Dear Donald and Jason,
Thank you for your scripts! Indeed SPM filter causes plenty of distortions if the cutoff is close to the period of the protocol. Interestingly, the distortions are grater if rest is as long as active state, in comparison to brief on states, even if spaced by a long rest condition.
I used your scripts and some information found on the link to plot time course of the BOLD signal form a preprocessed functional scan:
http://psych290z.wikispaces.com/file/view/spm_detrendScriptKK.m
I took a peak of visual activation and overplayed unfiltered and 128 s spm filtered data. What seems interesting for me is that average (red graf.) is the same, though the variation is greater in unfiltered data.
Maybe my speculations are pretty naive, but can this mean that at the second level it does not matter, if we filter data or not?
Another naive question, if order to check this empirically: after the 4-d volume is filtered, how to same it into a set of img or nii so that header information is not lost.
Thank you so much for a nice chance to improve neuroimaging education.
Sincerely yours,
Vladimir
Volodymyr B. Bogdanov, PhD
> --- On Fri, 4/12/13, MCLAREN, Donald <[log in to unmask]>
> wrote:
>
> From: MCLAREN, Donald <[log in to unmask]>
> Subject: Re: [SPM] A question about the high-pass filter
> To: [log in to unmask]
> Date: Friday, April 12, 2013, 6:17
> PM
>
> Just to add to what Jason's post, I modified his code to
> loop through different models and give you the flexibility
> for different block sizes (on and off specified
> separately).
>
> cc=filteringfun2(on,of,HP,TR,jitter,cycles); cc
>
> cc=filteringfun2(40,40,128,2,0,20); cc
> *Baseline varies from 0 to length of ON block at the
> beginning of each run. The variation in CC is caused by
> varying the baseline period at the beginning in TR
> increments.
>
>
> What becomes clear from trying different possibilities is
> that when the filter is 2* period, the correlation between
> signal and filtered signal is >.99. When the filter is
> just below the period (e.g. 60,60 and 128s filter), the
> correlation drops and drops rapidly as the period approaches
> the filter.
>
>
> You can also see odd signal patterns after filtering.
>
> Hopefully everyone has found this discussion as useful as I
> have found it.
>
>
>
> 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 Fri, Apr 12, 2013 at 9:22 AM, Torben Lund <[log in to unmask]>
> wrote:
>
> Hi Helmut (and others)
>
>
>
> Actually all the maps in my PDF were unfiltered (cut off
> period =inf). The message is that with reasonable fast
> paradigms the HP filter is not that important, but for slow
> paradigms you need it as a guard against false positives
> (due do scanner drift looking like the paradigm). If the
> maps had been filtered (cut off period=128s) the first 7
> would be almost unchanged, but number 8 the one with 120s on
> and 120s off would appear completely blank.
>
>
>
>
> Best
>
> Torben
>
>
>
>
>
>
>
>
>
>
>
>
>
> On Apr 12, 2013, at 12:39 PM, H. Nebl <[log in to unmask]>
> wrote:
>
>
>
> > Finally a question about Torben's example (thanks for
> sharing!): You varied the block length starting with 8 s up
> to 120 s, but did you always use 128 s as a HPF or did you
> adjust it? Otherwise it is not clear what you show ;-)
|