Thank you again for all this help!
Another question:
in the AROMA paper, in the regression after ICA-aroma a linear trend was
included among regressors. How can I do it? Is it the same if I use the
AFNI command 3dDetrend?
thank you
Angela
> Hi,
> also 6mm FWHM is quite a bit of smoothing.
> I would look at less smoothed data.
> Cheers,
> Andreas
>
> Von: Maarten Mennes <[log in to unmask]>
> Antworten an: FSL - FMRIB's Software Library <[log in to unmask]>
> Datum: Dienstag, 19. Mai 2015 13:39
> An: <[log in to unmask]>
> Betreff: Re: [FSL] ICA AROMA help
>
> Dear Angela,
>
> for that one subject with 3 components I would run Melodic before
> registering to the standard image, then inspect the report to see what the
> components look like. In addition I would take a good look at the raw data
> to see if there was anything unusual going on... I agree that 3-8
> components
> would not be typical...
>
> Maarten
>
> On Mon, May 18, 2015 at 11:14 PM, Angela Favaro <[log in to unmask]>
> wrote:
>> Thank you for this. This solved my problem.
>>
>> I have another general question on ICA AROMa (but also generally on
>> ICA).
>> I am analyzing a dataset of RS fMRI and comparing different methods of
>> motion denoising. However, I noticed that for some subjects ICA
>> processed
>> only a limited number of components (7 or 8 instead of more than 30-40
>> that I usually observe). In one subject the components were 3 and all 3
>> were motion components (mean FD was 0.05mm).
>> How must I interpret this? Is it an artifact? or did I make something
>> wrong in the preprocessing? my preprocessing consisted in: dropping
>> first
>> 5 volumes, deobliquing (3drefit), motion correction to average of
>> timeseries, skull stripping, registering to standard image (flirt),
>> smoothing (FWHM=6).
>>
>> THank you for any help or advice
>>
>> Angela
>>
>>> > Hi Angela,
>>> >
>>> > You need to change your $PATH variable so that the Anaconda version
>>> of
>>> > Python comes before the Framework version of Python. You can do this
>>> by
>>> > opening the file '/Users/angelafavaro/.bash_profile' in a text
>>> editor, and
>>> > pasting this line at the bottom of that file:
>>> >
>>> > export PATH=/Users/angelafavaro/minconda/bin:$PATH
>>> >
>>> > Alternately, in that file there is probably a line near the bottom
>>> which
>>> > adds the Framework python installation to the PATH -
>>> removing/commenting
>>> > out this line will have the same effect.
>>> >
>>> > Cheers,
>>> >
>>> > Paul
>>> >
>>> > On 15 May 2015 at 15:39, Angela Favaro <[log in to unmask]>
>>> wrote:
>>> >
>>>> >> Hi Paul,
>>>> >> thank you for your help! I ma not very expert in using python...
>>>> >> this is the output of what you requested:
>>>> >>
>>>> >> bash-3.2$ which python
>>>> >> /Library/Frameworks/Python.framework/Versions/2.7/bin/python
>>>> >> bash-3.2$ which pip
>>>> >> /Users/angelafavaro/miniconda/bin/pip
>>>> >> bash-3.2$ which conda
>>>> >> /Users/angelafavaro/miniconda/bin/conda
>>>> >> bash-3.2$ echo $PATH
>>>> >>
>>>> >>
>>>> /Applications/caret/bin_macosx:/sw/bin:/Library/Frameworks/Python.framework/
>>>> Versions/2.7/bin:/Users/angelafavaro/miniconda/bin:/sw/bin:/sw/sbin:/usr/loc
>>>> al/fsl/bin:/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin:/usr/X11/bin:/usr/X1
>>>> 1R6/bin:/usr/local/bic/bin:/Users/angelafavaro/abin
>>>> >> bash-3.2$
>>>> >>
>>>> >> thank you for any help
>>>> >> BW
>>>> >> Angela
>>>> >>
>>>> >>
>>>> >>
>>>> >>
>>>>> >> > Hi Angela,
>>>>> >> >
>>>>> >> > It is likely that you have python installed on your computer
>>>>> multiple
>>>>> >> > times, and the version which has numpy installed is not the
>>>>> version
>>>> >> that
>>>>> >> > is
>>>>> >> > being used by AROMA.
>>>>> >> >
>>>>> >> > Could you please provide us with the output of the following
>>>>> commands:
>>>>> >> >
>>>>> >> > which python
>>>>> >> > which pip
>>>>> >> > which conda
>>>>> >> > echo $PATH
>>>>> >> >
>>>>> >> > Thanks,
>>>>> >> >
>>>>> >> > Paul
>>>>> >> >
>>>>> >> > On 14 May 2015 at 21:13, Angela Favaro <[log in to unmask]>
>>>>> wrote:
>>>>> >> >
>>>>>> >> >> Hi Christian,
>>>>>> >> >> numpy was already installed using the command (conda install
>>>>>> numpy)
>>>>>> >> >>
>>>>>> >> >> bash-3.2$ pip install numpy
>>>>>> >> >> Requirement already satisfied (use --upgrade to upgrade):
>>>>>> numpy in
>>>>>> >> >> /Users/angelafavaro/miniconda/lib/python2.7/site-packages
>>>>>> >> >>
>>>>>> >> >> however, again ...
>>>>>> >> >>
>>>>>> >> >> --------------- 'ICA-based Automatic Removal Of Motion
>>>>>> Artifacts'
>>>>>> >> >> ---------------
>>>>>> >> >>
>>>>>> >> >> Step 1) MELODIC
>>>>>> >> >> Step 2) Automatic classification of the components
>>>>>> >> >> - registering the spatial maps to MNI
>>>>>> >> >> - extracting the CSF & Edge fraction features
>>>>>> >> >> Traceback (most recent call last):
>>>>>> >> >> File "ICA_AROMA.py", line 196, in <module>
>>>>>> >> >> edgeFract, csfFract = aromafunc.feature_spatial(fslDir,
>>>>>> outDir,
>>>>>> >> >> scriptDir, melIC_MNI)
>>>>>> >> >> File
>>>>>> >> >>
>>>>>> >> >>
>>>> >>
>>>>
> "/Users/angelafavaro/Desktop/risonanze/network/aroma/ICA_AROMA_functions.py",
>>>>>> >> >> line 334, in feature_spatial
>>>>>> >> >> import numpy as np
>>>>>> >> >> ImportError: No module named numpy
>>>>>> >> >>
>>>>>> >> >>
>>>>>> >> >> What can be the problem?
>>>>>> >> >> thank you
>>>>>> >> >> Angela
>>>>>> >> >>
>>>>>> >> >>
>>>>>> >> >>
>>>>>>> >> >> > Hi
>>>>>>> >> >> >
>>>>>>> >> >> > Looks like numpy isnĀ¹t installed on your system
>>>>>>> >> >> >
>>>>>>> >> >> > pip install numpy
>>>>>>> >> >> >
>>>>>>> >> >> > hth
>>>>>>> >> >> > Christian
>>>>>>> >> >> >
>>>>>>> >> >> >
>>>>>>>> >> >> >> On 13 May 2015, at 17:27, Angela Favaro
>>>>>>>> <[log in to unmask]>
>>>>>> >> >> wrote:
>>>>>>>> >> >> >>
>>>>>>>> >> >> >> Hi Maarten,
>>>>>>>> >> >> >> python 2.7
>>>>>>>> >> >> >>
>>>>>>>> >> >> >> thank you
>>>>>>>> >> >> >> Angela
>>>>>>>> >> >> >>
>>>>>>>> >> >> >>
>>>>>>>>> >> >> >>> Dear Angela,
>>>>>>>>> >> >> >>>
>>>>>>>>> >> >> >>> can you check whether you are using the correct python
> version
>>>>>>>>> >> >> >>>
>>>>>>>>> >> >> >>> run
>>>>>>>>> >> >> >>> which python
>>>>>>>>> >> >> >>> at the command line
>>>>>>>>> >> >> >>>
>>>>>>>>> >> >> >>> I also have the anaconda distribution on OSX 10.10,
>>>>>>>>> works
> like a
>>>>>> >> >> charm
>>>>>>>>> >> >> >>>
>>>>>>>>> >> >> >>> Maarten
>>>>>>>>> >> >> >>>
>>>>>>>>> >> >> >>> On Wed, May 13, 2015 at 12:15 AM, Angela Favaro
>>>>>>>>> >> >> >>> <[log in to unmask]>
>>>>>>>>> >> >> >>> wrote:
>>>>>>>>> >> >> >>>
>>>>>>>>>> >> >> >>>> Hi all,
>>>>>>>>>> >> >> >>>> I am not very expert in using python. The script of
> ICA-AROMA
>>>> >> give
>>>>>> >> >> me
>>>>>>>>>> >> >> >>>> this
>>>>>>>>>> >> >> >>>> error:
>>>>>>>>>> >> >> >>>>
>>>>>>>>>> >> >> >>>> --------------- 'ICA-based Automatic Removal Of
>>>>>>>>>> Motion
>>>> >> Artifacts'
>>>>>>>>>> >> >> >>>> ----------
>>>>>>>>>> >> >> >>>>
>>>>>>>>>> >> >> >>>> Step 1) MELODIC
>>>>>>>>>> >> >> >>>> Step 2) Automatic classification of the components
>>>>>>>>>> >> >> >>>> - registering the spatial maps to MNI
>>>>>>>>>> >> >> >>>> - extracting the CSF & Edge fraction features
>>>>>>>>>> >> >> >>>> Traceback (most recent call last):
>>>>>>>>>> >> >> >>>> File "ICA_AROMA.py", line 196, in <module>
>>>>>>>>>> >> >> >>>> edgeFract, csfFract =
>>>>>>>>>> aromafunc.feature_spatial(fslDir,
>>>> >> outDir,
>>>>>>>>>> >> >> >>>> scriptDir, melIC_MNI)
>>>>>>>>>> >> >> >>>> File "ICA_AROMA_functions.py", line 334, in
>>>>>>>>>> feature_spatial
>>>>>>>>>> >> >> >>>> import numpy as np
>>>>>>>>>> >> >> >>>> ImportError: No module named numpy
>>>>>>>>>> >> >> >>>> bash-3.2$
>>>>>>>>>> >> >> >>>>
>>>>>>>>>> >> >> >>>>
>>>>>>>>>> >> >> >>>> I had installed the numpy module (by using conda
>>>>>>>>>> install
> numpy).
>>>>>>>>>> >> >> >>>> I am using a OSX 10.6.8
>>>>>>>>>> >> >> >>>>
>>>>>>>>>> >> >> >>>> Can you help me in understanding the problem?
>>>>>>>>>> >> >> >>>>
>>>>>>>>>> >> >> >>>> thank you
>>>>>>>>>> >> >> >>>>
>>>>>>>>>> >> >> >>>> Angela
>>>>>>>>>> >> >> >>>>
>>>>>>>>>> >> >> >>>>
>>>>>>>>>> >> >> >>>>
>>>>>>>>>>> >> >> >>>>> Dear Angela,
>>>>>>>>>>> >> >> >>>>>
>>>>>>>>>>> >> >> >>>>> We recommend applying ICA-AROMA after spatial
>>>>>>>>>>> smoothing,
> this
>>>> >> is
>>>>>> >> >> in
>>>>>>>>>> >> >> >>>> order
>>>>>>>>>>> >> >> >>>>> to 'condition' the subsequent ICA run. It will
>>>>>>>>>>> prevent
> the
>>>>>> >> >> resulting
>>>>>>>>>>> >> >> >>>>> components from fragmenting too much. We have
>>>>>>>>>>> extensively
>>>> >> tested
>>>>>> >> >> the
>>>>>>>>>>> >> >> >>>>> suggested pipeline and the ability to detect the
>>>>>>>>>>> motion-related
>>>>>>>>>> >> >> >>>> artifacts
>>>>>>>>>>> >> >> >>>>> is hampered when leaving out the smoothing step.
>>>>>>>>>>> >> >> >>>>>
>>>>>>>>>>> >> >> >>>>> If you would like to continue using unsmooth data
>>>>>>>>>>> for
> your
>>>>>> >> >> analyses,
>>>>>>>>>> >> >> >>>> we
>>>>>>>>>>> >> >> >>>>> suggest running ICA-AROMA on the smoothed data,
>>>>>>>>>>> then
> applying
>>>>>>>>>> >> >> >>>> fsl_regfilt
>>>>>>>>>>> >> >> >>>>> with the identified components on the unsmooth data
> (AROMA will
>>>>>>>>>>> >> >> >>>>> return
>>>>>>>>>> >> >> >>>> a
>>>>>>>>>>> >> >> >>>>> list of identified components).
>>>>>>>>>>> >> >> >>>>>
>>>>>>>>>>> >> >> >>>>> hope this helps,
>>>>>>>>>>> >> >> >>>>> Maarten
>>>>>>>>>>> >> >> >>>>>
>>>>>>>>>>> >> >> >>>>> On Mon, May 11, 2015 at 9:40 PM, Angela Favaro
>>>>>>>>>> >> >> >>>> <[log in to unmask]>
>>>>>>>>>>> >> >> >>>>> wrote:
>>>>>>>>>>> >> >> >>>>>
>>>>>>>>>>>> >> >> >>>>>> Hi FSL expert,
>>>>>>>>>>>> >> >> >>>>>> I would like to test the effects of ICA AROMA on
>>>>>>>>>>>> my
> dataset
>>>>>> >> >> (RSFC).
>>>>>>>>>>>> >> >> >>>>>> In the website it is suggested to use the script
>>>>>>>>>>>> after
> spatial
>>>>>>>>>>>> >> >> >>>>>> smoothing.
>>>>>>>>>>>> >> >> >>>>>> Is smoothing a needed step to use ICA AROMA? in
>>>>>>>>>>>> some
>>>> >> particular
>>>>>>>>>> >> >> >>>> cases,
>>>>>>>>>>>> >> >> >>>>>> smoothing can create spurious correlations and I
>>>>>>>>>>>> want
> to know
>>>> >> if
>>>>>> >> >> the
>>>>>>>>>>>> >> >> >>>>>> ability of the method to denoise data is impaired
>>>>>>>>>>>> by
> omitting
>>>>>>>>>>>> >> >> >>>>>> spatial
>>>>>>>>>>>> >> >> >>>>>> smoothing of data.
>>>>>>>>>>>> >> >> >>>>>>
>>>>>>>>>>>> >> >> >>>>>> thank you
>>>>>>>>>>>> >> >> >>>>>>
>>>>>>>>>>>> >> >> >>>>>> Angela
>>>>>>>>>>>> >> >> >>>>>>
>>>>>>>>>>> >> >> >>>>>
>>>>>>>>>>> >> >> >>>>>
>>>>>>>>>>> >> >> >>>>>
>>>>>>>>>>> >> >> >>>>> --
>>>>>>>>>>> >> >> >>>>> Maarten Mennes, Ph.D.
>>>>>>>>>>> >> >> >>>>> Senior Researcher
>>>>>>>>>>> >> >> >>>>> Donders Institute for Brain, Cognition and
>>>>>>>>>>> Behaviour
>>>>>>>>>>> >> >> >>>>> Radboud University Nijmegen
>>>>>>>>>>> >> >> >>>>> Nijmegen
>>>>>>>>>>> >> >> >>>>> The Netherlands
>>>>>>>>>>> >> >> >>>>>
>>>>>>>>>>> >> >> >>>>> Google Scholar Author Link
>>>>>>>>>>> >> >> >>>>>
>>>>>>>>>>> <http://scholar.google.com/citations?user=pLlSTVgAAAAJ&hl=en>
>>>>>>>>>>> >> >> >>>>>
>>>>>>>>>> >> >> >>>>
>>>>>>>>> >> >> >>>
>>>>>>>>> >> >> >>>
>>>>>>>>> >> >> >>>
>>>>>>>>> >> >> >>> --
>>>>>>>>> >> >> >>> Maarten Mennes, Ph.D.
>>>>>>>>> >> >> >>> Senior Researcher
>>>>>>>>> >> >> >>> Donders Institute for Brain, Cognition and Behaviour
>>>>>>>>> >> >> >>> Radboud University Nijmegen
>>>>>>>>> >> >> >>> Nijmegen
>>>>>>>>> >> >> >>> The Netherlands
>>>>>>>>> >> >> >>>
>>>>>>>>> >> >> >>> Google Scholar Author Link
>>>>>>>>> >> >> >>> <http://scholar.google.com/citations?user=pLlSTVgAAAAJ&hl=en>
>>>>>>>>> >> >> >>>
>>>>>>> >> >> >
>>>>>>> >> >> >
>>>>>> >> >>
>>>>> >> >
>>>> >>
>>> >
>
>
>
> --
> Maarten Mennes, Ph.D.
> Senior Researcher
> Donders Institute for Brain, Cognition and Behaviour
> Radboud University Nijmegen
> Nijmegen
> The Netherlands
>
> Google Scholar Author Link
> <http://scholar.google.com/citations?user=pLlSTVgAAAAJ&hl=en>
>
>
>
|