug 26, 2008, at 11:11 AM, John C. Mosher, Ph.D. wrote:
> Dear Gustavo,
>
> At Cleveland Clinic, regarding Question 2, we see a doubling (not
> 3x) when running MaxFilter for a simple reason: Our recordings are
> stored in 16 bit integer on our Elekta 306, and MaxFilter saves out
> its results (as do other programs) as 32 bit floats. We recently
> began running in 32 bit integer mode and do not see significant file
> changes after running MaxFilter.
>
> Indeed, our version of Graph does not handle the 32 bit integer, so
> we ran MaxFilter with -nosss simply to get a quick conversion to 32
> bit float, which Graph does handle. We hear the next version of
> Graph will directly handle 32 bit integer.
>
> We're still testing continuous HPI and related questions of head
> movement, so I'm curious to hear the reactions to your Question 1.
>
> -- John Mosher, Cleveland Clinic, [log in to unmask], [log in to unmask]
>
> Gustavo Sudre <[log in to unmask]> wrote: Thank you Burkhard and Danny.
>
> It's unfortunate that it cannot be done. Is there a reason for it,
> based on how the algorithms are implemented?
>
> Also, correct me if I am wrong, but the previous command I was using
> (maxfilter -f $file -frame head -origin 0 0 40 -trans default)
> realigned all files to the same origin as well. What would be the
> advantages of realigning it to the first file?
>
> Best,
>
> Gus
>
> On Aug 26, 2008, at 10:19 AM, Danny Mitchell wrote:
>
> > Hi Gustavo,
> >
> > I think the answer to question 1 is no: transformation of the data
> > requires SSS. The command you would want for aligning two sessions
> > would be something like:
> > maxfilter -f $session2.fif -trans session1.fif
> > (-nosss was designed for testing and is not recommended...)
> >
> > Not sure about the answer to question 2...
> >
> > Danny
> >
> > 2008/8/26 Gustavo Sudre
> > Hello,
> >
> > I have a couple questions about Maxfilter:
> >
> > 1) Is there a way to move my data from different sessions (same
> > subject) to the same coordinate system without using the SSS
> > algorithm? I saw the -nosss option in the manual, but it was unclear
> > on exactly what it does. I currently use something like maxfilter -f
> > $file -frame head -origin 0 0 40 -trans default , for running it
> > with SSS. The goal is to treat the data as if they came from only
> > one session, with the fewer modifications to it as possible.
> >
> > 2) After running Maxfilter (with SSS), my files increase size by a
> > factor of 3x. What kind of information is stored in the FIFF files
> > now that they are so big? Is it just a matter of the precision that
> > the data is now stored?
> >
> > Thanks,
> >
> > Gus
> >
>
|