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 $PATH
>>
>> /Applications/caret/bin_macosx:/sw/bin:/Library/Frameworks/Python.framework/Versions/2.7/bin:/Users/angelafavaro/miniconda/bin:/sw/bin:/sw/sbin:/usr/local/fsl/bin:/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin:/usr/X11/bin:/usr/X11R6/bin:/usr/local/bic/bin:/Users/angelafavaro/abin
>> bash-3.2$
>>
>> 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 isn’t 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>
>> >> >>>
>> >> >
>> >> >
>> >>
>> >
>>
>
|