Thanks Christian. I did not know if there was a stand alone GS decomposition routine in
FSL so I instead used matlab's orth function instead to find the orthonormal basis of my
signal and nuisance time series. It of course produced a completely different orthonormal
basis than GS. It uses singular value decomposition. I get an interesting set of results
with this but have no idea if they have any sort of meaning.
I know this isn't an FSL question per se, but do you have any idea if there is a statistical
interpretation to the time series produced by the SVD algorithm?
Thanks again,
Tal.
On Thu, 13 Aug 2009 21:34:24 +0100, Christian F. Beckmann
<[log in to unmask]> wrote:
>Hi
>
>It's identical, see the wiki description of GS, i.e. when using GS on
>two vectors v1 and v2, the new vectors are v1 and v2-proj(v1,v2)
>
>That is, you create the second vector by removing any correlation with
>the first.
>And yes, you don;t want to fully orthogonalise everything with
>everythinge else - in fact if you simply add the nuisance regressors
>then things will already be taken of.
>hth
>Christian
>
>
>
>
>
>On 13 Aug 2009, at 17:22, Tal Herbsman wrote:
>
>> A couple of of further questions, please forgive the ignorance/
>> confusion, it's been a while
>> since linear algebra:
>>
>> My understanding is that the GS de gives an orthonormal basis for
>> the function space. It's
>> not intuitively clear to me that this the same thing as covarying
>> nuisance signals out of a
>> time series.
>>
>> My application is not VBM whole brain correlation but rather a set
>> of n ROI based
>> timeseries cross-correlated with each other. Would GS decomposition
>> of [ROI1 ... ROIn
>> Nuisance1 ... NuisanceM] not ensure that I get no cross correlation
>> (as opposed to cross
>> correlating each ROI decomposed with only the nuisance variables?
>>
>> Tal.
>>
>> On Thu, 13 Aug 2009 10:25:22 +0100, Christian F. Beckmann
>> <[log in to unmask]> wrote:
>>
>>> Hi
>>>
>>> Yes, the description of the method is very inclusive. Only issue I
>>> have is that I'd advice using the Eigen-timeseries rather than the
>>> mean timeseries within the ROIs for further analysis
>>> hth
>>> Christian
>>>
>>>
>>>
>>> On 13 Aug 2009, at 04:02, Xue, Feng wrote:
>>>
>>>> Here is an good example for ROI time series data correlation in RS-
>>>> fMRI:
>>>>
>>>> "Mapping the functional connectivity of anterior cingulate cortex"
>>>> by Daniel S. Marguliesa, A.M. Clare Kellya, Lucina Q. Uddina, Bharat
>>>> B. Biswalb, F. Xavier Castellanosa, c, , and Michael P. Milham
>>>> NeuroImage
>>>> Volume 37, Issue 2, 15 August 2007, Pages 579-588
>>>>
>>>> On Thu, Aug 13, 2009 at 12:19 AM, Tal Herbsman<[log in to unmask]>
>>>> wrote:
>>>>> Dear FSL users, I would appreciate some advice on this: I am
>>>>> interested in doing
>>>>> correlation analysis on time series data extracted w/ fslmeants of
>>>>> various anatomical
>>>>> ROIs from FEAT preprocessed fMRI data. I have a set of nuisance
>>>>> time series (CSF, WM
>>>>> and whole brain global drifts). What would be the correct way of
>>>>> regressing these
>>>>> nuisance covariates out of my time series? I can think of two ways
>>>>> but am not sure which
>>>>> (if any) is correct, nor how to implement them:
>>>>>
>>>>> 1. orthogonalize the ROI time series with respect to global drift
>>>>> time series. Is the FSL
>>>>> orthogonalization available as a separate command line program?
>>>>>
>>>>> 2. use GLM to figure out PEs for nuisance regressors and subtract
>>>>> from signal. Not sure
>>>>> how to go about doing this.
>>>>>
>>>>> Any ideas?
>>>>>
>>>>> Tal.
>>>>>
>>>>>
>>>>> Tal Herbsman, MD
>>>>>
>>>>> Assistant Professor
>>>>> Department of Psychiatry
>>>>> University of Wisconsin School of Medicine and Public Health
>>>>> 6001 Research Park Blvd.
>>>>> Madison, WI 53719-1176
>>>>>
>>>>> Office: (608) 263-6100
>>>>> Fax: (608) 263-0265
>>>>> [log in to unmask]
>>>>>
>>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> Best Regards
>>>>
>>>>
>>>> National Key Laboratory of Cognitive Neuroscience and Learning
>>>> Beijing Normal University
>>>> Beijing, China. 100875
>>>> Tel: +86-13810154455
>>>> web: http://psychbrain.bnu.edu.cn
>>>> ==============================================
>>>> Welcome to MuDuo JinSheng BBS @ Beijing Normal University
>>>> telnet://bbs.mdjs.org
>>>> http://bbs.mdjs.org
>>>
>>> _______________________________________________
>>> Christian F. Beckmann, DPhil
>>> Senior Lecturer, Clinical Neuroscience Department
>>> Division of Neuroscience and Mental Health
>>> Imperial College London, Hammersmith Campus
>>> Rm 419, Burlington Danes Bldg, Du Cane Road, London W12 0NN, UK
>>> Tel.: +44 (0)20 7594 6685 --- Fax: +44 (0)20 7594 6548
>>> Email: [log in to unmask]
>>> http://www.imperial.ac.uk/medicine/people/c.beckmann/
>>>
>>> Senior Research Fellow, FMRIB Centre
>>> University of Oxford
>>> JR Hospital - Oxford OX3 9DU
>>> Tel.: +44 (0)1865 222551 --- Fax: +44 (0)1865 222717
>>> Email: [log in to unmask]
>>> http://www.fmrib.ox.ac.uk/~beckmann
>
>_______________________________________________
>Christian F. Beckmann, DPhil
>Senior Lecturer, Clinical Neuroscience Department
>Division of Neuroscience and Mental Health
>Imperial College London, Hammersmith Campus
>Rm 419, Burlington Danes Bldg, Du Cane Road, London W12 0NN, UK
>Tel.: +44 (0)20 7594 6685 --- Fax: +44 (0)20 7594 6548
>Email: [log in to unmask]
>http://www.imperial.ac.uk/medicine/people/c.beckmann/
>
>Senior Research Fellow, FMRIB Centre
>University of Oxford
>JR Hospital - Oxford OX3 9DU
>Tel.: +44 (0)1865 222551 --- Fax: +44 (0)1865 222717
>Email: [log in to unmask]
>http://www.fmrib.ox.ac.uk/~beckmann
|