> On 3 Jul 2015, at 09:08, FSL automatic digest system <[log in to unmask]> wrote:
>
> There are 13 messages totaling 3363 lines in this issue.
>
> Topics of the day:
>
> 1. Dose FSL do "linear trend removal" in preprocessing?
> 2. Single subject analysis-RS, FSL error report
> 3. Fwd: [FSL] MELODIC experiment design (2)
> 4. A question about FIRSR
> 5. Fw: [FSL] Dose FSL do "linear trend removal" in preprocessing?
> 6. Issue with coregistration of spinal cord images
> 7. Melodic and Resting State
> 8. ASL partial volume correction
> 9. FSL website down until further notice
> 10. FSLVBM with pediatric population
> 11. Assistant Professor position (tenure-track)
> 12. Research Assistant/Lab Manager position
>
> ----------------------------------------------------------------------
>
> Date: Thu, 2 Jul 2015 09:25:14 +0800
> From: zhang mingxia <[log in to unmask]>
> Subject: Re: Dose FSL do "linear trend removal" in preprocessing?
>
> Dear FSL experts,
>
> Thank you for your patient answer! Acturally, I am not doing the
> connectivity analysis. Instead, I am investigating whether the timecourses
> of several ROIs are the same. I just used the standard way of preprocessing
> by FEAT(motion correction, smooth, high pass filter). Your email reminds me
> that the noise should be regressed out before timecourse extraction. Any
> other suggestions?
>
> I have two more question:
>
> 1. My design is A(150s)B(120s)A(150s), How the "high pass filter cutoff"
> should be set? How fslmaths -bptf highpass(what's this value should)
> lowpass(what's this value should) be set? I saw in the report_log that the
> lowpass is alway -1, why? And the value for highpass is the "high pass
> filter cutoff"/4?
>
> 2.An unrelated question: Is there some package in FSL for resting state
> connectivity analysis? I also did resting state connectivity analysis by
> FSL. However, after regressed out the noise, there were nagative value in
> the image data, so the further analysis(correlation of the timecoure of ROI
> with other voxels) could not be processed by FEAT. How should I resolve
> this issue?
>
> Thanks for advance!
>
> Mingxia Zhang
>
>> On Wed, Jul 1, 2015 at 7:21 PM, Iwo Bohr <[log in to unmask]> wrote:
>>
>> That's true but for e.g. connectivity analysis the standard way of
>> preprocessing ts is further signal cleaning as specified in my previous
>> email (noise regressing out) and also other than de-trending procedure:
>> low-pass filtering.
>>
>> To be honest it sounds like doing a similar thing in different ways (high
>> pass filtering and regressing out) but this is how it is. Maybe the
>> difference is that regressing out gets rid of other noise than only slow
>> drift noise...
>>
>> Iwo
>>
>> ------------------------------
>> *From:* Mark Jenkinson <[log in to unmask]>
>> *To:* [log in to unmask]
>> *Sent:* Wednesday, 1 July 2015, 12:04
>> *Subject:* Re: [FSL] Dose FSL do "linear trend removal" in preprocessing?
>>
>> Hi,
>>
>> The high pass filtering in FSL removes the linear trends.
>>
>> All the best,
>> Mark
>>
>>
>>
>>
>> From: FSL - FMRIB's Software Library <[log in to unmask]> on behalf of
>> zhang mingxia <[log in to unmask]>
>> Reply-To: FSL - FMRIB's Software Library <[log in to unmask]>
>> Date: Wednesday, 1 July 2015 11:29
>> To: "[log in to unmask]" <[log in to unmask]>
>> Subject: [FSL] Dose FSL do "linear trend removal" in preprocessing?
>>
>> Dear FSL experts,
>>
>> I have done the preprocessing by FSL GUI and will further extract the
>> time course for further analyis. I think the linear trend in the fMRI data
>> will badly influence the results, but I didn't see FSL did "linear trend
>> removal". Can FSL do "linear trend removal"? How can I do?
>>
>> Thanks in advance!
>>
>> Mingxia Zhang
>
> ------------------------------
>
> Date: Thu, 2 Jul 2015 09:53:40 +0530
> From: Ramesh Babu <[log in to unmask]>
> Subject: Re: Single subject analysis-RS, FSL error report
>
> Dear Paul,
> Thank you very much. Now I can create roi and also mask of dmn. Initially
> it was giving some error report saying that there is no image called
> my_mask. Then I typed dmn_component and then created a mask also. Now with
> confident I can scan further and do it in multi subjects.
>
> I was trying to attach png image file dmn component, but I couldn't. I
> request to increase the message size (which was 200kb) allotted for me. I
> will try to attache the image in next mail.
>
> Thank you very much.
> Ramesh
>
> On Tue, Jun 30, 2015 at 9:30 PM, paul mccarthy <[log in to unmask]>
> wrote:
>
>> Hi Ramesh,
>>
>> Ok - what you should do is look at the components in melodic_IC_hr.nii.gz
>> in FSLView, and identify the component with activity which most closely
>> resembles the DMN.
>>
>> Then, you can use fslroi to extract that component from melodic_IC_hr. For
>> example, if you find that component 20 (volume 20 in FSLView) is the DMN
>> component, you can extract it using:
>>
>> fslroi melodic_IC_hr dmn_component 20 1
>>
>> If necessary, you can then mask the dmn_component image using fslmaths:
>>
>> fslmaths dmn_component -mas my_mask dmn_component_masked
>>
>> Cheers,
>>
>> Paul
>>
>>> On 29 June 2015 at 18:19, Ramesh Babu <[log in to unmask]> wrote:
>>>
>>> Dear Paul,
>>>
>>> I want to find out DMN in single subject rsfMRI. I have done analysi by
>>> using melodic and I got all necessary out put files including
>>> "melodic_IC_hr.nii.gz" file. There are 29 volumes of images are present in
>>> the file which shows activities in different regions. By using
>>> WFU_PickAtlas I have created a single mask image which includes multiple
>>> rois (DMN). By using fslmaths command I have created threshold mask from
>>> the mask created earlier. In fsl view it appears fine.
>>>
>>> Further I want to know how to apply this mask so that I will get only
>>> DMN. (Just want to see dmn in single subject rsfMRI).
>>>
>>> Thanks
>>> Ramesh
>>>
>>> On Mon, Jun 29, 2015 at 9:44 PM, paul mccarthy <[log in to unmask]>
>>> wrote:
>>>
>>>> Hi Ramesh,
>>>>
>>>> It is not clear to me exactly what you are trying to do. Are you just
>>>> interested in these two scans from a single subject, or is this part of a
>>>> larger (multi-subject) study?
>>>>
>>>> Thanks,
>>>>
>>>> Paul
>>>>
>>>>
>>>>
>>>>
>>>> On 27 June 2015 at 11:21, Ramesh Babu <[log in to unmask]>
>>>> wrote:
>>>>
>>>>> Dear Paul,
>>>>>
>>>>> Earlier I had a doubt about the activities seen in my data set. After
>>>>> spending some time in FSL view, now I came to know how to look at different
>>>>> volumes of the maps.
>>>>>
>>>>> I have 2 rsfMRI scans, one meditators and one control scans. I have
>>>>> analysed separately in melodic as mentioned in the manual and obtained
>>>>> transformed map. By looking through FSL view, I found activities in the DMN
>>>>> area. The amount of the activity (size of the BOLD activity) is different
>>>>> in meditators and control.
>>>>>
>>>>> Should I create mask related to DMN? Is it necessary for single subject
>>>>> analysis? If it is necessary plz guide me how to create mask?
>>>>>
>>>>> Thanks
>>>>> Ramesh
>>>>>
>>>>>
>>>>>
>>>>> Since it is single subject analysis I didn't to any statistical
>>>>> analysis. Should I create ROI mask to define DMN network in these areas?
>>>>>
>>>>> On Fri, Jun 26, 2015 at 3:27 AM, Ramesh Babu <
>>>>> [log in to unmask]> wrote:
>>>>>
>>>>>> Dear Paul,
>>>>>> Thank you. I have learned a lesson that I should not copy paste in the
>>>>>> terminal. Thanks
>>>>>> In the downloaded data (short.nii, 100 volumes) I saw some activity in
>>>>>> the DMN. But in my data (has 120 volumes) it shows activity only in the
>>>>>> temporal region. Should I change the threshold to .5 (default value) in the
>>>>>> post stat window of melodic? or any other suggestion?
>>>>>>
>>>>>> Thanks
>>>>>> Ramesh
>>>>>>
>>>>>> On Fri, Jun 26, 2015 at 2:11 AM, paul mccarthy <
>>>>>> [log in to unmask]> wrote:
>>>>>>
>>>>>>> Hi Ramesh,
>>>>>>>
>>>>>>> This looks like a copy+paste issue again - try typing the fslview
>>>>>>> command manually, and it should work.
>>>>>>>
>>>>>>> Cheers,
>>>>>>>
>>>>>>> Paul
>>>>>>>
>>>>>>> On 25 June 2015 at 19:01, Ramesh Babu <[log in to unmask]>
>>>>>>> wrote:
>>>>>>>
>>>>>>>> Dear Paul,
>>>>>>>> When I type "fslview $FSLDIR/data/standard/MNI152_T1_2mm
>>>>>>>> melodic_IC_hr.nii.gz ‐l "Red‐Yellow" ‐b 5,10" it opens the fsl view window,
>>>>>>>> but it shows missing image/header. Then after clicking OK 3-4 times, it
>>>>>>>> opens the image. Should I ignore this or is there any solution for this
>>>>>>>> problem?
>>>>>>>>
>>>>>>>> Thanks
>>>>>>>> Ramesh
>>>>>>>>
>>>>>>>> On Wed, Jun 24, 2015 at 9:29 PM, paul mccarthy <
>>>>>>>> [log in to unmask]> wrote:
>>>>>>>>
>>>>>>>>> Hi Ramesh (sorry for getting your name wroing previously!),
>>>>>>>>>
>>>>>>>>> That is a complex question - there are many ways to identify the
>>>>>>>>> DMN. Probably the most common would be to use MELODIC on your data, and
>>>>>>>>> find the component which most closely resembles the DMN.
>>>>>>>>>
>>>>>>>>> You could use the component from the Smith et. al. 2011 study as a
>>>>>>>>> template:
>>>>>>>>>
>>>>>>>>> http://www.fmrib.ox.ac.uk/analysis/brainmap+rsns/
>>>>>>>>>
>>>>>>>>> Cheers,
>>>>>>>>>
>>>>>>>>> Paul
>>>>>>>>>
>>>>>>>>> On 24 June 2015 at 16:45, Ramesh Babu <[log in to unmask]>
>>>>>>>>> wrote:
>>>>>>>>>
>>>>>>>>>> Dear Paul,
>>>>>>>>>> Now I got the result. Earlier I have made a mistake by typing "\"
>>>>>>>>>> in the command line. After removing this, I got the result. I need one more
>>>>>>>>>> help from you. How to identify DMN by using fsl atlas. is there any other
>>>>>>>>>> way to find out DMN?
>>>>>>>>>>
>>>>>>>>>> Thanks for your help.
>>>>>>>>>> Ramesh
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> On Wed, Jun 24, 2015 at 8:39 PM, Ramesh Babu <
>>>>>>>>>> [log in to unmask]> wrote:
>>>>>>>>>>
>>>>>>>>>>> Dear Paul,
>>>>>>>>>>>
>>>>>>>>>>> Still I am facing the same problem.
>>>>>>>>>>>
>>>>>>>>>>> [fsl@fslvm6 filtered_func_data.ica]$ flirt -in melodic_IC -ref
>>>>>>>>>>> $FSLDIR/data/standard/MNI152_T1_2mm \ -out melodic_IC_hr -applyxfm -init
>>>>>>>>>>> ../reg/example_func2standard.mat
>>>>>>>>>>>
>>>>>>>>>>> Unrecognised option
>>>>>>>>>>>
>>>>>>>>>>> [fsl@fslvm6 filtered_func_data.ica]$
>>>>>>>>>>>
>>>>>>>>>>> Please see the list of file generated after first step which I
>>>>>>>>>>> have given below.
>>>>>>>>>>> [fsl@fslvm6 filtered_func_data.ica]$ ls
>>>>>>>>>>>
>>>>>>>>>>> eigenvalues_percent mean.nii.gz melodic_ICstats
>>>>>>>>>>> melodic_Tmodes
>>>>>>>>>>> log.txt melodic_FTmix melodic_mix report
>>>>>>>>>>> mask.nii.gz melodic_IC.nii.gz melodic_PPCA stats
>>>>>>>>>>>
>>>>>>>>>>> [fsl@fslvm6 filtered_func_data.ica]$
>>>>>>>>>>>
>>>>>>>>>>> Even after typing (not copy pasting), I am getting the same error
>>>>>>>>>>> message (Unrecognized options).
>>>>>>>>>>>
>>>>>>>>>>> It will be very helpful for me if you provide the solution for
>>>>>>>>>>> it.
>>>>>>>>>>>
>>>>>>>>>>> Thanks
>>>>>>>>>>> Ramesh
>>>>>>>>>>>
>>>>>>>>>>> On Tue, Jun 23, 2015 at 3:31 PM, paul mccarthy <
>>>>>>>>>>> [log in to unmask]> wrote:
>>>>>>>>>>>
>>>>>>>>>>>> Hi Ramesh,
>>>>>>>>>>>>
>>>>>>>>>>>> Did you copy+paste the flirt command from the web page? Instead,
>>>>>>>>>>>> manually type out the command, and see if that works.
>>>>>>>>>>>>
>>>>>>>>>>>> Cheers,
>>>>>>>>>>>>
>>>>>>>>>>>> Paul
>>>>>>>>>>>>
>>>>>>>>>>>> On 23 June 2015 at 07:32, Ramesh Babu <
>>>>>>>>>>>> [log in to unmask]> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>>> Dear FSL member,
>>>>>>>>>>>>>
>>>>>>>>>>>>> I have downloaded example data set for resting state fMRI
>>>>>>>>>>>>> analysis in FSL. I am following step by step procedure mentioned in the
>>>>>>>>>>>>> manual "Model Free FMRI Analysis Practical-Resting State data analysis
>>>>>>>>>>>>> (Optional)". I am getting an error which I have pasted below.
>>>>>>>>>>>>>
>>>>>>>>>>>>> "[fsl@fslvm6 filtered_func_data.ica]$ flirt ‐in melodic_IC
>>>>>>>>>>>>> ‐ref $FSLDIR/data/standard/MNI152_T1_2mm ‐out melodic_IC_hr ‐applyxfm ‐init
>>>>>>>>>>>>> ../reg/example_func2standard.mat
>>>>>>>>>>>>>
>>>>>>>>>>>>> Unrecognised option
>>>>>>>>>>>>>
>>>>>>>>>>>>> [fsl@fslvm6 filtered_func_data.ica]$"
>>>>>>>>>>>>>
>>>>>>>>>>>>> I didn't understand why it shows unrecognized option. Once I
>>>>>>>>>>>>> get DMN in this sample data then I will proceed scanning in similar way for
>>>>>>>>>>>>> my project.
>>>>>>>>>>>>>
>>>>>>>>>>>>> Since I am not familiar with scripting, it will be helpful for
>>>>>>>>>>>>> me if you give step by step usage of FLIRT GUI.
>>>>>>>>>>>>>
>>>>>>>>>>>>> Thanks
>>>>>>>>>>>>> Ramesh Babu
>
> ------------------------------
>
> Date: Thu, 2 Jul 2015 10:34:12 +0100
> From: "Anderson M. Winkler" <[log in to unmask]>
> Subject: Re: Fwd: [FSL] MELODIC experiment design
>
> Hi Sarah,
>
> The design and contrasts are fine. The order in which they appear doesn't
> matter, and for any given contrast, the EVs marked as 0 are nuisance. So,
> for C4, that tests EV5, all other behavioural scores, plus age and sex, are
> nuisance.
>
> For negative, repeat the same contrasts, replacing 1 with -1 (the contrasts
> would run from C1 to C16 then).
>
> As there are many contrasts, this introduces a multiple testing problem.
> Consider correcting for that (not yet in randomise, but available in PALM
> <http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/PALM> with the option "-corrcon").
>
> All the best,
>
> Anderson
>
>
>> On 1 July 2015 at 15:04, Sarah Izen <[log in to unmask]> wrote:
>>
>> Sending again because I'm not sure this message went through the first
>> time...
>>
>>
>> ---------- Forwarded message ----------
>> From: Sarah Izen <[log in to unmask]>
>> Date: Thu, Jun 11, 2015 at 12:36 PM
>> Subject: Re: [FSL] MELODIC experiment design
>> To: FSL - FMRIB's Software Library <[log in to unmask]>
>>
>>
>> Hi Anderson (or anyone else who wants to answer!),
>>
>> Hoping you won't mind another (rather basic) question about my design. I
>> only have one group and initially I just wanted to examine how functional
>> connectivity changed as a function of one behavioral measure. Now I want to
>> expand my analysis to have six behavioral measures plus sex and age. I'm a
>> little unsure how to model the contrasts. Essentially, I am just interested
>> to see if any of these variables has an effect on functional connectivity.
>> Here is what I have so far.
>>
>> 9 EVs:
>> EV1: group mean
>> EV2: sex
>> EV3: age
>> EV4: behavioral measure 1
>> EV5: behavioral measure 2
>> EV6: behavioral measure 3
>> EV7: behavioral measure 4
>> EV8: behavioral measure 5
>> EV9: behavioral measure 6
>>
>> I am quite new at all of this so I'm really not sure if I'm going in the
>> right direction but here is my initial matrix:
>>
>> EV1 EV2 EV3 EV4 EV5 EV6 EV7 EV8
>> EV9
>>
>> C1 0 1 0 0 0 0
>> 0 0 0
>> C2 0 0 1 0 0 0
>> 0 0 0
>> C3 0 0 0 1 0 0
>> 0 0 0
>> C4 0 0 0 0 1 0
>> 0 0 0
>> C5 0 0 0 0 0 1
>> 0 0 0
>> C6 0 0 0 0 0 0
>> 1 0 0
>> C7 0 0 0 0 0 0
>> 0 1 0
>> C8 0 0 0 0 0 0
>> 0 0 1
>>
>> It may turn out to be the case that we don't actually care about sex and
>> age and so wouldn't need a contrast for them, just for them to be regressed
>> out, but I'm kind of interested to see initially if this has any
>> significance. So, I have two questions. Does the order of the EVs in the
>> design matrix matter at all? Also, if I want to look at negative contrasts,
>> does it make more sense to double the amount of contrasts and have a -1 for
>> each or to run a separate analysis, also with 8 contrasts, but with only
>> -1s?
>>
>> Thanks!
>>
>> Sarah
>>
>> On Thu, Apr 23, 2015 at 3:29 AM, Anderson M. Winkler <
>> [log in to unmask]> wrote:
>>
>>> Hi Sarah,
>>>
>>> Please, see below:
>>>
>>>> On 22 April 2015 at 21:23, Sarah Izen <[log in to unmask]> wrote:
>>>>
>>>> Hi,
>>>>
>>>> Thanks so much. I really appreciate your help. My lab mostly uses SPM so
>>>> FSL is totally new territory for us. I may have been confused the "required
>>>> effect" number in the .con file. Is the required effect number the same as
>>>> the t-stat? For some reason I am getting a really high required effect
>>>> (3.879). I have 24 participants.
>>>
>>> You can pretty much ignore this value. It's difficult to interpret it in
>>> the dual_regression context and isn't used by randomise.
>>>
>>>
>>>>
>>>> Also, I can't find any documentation on what type of FWE is used.
>>>
>>>
>>> Have you checked the randomise webpage?
>>> http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/Randomise
>>>
>>>
>>>
>>>> It's obvious from my files that the FWE is getting rid of A LOT of
>>>> stuff. I'm also wondering what exactly TFCE does. Is it an enhancement or
>>>> some type of multiple comparisons correction?
>>>
>>>
>>> TFCE is a way to identify strong and spatially distributed signals. It is
>>> described in this paper: http://www.ncbi.nlm.nih.gov/pubmed/18501637
>>> It is an enhancement. TFCE results need correction too.
>>>
>>>
>>> If it's enhancement, why do my clusters look smaller in my tfce p image
>>>> than in the raw tstat image?
>>>
>>>
>>> The size of the surviving regions depends on significance. Are you
>>> looking at the same correction (i.e., fwep for both, or uncp for both?). At
>>> any rate, while TFCE tends to be more powerful, the local configuration of
>>> the signal in the neighbours and support region don't guarantee more power
>>> always.
>>>
>>>
>>>
>>>> Also, if the TFCE essentially has the same goal as smoothing, was it
>>>> redundant to have done spatial smoothing during prestats?
>>>
>>>
>>> TFCE does not have the same goals as smoothing. It's something entirely
>>> different.
>>>
>>>
>>>
>>>> I've heard that the TFCE has a more stringent correction. If that's the
>>>> case, why do we also do a FWE?
>>>
>>>
>>> TFCE isn't a correction, but a way to strengthen the ability to identify
>>> spatial signals. It isn't more or less stringent, as the stringency depends
>>> on the permutation test, which ensures exactness of p-values. It tends to
>>> be more powerful, though.
>>>
>>>
>>>
>>>>
>>>> Also, is there a way to look at numbers of voxels in clusters in FSLView?
>>>
>>> I'm not sure in FSLview, but you can get that easily with the command
>>> "cluster", or thresholding with "fslmaths" then using "fslstats".
>>>
>>> All the best,
>>>
>>> Anderson
>>>
>>>
>>>
>>>
>>>>
>>>> Thanks,
>>>> Sarah
>>>>
>>>> On Mon, Apr 20, 2015 at 1:49 AM, Anderson M. Winkler <
>>>> [log in to unmask]> wrote:
>>>>
>>>>> Hi Sarah,
>>>>>
>>>>> Please, see below:
>>>>>
>>>>>
>>>>>> On 16 April 2015 at 20:58, Sarah Izen <[log in to unmask]> wrote:
>>>>>>
>>>>>> Hi Anderson,
>>>>>>
>>>>>> Thanks so much for all your help. I ran the dual regression and
>>>>>> randomise and I guess it automatically does the TFCE per the command I used
>>>>>> but I'm wondering if you could give me some insight as to which randomise
>>>>>> output might be the best for my situation. What are some reasons to choose
>>>>>> TFCE or cluster-based, and further, cluster-based extent or mass?
>>>>>
>>>>> The dual_regression does by default TFCE only, but cluster extent
>>>>> and/or mass may be included too. Each of these statistics tell slightly
>>>>> different things about the data, following their respective definitions.
>>>>> TFCE doesn't require a cluster-forming threshold, which is generally a
>>>>> benefit. In principle any of these can be used, as long as properly
>>>>> corrected for multiple testing.
>>>>>
>>>>>
>>>>>
>>>>>>
>>>>>> Additionally, for some reason my tstat map does not seem to be
>>>>>> matching up with my uncorrected p map. The tstat map shows mostly the same
>>>>>> activations but with many more voxels activated. Do you have any idea why
>>>>>> this could be? I compared them setting my tstat min to 1.234 which should
>>>>>> be my cutoff for .05 significance. I am attaching a screenshot of the two
>>>>>> of them next to each other.
>>>>>
>>>>> The map on the left is a t-statistic. If various assumptions about the
>>>>> data were guaranteed to be valid, and the sample size very large, the
>>>>> cutoff for p=0.05 would be about t=1.64 (so, higher than 1.234). The map on
>>>>> the right has the p-values not for the t-statistic, but for the TFCE
>>>>> computed using the t-statistic.
>>>>>
>>>>> To make a more fair comparison (but still rough), threshold the image
>>>>> on the left at 1.64 (or compute the exact value given your degrees of
>>>>> freedom of the test), and replace the image on the right for the
>>>>> dr_stage3_ic0016_vox_p_tstat1, threshold at 0.95. This file isn't produced
>>>>> by default, and if you don't have it, you'd have to run randomise again,
>>>>> now including the option -x, and if you are using the latest FSL, also
>>>>> adding the option --uncorrp (you can edit the dual_regression script, there
>>>>> is a clear indication for this towards the end of the script).
>>>>>
>>>>> You don't actually need to do this, it's just in case you'd like to
>>>>> understand better what is going on.
>>>>>
>>>>>
>>>>>
>>>>>>
>>>>>> If I compare a component of my input data to that same component's
>>>>>> randomise output files the activation doesn't always match up.
>>>>>
>>>>> There is no guarantee that they would match. If you only want to see
>>>>> the voxels that are part of a given network, consider using a mask. See
>>>>> this question in the FAQ:
>>>>> http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/DualRegression/Faq
>>>>>
>>>>>
>>>>>
>>>>>> Does the activation in the output files signify areas that are
>>>>>> functionally connected to the networks I used as my input data, or is the
>>>>>> output data supposed to show areas of the input network that differ based
>>>>>> on my covariate (behavioral scores)?
>>>>>
>>>>> Neither I'm afraid. The maps show differences between groups, taking
>>>>> the behavioural as nuisance, or where the nuisance is associated
>>>>> (correlated) with the networks having the group as nuisance. It depends on
>>>>> the contrast used with the GLM.
>>>>>
>>>>> All the best,
>>>>>
>>>>> Anderson
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>>
>>>>>> Again, thanks so much. I really appreciate you taking the time to help
>>>>>> me!
>>>>>>
>>>>>> Sarah
>>>>>>
>>>>>> On Thu, Apr 16, 2015 at 2:02 AM, Anderson M. Winkler <
>>>>>> [log in to unmask]> wrote:
>>>>>>
>>>>>>> Hi Sarah,
>>>>>>>
>>>>>>> Very quick:
>>>>>>> - No need for -1
>>>>>>> - Use the example from the manual (2 EVs, being one intercept and one
>>>>>>> for behavioural scores).
>>>>>>> - No need to demean the behavioural EV, unless you wanted to test the
>>>>>>> intercept -- see Jeanette Mumford's page:
>>>>>>> http://mumford.fmripower.org/mean_centering
>>>>>>> - No need for -D
>>>>>>>
>>>>>>> All the best,
>>>>>>>
>>>>>>> Anderson
>>>>>>>
>>>>>>>
>>>>>>>> On 15 April 2015 at 18:44, Sarah Izen <[log in to unmask]> wrote:
>>>>>>>>
>>>>>>>> Hi Anderson,
>>>>>>>>
>>>>>>>> Thanks for your help. Just want to make sure I fully understand what
>>>>>>>> you're saying. So would it be correct that I should not use the -1 option
>>>>>>>> since I have the additional covariate?
>>>>>>>>
>>>>>>>> I ran dual regression and randomise with this command:
>>>>>>>>
>>>>>>>> dual_regression <4d_input> 1 design.mat design.con 500 <output>
>>>>>>>> <inputs>
>>>>>>>>
>>>>>>>> Is this correct? I'm a little concerned about my results because my
>>>>>>>> corrected p-value results for each component don't correspond with
>>>>>>>> activation in the input components at all. Should my corrected p-value
>>>>>>>> images show areas of activation of my input components that are
>>>>>>>> significantly correlated with my EV or is it showing areas that are
>>>>>>>> functionally connected to the areas activated in my input components?
>>>>>>>>
>>>>>>>> I also just want to make sure I set up my design files in the right
>>>>>>>> way because I was a little confused by the GLM wiki page. If I have one
>>>>>>>> group with one EV do I set up the design with two EVs (i.e. group mean and
>>>>>>>> behavioral scores) or just one EV (behavioral scores)? Am I correct in
>>>>>>>> thinking I can just do the one EV if I add the -D option at the end of my
>>>>>>>> dual regression command?
>>>>>>>>
>>>>>>>> Sorry for so many questions! I really appreciate your help.
>>>>>>>>
>>>>>>>>
>>>>>>>> Sarah
>>>>>>>>
>>>>>>>> On Fri, Apr 10, 2015 at 2:25 AM, Anderson M. Winkler <
>>>>>>>> [log in to unmask]> wrote:
>>>>>>>>
>>>>>>>>> Hi Sarah,
>>>>>>>>>
>>>>>>>>> Please, see below:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>> On 8 April 2015 at 21:56, Sarah Izen <[log in to unmask]> wrote:
>>>>>>>>>>
>>>>>>>>>> Hi,
>>>>>>>>>>
>>>>>>>>>> Thanks so much - I realized this right after I emailed the list. I
>>>>>>>>>> just have a couple of questions about my design. I have one group and one
>>>>>>>>>> EV (scores on behavioral tests). I just want to make sure I am on the right
>>>>>>>>>> track with my commands.
>>>>>>>>>>
>>>>>>>>>> For dual regression: dual_regression <4d_input_data> 1
>>>>>>>>>> <design.mat> <design.con> 0 <output_directory> <inputs>
>>>>>>>>>>
>>>>>>>>>> For randomise: randomise -i <input> -o <output> -1
>>>>>>>>>>
>>>>>>>>>> Are these correct? I'm not positive about the randomise command -
>>>>>>>>>> I know the user guide says you do not need to specify design.mat and
>>>>>>>>>> design.con files for a one sample test but just want to make sure I'm doing
>>>>>>>>>> it right.
>>>>>>>>>
>>>>>>>>> It would be right if it wasn't for the covariate. The simpler
>>>>>>>>> randomise call above will use a 1-sample t-test without any other EV. For
>>>>>>>>> your design you need the full design.mat and design.con files, as in the
>>>>>>>>> example in the GLM manual.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>> Additionally, I know commands can be written to do both dual
>>>>>>>>>> regression and randomise at the same time but I'm not sure how to do that
>>>>>>>>>> with my design since I'm assuming dual regression still needs the two
>>>>>>>>>> design files that randomise doesn't need? I can run them separately but
>>>>>>>>>> just wondered if there was an easier way.
>>>>>>>>>
>>>>>>>>> As you have the extra EV, you won't be using the randomise call as
>>>>>>>>> above, so the easier way is to provide the design files to the
>>>>>>>>> dual_regression script.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Also, with regards to interpreting corrected p-value images for my
>>>>>>>>>> experiment, if I am looking at behavioral scores does that mean that areas
>>>>>>>>>> of activation in corrected pvalue images are areas that vary with regard to
>>>>>>>>>> the subjects' behavioral scores? Just want to make sure I am interpreting
>>>>>>>>>> this in the right way.
>>>>>>>>>
>>>>>>>>> If the contrast tests the behavioural scores, the results are for
>>>>>>>>> the behavioural scores. Otherwise it will be testing just the mean.
>>>>>>>>>
>>>>>>>>> All the best,
>>>>>>>>>
>>>>>>>>> Anderson
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Thanks so much for your help,
>>>>>>>>>> Sarah
>>>>>>>>>>
>>>>>>>>>> On Tue, Apr 7, 2015 at 1:41 AM, Stephen Smith <
>>>>>>>>>> [log in to unmask]> wrote:
>>>>>>>>>>
>>>>>>>>>>> Hi - have another look at the FSL wiki doc on ICA/MELODIC,
>>>>>>>>>>> dual-regression and randomise. dualreg+randomise should still be the way
>>>>>>>>>>> to go with your design.
>>>>>>>>>>> Cheers.
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> On 6 Apr 2015, at 17:28, Sarah Izen <[log in to unmask]> wrote:
>>>>>>>>>>>
>>>>>>>>>>> Hi all,
>>>>>>>>>>>
>>>>>>>>>>> I just want to make sure I'm on the right track with my analysis.
>>>>>>>>>>> I am very new to FSL! I have a set of resting state data that I ran using
>>>>>>>>>>> MELODIC. I preprocessed everything, denoised, ran melodic, etc. Now, I want
>>>>>>>>>>> to compare my networks to scores on behavioral tests. From what I was told,
>>>>>>>>>>> I can do this with GLM - single group with additional covariate. But, as I
>>>>>>>>>>> try to set it up I see that I have to input lower level FEAT directories.
>>>>>>>>>>> How can I used my melodic data as an input? I've only really ever seen
>>>>>>>>>>> melodic used with dual regression but that does not seem appropriate in my
>>>>>>>>>>> case since I only have one group of participants.
>>>>>>>>>>>
>>>>>>>>>>> Thanks,
>>>>>>>>>>> Sarah
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> ---------------------------------------------------------------------------
>>>>>>>>>>> Stephen M. Smith, Professor of Biomedical Engineering
>>>>>>>>>>> Associate Director, Oxford University FMRIB Centre
>>>>>>>>>>>
>>>>>>>>>>> FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
>>>>>>>>>>> +44 (0) 1865 222726 (fax 222717)
>>>>>>>>>>> [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
>>>>>>>>>>>
>>>>>>>>>>> ---------------------------------------------------------------------------
>>>>>>>>>>>
>>>>>>>>>>> Stop the cultural destruction of Tibet <http://smithinks.net>
>
> ------------------------------
>
> Date: Thu, 2 Jul 2015 19:18:18 +0900
> From: jh kim <[log in to unmask]>
> Subject: A question about FIRSR
>
> Dear FSL developers and experts
>
> I've done between-group comparison (controls vs. patients) of hippocampal
> shape using FIRST implemented in FSL version 5.0.6.
> I strictly followed every steps described in online userguide.
>
> I've checked the correct registration in all subjects before statistical
> analysis by using the following command:
> ${FSLDIR}/bin/slicesdir -p
> ${FSLDIR}/data/standard/MNI152_T1_1mm.nii.gz *_to_std_sub.nii.gz
> Correct segmentation of hippocampus and the other subcortical structures
> were also checked by using 'first_roi_slicedir'.
>
> One of the reviewers commented the following query:
> "How is across-subjects correspondence established in the FIRST surface
> analysis? In other words, how did authors ensure that vertex locations were
> corresponding across individuals?"
>
> What is the appropriate answer to this question?
> Any comments would be much appreciated.
>
> Thank you in advance for your help.
>
>
> Kim
>
> ------------------------------
>
> Date: Thu, 2 Jul 2015 10:08:26 -0400
> From: Sarah Izen <[log in to unmask]>
> Subject: Re: Fwd: [FSL] MELODIC experiment design
>
> Thanks so much! I'll take a look at PALM.
>
> On Thu, Jul 2, 2015 at 5:34 AM, Anderson M. Winkler <[log in to unmask]>
> wrote:
>
>> Hi Sarah,
>>
>> The design and contrasts are fine. The order in which they appear doesn't
>> matter, and for any given contrast, the EVs marked as 0 are nuisance. So,
>> for C4, that tests EV5, all other behavioural scores, plus age and sex, are
>> nuisance.
>>
>> For negative, repeat the same contrasts, replacing 1 with -1 (the
>> contrasts would run from C1 to C16 then).
>>
>> As there are many contrasts, this introduces a multiple testing problem.
>> Consider correcting for that (not yet in randomise, but available in PALM
>> <http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/PALM> with the option "-corrcon").
>>
>> All the best,
>>
>> Anderson
>>
>>
>>> On 1 July 2015 at 15:04, Sarah Izen <[log in to unmask]> wrote:
>>>
>>> Sending again because I'm not sure this message went through the first
>>> time...
>>>
>>>
>>> ---------- Forwarded message ----------
>>> From: Sarah Izen <[log in to unmask]>
>>> Date: Thu, Jun 11, 2015 at 12:36 PM
>>> Subject: Re: [FSL] MELODIC experiment design
>>> To: FSL - FMRIB's Software Library <[log in to unmask]>
>>>
>>>
>>> Hi Anderson (or anyone else who wants to answer!),
>>>
>>> Hoping you won't mind another (rather basic) question about my design. I
>>> only have one group and initially I just wanted to examine how functional
>>> connectivity changed as a function of one behavioral measure. Now I want to
>>> expand my analysis to have six behavioral measures plus sex and age. I'm a
>>> little unsure how to model the contrasts. Essentially, I am just interested
>>> to see if any of these variables has an effect on functional connectivity.
>>> Here is what I have so far.
>>>
>>> 9 EVs:
>>> EV1: group mean
>>> EV2: sex
>>> EV3: age
>>> EV4: behavioral measure 1
>>> EV5: behavioral measure 2
>>> EV6: behavioral measure 3
>>> EV7: behavioral measure 4
>>> EV8: behavioral measure 5
>>> EV9: behavioral measure 6
>>>
>>> I am quite new at all of this so I'm really not sure if I'm going in the
>>> right direction but here is my initial matrix:
>>>
>>> EV1 EV2 EV3 EV4 EV5 EV6 EV7 EV8
>>> EV9
>>>
>>> C1 0 1 0 0 0 0
>>> 0 0 0
>>> C2 0 0 1 0 0 0
>>> 0 0 0
>>> C3 0 0 0 1 0 0
>>> 0 0 0
>>> C4 0 0 0 0 1 0
>>> 0 0 0
>>> C5 0 0 0 0 0 1
>>> 0 0 0
>>> C6 0 0 0 0 0 0
>>> 1 0 0
>>> C7 0 0 0 0 0 0
>>> 0 1 0
>>> C8 0 0 0 0 0 0
>>> 0 0 1
>>>
>>> It may turn out to be the case that we don't actually care about sex and
>>> age and so wouldn't need a contrast for them, just for them to be regressed
>>> out, but I'm kind of interested to see initially if this has any
>>> significance. So, I have two questions. Does the order of the EVs in the
>>> design matrix matter at all? Also, if I want to look at negative contrasts,
>>> does it make more sense to double the amount of contrasts and have a -1 for
>>> each or to run a separate analysis, also with 8 contrasts, but with only
>>> -1s?
>>>
>>> Thanks!
>>>
>>> Sarah
>>>
>>> On Thu, Apr 23, 2015 at 3:29 AM, Anderson M. Winkler <
>>> [log in to unmask]> wrote:
>>>
>>>> Hi Sarah,
>>>>
>>>> Please, see below:
>>>>
>>>>> On 22 April 2015 at 21:23, Sarah Izen <[log in to unmask]> wrote:
>>>>>
>>>>> Hi,
>>>>>
>>>>> Thanks so much. I really appreciate your help. My lab mostly uses SPM
>>>>> so FSL is totally new territory for us. I may have been confused the
>>>>> "required effect" number in the .con file. Is the required effect number
>>>>> the same as the t-stat? For some reason I am getting a really high required
>>>>> effect (3.879). I have 24 participants.
>>>>
>>>> You can pretty much ignore this value. It's difficult to interpret it in
>>>> the dual_regression context and isn't used by randomise.
>>>>
>>>>
>>>>>
>>>>> Also, I can't find any documentation on what type of FWE is used.
>>>>
>>>>
>>>> Have you checked the randomise webpage?
>>>> http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/Randomise
>>>>
>>>>
>>>>
>>>>> It's obvious from my files that the FWE is getting rid of A LOT of
>>>>> stuff. I'm also wondering what exactly TFCE does. Is it an enhancement or
>>>>> some type of multiple comparisons correction?
>>>>
>>>>
>>>> TFCE is a way to identify strong and spatially distributed signals. It
>>>> is described in this paper: http://www.ncbi.nlm.nih.gov/pubmed/18501637
>>>> It is an enhancement. TFCE results need correction too.
>>>>
>>>>
>>>> If it's enhancement, why do my clusters look smaller in my tfce p image
>>>>> than in the raw tstat image?
>>>>
>>>>
>>>> The size of the surviving regions depends on significance. Are you
>>>> looking at the same correction (i.e., fwep for both, or uncp for both?). At
>>>> any rate, while TFCE tends to be more powerful, the local configuration of
>>>> the signal in the neighbours and support region don't guarantee more power
>>>> always.
>>>>
>>>>
>>>>
>>>>> Also, if the TFCE essentially has the same goal as smoothing, was it
>>>>> redundant to have done spatial smoothing during prestats?
>>>>
>>>>
>>>> TFCE does not have the same goals as smoothing. It's something entirely
>>>> different.
>>>>
>>>>
>>>>
>>>>> I've heard that the TFCE has a more stringent correction. If that's the
>>>>> case, why do we also do a FWE?
>>>>
>>>>
>>>> TFCE isn't a correction, but a way to strengthen the ability to identify
>>>> spatial signals. It isn't more or less stringent, as the stringency depends
>>>> on the permutation test, which ensures exactness of p-values. It tends to
>>>> be more powerful, though.
>>>>
>>>>
>>>>
>>>>>
>>>>> Also, is there a way to look at numbers of voxels in clusters in
>>>>> FSLView?
>>>>
>>>> I'm not sure in FSLview, but you can get that easily with the command
>>>> "cluster", or thresholding with "fslmaths" then using "fslstats".
>>>>
>>>> All the best,
>>>>
>>>> Anderson
>>>>
>>>>
>>>>
>>>>
>>>>>
>>>>> Thanks,
>>>>> Sarah
>>>>>
>>>>> On Mon, Apr 20, 2015 at 1:49 AM, Anderson M. Winkler <
>>>>> [log in to unmask]> wrote:
>>>>>
>>>>>> Hi Sarah,
>>>>>>
>>>>>> Please, see below:
>>>>>>
>>>>>>
>>>>>>> On 16 April 2015 at 20:58, Sarah Izen <[log in to unmask]> wrote:
>>>>>>>
>>>>>>> Hi Anderson,
>>>>>>>
>>>>>>> Thanks so much for all your help. I ran the dual regression and
>>>>>>> randomise and I guess it automatically does the TFCE per the command I used
>>>>>>> but I'm wondering if you could give me some insight as to which randomise
>>>>>>> output might be the best for my situation. What are some reasons to choose
>>>>>>> TFCE or cluster-based, and further, cluster-based extent or mass?
>>>>>>
>>>>>> The dual_regression does by default TFCE only, but cluster extent
>>>>>> and/or mass may be included too. Each of these statistics tell slightly
>>>>>> different things about the data, following their respective definitions.
>>>>>> TFCE doesn't require a cluster-forming threshold, which is generally a
>>>>>> benefit. In principle any of these can be used, as long as properly
>>>>>> corrected for multiple testing.
>>>>>>
>>>>>>
>>>>>>
>>>>>>>
>>>>>>> Additionally, for some reason my tstat map does not seem to be
>>>>>>> matching up with my uncorrected p map. The tstat map shows mostly the same
>>>>>>> activations but with many more voxels activated. Do you have any idea why
>>>>>>> this could be? I compared them setting my tstat min to 1.234 which should
>>>>>>> be my cutoff for .05 significance. I am attaching a screenshot of the two
>>>>>>> of them next to each other.
>>>>>>
>>>>>> The map on the left is a t-statistic. If various assumptions about the
>>>>>> data were guaranteed to be valid, and the sample size very large, the
>>>>>> cutoff for p=0.05 would be about t=1.64 (so, higher than 1.234). The map on
>>>>>> the right has the p-values not for the t-statistic, but for the TFCE
>>>>>> computed using the t-statistic.
>>>>>>
>>>>>> To make a more fair comparison (but still rough), threshold the image
>>>>>> on the left at 1.64 (or compute the exact value given your degrees of
>>>>>> freedom of the test), and replace the image on the right for the
>>>>>> dr_stage3_ic0016_vox_p_tstat1, threshold at 0.95. This file isn't produced
>>>>>> by default, and if you don't have it, you'd have to run randomise again,
>>>>>> now including the option -x, and if you are using the latest FSL, also
>>>>>> adding the option --uncorrp (you can edit the dual_regression script, there
>>>>>> is a clear indication for this towards the end of the script).
>>>>>>
>>>>>> You don't actually need to do this, it's just in case you'd like to
>>>>>> understand better what is going on.
>>>>>>
>>>>>>
>>>>>>
>>>>>>>
>>>>>>> If I compare a component of my input data to that same component's
>>>>>>> randomise output files the activation doesn't always match up.
>>>>>>
>>>>>> There is no guarantee that they would match. If you only want to see
>>>>>> the voxels that are part of a given network, consider using a mask. See
>>>>>> this question in the FAQ:
>>>>>> http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/DualRegression/Faq
>>>>>>
>>>>>>
>>>>>>
>>>>>>> Does the activation in the output files signify areas that are
>>>>>>> functionally connected to the networks I used as my input data, or is the
>>>>>>> output data supposed to show areas of the input network that differ based
>>>>>>> on my covariate (behavioral scores)?
>>>>>>
>>>>>> Neither I'm afraid. The maps show differences between groups, taking
>>>>>> the behavioural as nuisance, or where the nuisance is associated
>>>>>> (correlated) with the networks having the group as nuisance. It depends on
>>>>>> the contrast used with the GLM.
>>>>>>
>>>>>> All the best,
>>>>>>
>>>>>> Anderson
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>>
>>>>>>> Again, thanks so much. I really appreciate you taking the time to
>>>>>>> help me!
>>>>>>>
>>>>>>> Sarah
>>>>>>>
>>>>>>> On Thu, Apr 16, 2015 at 2:02 AM, Anderson M. Winkler <
>>>>>>> [log in to unmask]> wrote:
>>>>>>>
>>>>>>>> Hi Sarah,
>>>>>>>>
>>>>>>>> Very quick:
>>>>>>>> - No need for -1
>>>>>>>> - Use the example from the manual (2 EVs, being one intercept and
>>>>>>>> one for behavioural scores).
>>>>>>>> - No need to demean the behavioural EV, unless you wanted to test
>>>>>>>> the intercept -- see Jeanette Mumford's page:
>>>>>>>> http://mumford.fmripower.org/mean_centering
>>>>>>>> - No need for -D
>>>>>>>>
>>>>>>>> All the best,
>>>>>>>>
>>>>>>>> Anderson
>>>>>>>>
>>>>>>>>
>>>>>>>>> On 15 April 2015 at 18:44, Sarah Izen <[log in to unmask]> wrote:
>>>>>>>>>
>>>>>>>>> Hi Anderson,
>>>>>>>>>
>>>>>>>>> Thanks for your help. Just want to make sure I fully understand
>>>>>>>>> what you're saying. So would it be correct that I should not use the -1
>>>>>>>>> option since I have the additional covariate?
>>>>>>>>>
>>>>>>>>> I ran dual regression and randomise with this command:
>>>>>>>>>
>>>>>>>>> dual_regression <4d_input> 1 design.mat design.con 500 <output>
>>>>>>>>> <inputs>
>>>>>>>>>
>>>>>>>>> Is this correct? I'm a little concerned about my results because my
>>>>>>>>> corrected p-value results for each component don't correspond with
>>>>>>>>> activation in the input components at all. Should my corrected p-value
>>>>>>>>> images show areas of activation of my input components that are
>>>>>>>>> significantly correlated with my EV or is it showing areas that are
>>>>>>>>> functionally connected to the areas activated in my input components?
>>>>>>>>>
>>>>>>>>> I also just want to make sure I set up my design files in the right
>>>>>>>>> way because I was a little confused by the GLM wiki page. If I have one
>>>>>>>>> group with one EV do I set up the design with two EVs (i.e. group mean and
>>>>>>>>> behavioral scores) or just one EV (behavioral scores)? Am I correct in
>>>>>>>>> thinking I can just do the one EV if I add the -D option at the end of my
>>>>>>>>> dual regression command?
>>>>>>>>>
>>>>>>>>> Sorry for so many questions! I really appreciate your help.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Sarah
>>>>>>>>>
>>>>>>>>> On Fri, Apr 10, 2015 at 2:25 AM, Anderson M. Winkler <
>>>>>>>>> [log in to unmask]> wrote:
>>>>>>>>>
>>>>>>>>>> Hi Sarah,
>>>>>>>>>>
>>>>>>>>>> Please, see below:
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>> On 8 April 2015 at 21:56, Sarah Izen <[log in to unmask]> wrote:
>>>>>>>>>>>
>>>>>>>>>>> Hi,
>>>>>>>>>>>
>>>>>>>>>>> Thanks so much - I realized this right after I emailed the list.
>>>>>>>>>>> I just have a couple of questions about my design. I have one group and one
>>>>>>>>>>> EV (scores on behavioral tests). I just want to make sure I am on the right
>>>>>>>>>>> track with my commands.
>>>>>>>>>>>
>>>>>>>>>>> For dual regression: dual_regression <4d_input_data> 1
>>>>>>>>>>> <design.mat> <design.con> 0 <output_directory> <inputs>
>>>>>>>>>>>
>>>>>>>>>>> For randomise: randomise -i <input> -o <output> -1
>>>>>>>>>>>
>>>>>>>>>>> Are these correct? I'm not positive about the randomise command -
>>>>>>>>>>> I know the user guide says you do not need to specify design.mat and
>>>>>>>>>>> design.con files for a one sample test but just want to make sure I'm doing
>>>>>>>>>>> it right.
>>>>>>>>>>
>>>>>>>>>> It would be right if it wasn't for the covariate. The simpler
>>>>>>>>>> randomise call above will use a 1-sample t-test without any other EV. For
>>>>>>>>>> your design you need the full design.mat and design.con files, as in the
>>>>>>>>>> example in the GLM manual.
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>> Additionally, I know commands can be written to do both dual
>>>>>>>>>>> regression and randomise at the same time but I'm not sure how to do that
>>>>>>>>>>> with my design since I'm assuming dual regression still needs the two
>>>>>>>>>>> design files that randomise doesn't need? I can run them separately but
>>>>>>>>>>> just wondered if there was an easier way.
>>>>>>>>>>
>>>>>>>>>> As you have the extra EV, you won't be using the randomise call as
>>>>>>>>>> above, so the easier way is to provide the design files to the
>>>>>>>>>> dual_regression script.
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Also, with regards to interpreting corrected p-value images for
>>>>>>>>>>> my experiment, if I am looking at behavioral scores does that mean that
>>>>>>>>>>> areas of activation in corrected pvalue images are areas that vary with
>>>>>>>>>>> regard to the subjects' behavioral scores? Just want to make sure I am
>>>>>>>>>>> interpreting this in the right way.
>>>>>>>>>>
>>>>>>>>>> If the contrast tests the behavioural scores, the results are for
>>>>>>>>>> the behavioural scores. Otherwise it will be testing just the mean.
>>>>>>>>>>
>>>>>>>>>> All the best,
>>>>>>>>>>
>>>>>>>>>> Anderson
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> Thanks so much for your help,
>>>>>>>>>>> Sarah
>>>>>>>>>>>
>>>>>>>>>>> On Tue, Apr 7, 2015 at 1:41 AM, Stephen Smith <
>>>>>>>>>>> [log in to unmask]> wrote:
>>>>>>>>>>>
>>>>>>>>>>>> Hi - have another look at the FSL wiki doc on ICA/MELODIC,
>>>>>>>>>>>> dual-regression and randomise. dualreg+randomise should still be the way
>>>>>>>>>>>> to go with your design.
>>>>>>>>>>>> Cheers.
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> On 6 Apr 2015, at 17:28, Sarah Izen <[log in to unmask]> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>> Hi all,
>>>>>>>>>>>>
>>>>>>>>>>>> I just want to make sure I'm on the right track with my
>>>>>>>>>>>> analysis. I am very new to FSL! I have a set of resting state data that I
>>>>>>>>>>>> ran using MELODIC. I preprocessed everything, denoised, ran melodic, etc.
>>>>>>>>>>>> Now, I want to compare my networks to scores on behavioral tests. From what
>>>>>>>>>>>> I was told, I can do this with GLM - single group with additional
>>>>>>>>>>>> covariate. But, as I try to set it up I see that I have to input lower
>>>>>>>>>>>> level FEAT directories. How can I used my melodic data as an input? I've
>>>>>>>>>>>> only really ever seen melodic used with dual regression but that does not
>>>>>>>>>>>> seem appropriate in my case since I only have one group of participants.
>>>>>>>>>>>>
>>>>>>>>>>>> Thanks,
>>>>>>>>>>>> Sarah
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> ---------------------------------------------------------------------------
>>>>>>>>>>>> Stephen M. Smith, Professor of Biomedical Engineering
>>>>>>>>>>>> Associate Director, Oxford University FMRIB Centre
>>>>>>>>>>>>
>>>>>>>>>>>> FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
>>>>>>>>>>>> +44 (0) 1865 222726 (fax 222717)
>>>>>>>>>>>> [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
>>>>>>>>>>>>
>>>>>>>>>>>> ---------------------------------------------------------------------------
>>>>>>>>>>>>
>>>>>>>>>>>> Stop the cultural destruction of Tibet <http://smithinks.net>
>
> ------------------------------
>
> Date: Thu, 2 Jul 2015 14:19:38 +0000
> From: Iwo Bohr <[log in to unmask]>
> Subject: Re: Fw: [FSL] Dose FSL do "linear trend removal" in preprocessing?
>
> Yet another little comment for completeness on extracted BOLD cleaning :
> motion parameters as generated in preprocessing stage of Feat are also conventionally included in step 2 (see below) to account for residual effects of head motion.
> Iwo
>
> From: Iwo Bohr <[log in to unmask]>
> To: [log in to unmask]
> Sent: Wednesday, 1 July 2015, 12:10
> Subject: [FSL] Fw: [FSL] Dose FSL do "linear trend removal" in preprocessing?
>
> One remark re point 1: if you did some preprocessing using FEAT then probably it included high-pass filtering in which case you need to do only low pass filtering with "fslmaths - bptf"
>
> Iwo
>
>
>
> ----- Forwarded Message -----
> From: Iwo Bohr <[log in to unmask]>
> To: FSL - FMRIB's Software Library <[log in to unmask]>
> Sent: Wednesday, 1 July 2015, 12:04
> Subject: Re: [FSL] Dose FSL do "linear trend removal" in preprocessing?
>
> Dear Zhang,
> I don't know what you exactly want to do with the extracted ts but normally there are ways of dealing with slow trends and high frequency noise as well, such as summarized it the following:
> 1. band pass filtering using: "fslmaths - bptf" setting appropriately low pass and high pass thresholds
> 2. regressing out BOLD ts from ventricles and WM from your ts of interest using: " fsl_glm" tool with your ventricle and WM ts put as argument in a txt file to "-d" (design) option
> Just an outline of one possibility. A more advanced option is using FIX tool (automatically removing noise from signal) which is still under development and not integrated with the actual package...
>
> Best,Iwo
>
>
> From: zhang mingxia <[log in to unmask]>
> To: [log in to unmask]
> Sent: Wednesday, 1 July 2015, 11:29
> Subject: [FSL] Dose FSL do "linear trend removal" in preprocessing?
>
> Dear FSL experts,
> I have done the preprocessing by FSL GUI and will further extract the time course for further analyis. I think the linear trend in the fMRI data will badly influence the results, but I didn't see FSL did "linear trend removal". Can FSL do "linear trend removal"? How can I do?
> Thanks in advance!
> Mingxia Zhang
>
>
>
>
>
> ------------------------------
>
> Date: Thu, 2 Jul 2015 17:27:54 +0200
> From: Andreas Bartsch <[log in to unmask]>
> Subject: Re: Issue with coregistration of spinal cord images
>
> PS:
> flirt -in B0.nii.gz -ref T2.nii.gz -init init.mat -dof 6 -nosearch -out
> B02T2 -v
> with the attached init.mat gives a reasonable result.
> You can refine that using Nudge(_gui) and approximately -6mm y-translation.
> However, your B0 has FatSat while your T2 is not fat-saturated.
> I >think< you better generate a mask arount the spinal cord and focus to
> register the myelon as best as you can (if that is what you venture to do).
> Cheers,
> Andreas
>
>
> Von: andreas <[log in to unmask]>
> Datum: Mittwoch, 1. Juli 2015 21:27
> An: FSL - FMRIB's Software Library <[log in to unmask]>
> Betreff: Re: [FSL] Issue with coregistration of spinal cord images
>
> Hi,
>
> well - these images differ in their initial orientation but you may use
> flirt -in B0.nii.gz -ref T2.nii.gz -applyxfm -usesqform -out B02T2 -v
> to get an reasonable initial starting point and go from there.
> Cheers,
> Andreas
>
> Von: FSL - FMRIB's Software Library <[log in to unmask]> on behalf of
> "Raphael F.C." <[log in to unmask]>
> Antworten an: FSL - FMRIB's Software Library <[log in to unmask]>
> Datum: Dienstag, 30. Juni 2015 02:40
> An: <[log in to unmask]>
> Betreff: Re: [FSL] Issue with coregistration of spinal cord images
>
> Hi, Andreas! Thank you very much for your prompt response!
>
> I tried doing as you suggested (to coregister the B0 image to the T2), but
> the output is still problematic.
>
> The link below give access to three files:
> 1) The B0 image;
> 2) The T2 image;
> 3) The output of the following:
> flirt -in B0.nii.gz -ref T2.nii -out B0coreg.nii -omat invol2refvol.mat -dof
> 6
> (Coregistration of B0 image to the T2).
>
> https://drive.google.com/open?id=0B-D-mq6qMD2Afk14MHh5alRfYXh3YzIyVEJFSDZvLV
> BJUUxDUWhHVUVVSFQxeVNWSEx5WFE&authuser=0
>
> Is there is anything I am doing wrong?
>
>
>
> 2015-06-29 2:40 GMT-03:00 Andreas Bartsch <[log in to unmask]>:
>> Hi,
>>
>> try the other way around, i.e. registering the lower-resolution B0 to the
>> higher resolution T2 image.
>> This >may< work if your B0 image is distortion corrected or has low
>> distortions (e.g., by segmented EPI aquisitions / RESOLVE).
>> If your B0 has lots of distortions you may try FNIRT but your chances are low
>> to get a good registration.
>> Cheers,
>> Andreas
>>
>> Von: FSL - FMRIB's Software Library <[log in to unmask]> on behalf of
>> "Raphael F.C." <[log in to unmask]>
>> Antworten an: FSL - FMRIB's Software Library <[log in to unmask]>
>> Datum: Montag, 29. Juni 2015 05:15
>> An: <[log in to unmask]>
>> Betreff: [FSL] Issue with coregistration of spinal cord images
>>
>> Dear all,
>>
>> I am trying to corregister the B0 image from a spinal cord DTI with a T2
>> structural image.
>>
>> I first tried using the following code:
>> flirt -in T2.nii -ref B0.nii.gz -out T2correg.nii -omat invol2refvol.mat
>> -dof 6
>> But the output image was a weird pyramid in the corner of the volume.
>> Then I tried including the cost function, because maybe the weight could
>> enhance the result:
>> flirt -in T2.nii -ref B0.nii.gz -out T2correg.nii -omat invol2refvol.mat
>> -dof 6
>> But I got just one peace of the image as the output.
>>
>> I am sorry for such a basic issue, but are the codelines above ok? Is there
>> any limitation which would forbid to use flirt with non-brain images?
>>
>> Any lights on how to make it work? I would appreciate very much any help.
>>
>> Thank you very much,
>> --
>> Raphael F. Casseb
>> Medical Physicist, Ph.D. Student
>> Medical Physics Lab - State University of Campinas
>> Contact: +55 19 3521-8246 <tel:%2B55%2019%203521-8246>
>
>
>
> --
> Raphael F. Casseb
> Medical Physicist, Ph.D. Student
> Medical Physics Lab - State University of Campinas
> Contact: +55 19 3521-8246
>
>
> ------------------------------
>
> Date: Thu, 2 Jul 2015 11:28:25 -0400
> From: Sarah Izen <[log in to unmask]>
> Subject: Re: Melodic and Resting State
>
> Hi Stephen,
>
> I ran everything with the set of 20 resting state networks, but then
> noticed that there's also a file available with only the ten most well
> matched networks. In your opinion, does it make a difference which set I
> use? If I ran everything with the set of 20, could I just focus on the 10
> well-matched networks in my interpretation of the results, or does it make
> more sense to always only use the set of 10?
>
> Thanks,
> Sarah
>
>> On Wed, Jun 3, 2015 at 11:04 AM, Sarah Izen <[log in to unmask]> wrote:
>>
>> Thanks very much!
>>
>> On Wed, Jun 3, 2015 at 8:22 AM, Stephen Smith <[log in to unmask]>
>> wrote:
>>
>>> Hi - I would use the RSN20 - they are higher intrinsic resolution /
>>> detail than the BrainMap maps.
>>>
>>> Yes ideally if you can use within-run ICA to denoise then that's a good
>>> thing to do before dual-regression.
>>>
>>> Cheers
>>>
>>>
>>>
>>> On 2 Jun 2015, at 09:58, Sarah Izen <[log in to unmask]> wrote:
>>>
>>> Hi Everyone,
>>>
>>> I'm using FSL to analyze resting state functional connectivity. After
>>> performing melodic and determining that my networks were not well defined,
>>> even after trying different numbers of components, I've decided to use
>>> predefined networks for my analysis. I have been looking at the BrainMap 20
>>> database and see there are two sets of data - the BrainMap 20 and the
>>> Resting State Network 20. From what I can gather, these two sets of
>>> networks have been found to be quite similar. So, is there any reason to
>>> choose one over the other for my analysis? Is one more widely used or more
>>> valid for any reason?
>>>
>>> Secondly, I just want to make sure it is correct that I am running
>>> melodic in order to denoise prior to conducting my dual regression instead
>>> of just inputting raw resting state data. Please advise.
>>>
>>> Thanks!
>>> Sarah
>>>
>>>
>>>
>>>
>>> ---------------------------------------------------------------------------
>>> Stephen M. Smith, Professor of Biomedical Engineering
>>> Associate Director, Oxford University FMRIB Centre
>>>
>>> FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
>>> +44 (0) 1865 222726 (fax 222717)
>>> [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
>>>
>>> ---------------------------------------------------------------------------
>>>
>>> Stop the cultural destruction of Tibet <http://smithinks.net>
>
> ------------------------------
>
> Date: Thu, 2 Jul 2015 16:31:09 +0100
> From: Olga Boukrina <[log in to unmask]>
> Subject: Re: ASL partial volume correction
>
> Thank you very much! Adding --pvcorr was the missing step.
>
> Olga
>
> ------------------------------
>
> Date: Thu, 2 Jul 2015 16:12:15 +0000
> From: Duncan Mortimer <[log in to unmask]>
> Subject: FSL website down until further notice
>
> Hi,
>
> We are experiencing some problems with the FSL webserver at the moment and are working to get this resolved. I don't expect this to be working again before mid-day on Friday (UK time). Sorry for any inconvenience this causes.
>
> Regards,
>
> Duncan
> --
> Duncan Mortimer
> Computing Officer, FMRIB Centre, University of Oxford
> John Radcliffe Hospital, Oxford, UK
>
> ------------------------------
>
> Date: Thu, 2 Jul 2015 16:13:57 +0000
> From: "RICHARDS, JOHN" <[log in to unmask]>
> Subject: Re: FSLVBM with pediatric population
>
> 1. One issue may be the different size head, different topography of the pediatric heads, or different GM/WM values. Also, the bet may be “tuned” to older brain voxel values.
>
> 2. I have a set of average MRI templates for pediatric populations. We substitute an age-appropriate MRI template (head, brain, gm/wm priors) for the default FSL brain (usually the MNI 152). I use a modified version of the FSL pipeline (register participant head to template; inverse transform the template brain to participant; expand the transformed brain by 2-3 voxels; mask, then BET). We have pretty much 100% success with this with 7 yr olds. For younger participants (e.g., infants, preschool) or for non-standard participants (e.g., Chinese children), an age-appropriate MRI template is essential.
>
>
> See http://jerlab.psych.sc.edu/NeurodevelopmentalMRIDatabase/
> Richards, J.E., Sanchez, C., Phillips-Meek, M., & Xie, W. (2015) <http://jerlab.psych.sc.edu/hdeegerp/PWRPublications/abstracts.php?&TablePrimaryKey=410>A database of age-appropriate average MRI templates. Neuroimage, doi:10.1016/j.neuroimage.2015.04.055
>
>
> And my www site for other references to this.
>
> John
>
> ***********************************************
> John E. Richards
> Carolina Distinguished Professor
> Department of Psychology
> University of South Carolina
> Columbia, SC 29208
> Dept Phone: 803 777 2079
> Fax: 803 777 9558
> Email: [log in to unmask] <applewebdata:[log in to unmask]>
> HTTP: jerlab.psych.sc.edu
> *************************************************
>
>
>
>
>
>
>
>
>> On 7/1/15, 10:07 AM, "FSL - FMRIB's Software Library on behalf of Rachel Ellenbogen" <[log in to unmask] on behalf of [log in to unmask]> wrote:
>>
>> Hello,
>>
>> I am trying to us FSL to do VBM analysis on a population of pediatric (ages 7-17) MRIs. I have been running into a lot of issues with the first BET step. I have tried adjusting the parameters (f, g, c) and have used robust, neck, and eye removal options but haven't had much success, especially in the MRI's for the younger kids. The program often cuts of large portions of the frontal lobe. Has anyone had success with BET/extracting skull from a pediatric population? I am thankful for any suggestions or work arounds!
>>
>> Thank you!
>>
>> Rachel
>
> ------------------------------
>
> Date: Thu, 2 Jul 2015 17:14:41 +0100
> From: Will Graves <[log in to unmask]>
> Subject: Assistant Professor position (tenure-track)
>
> The Department of Psychology at Rutgers University-Newark anticipates hiring at the assistant professor (tenure track) level. We seek applications from individuals with specializations in areas of Psychology and Neuroscience that primarily use fMRI as a methodology. Examples of research areas of special interest include social, developmental and cognitive neuroscience.
> Applicants that incorporate brain connectivity concepts in their research are encouraged to apply. Applicants will have access to the new Rutgers University Brain Imaging Center (RUBIC; Siemens 3T Trio). This position requires a Ph.D. in Psychology, Neuroscience, or related field.
> Highest priority will be given to applicants who demonstrate excellence in teaching at the graduate and undergraduate levels, provide research mentorship to students, and have research programs that can be supported by external funding. Applicants should submit a CV, statement of research and teaching interests, 3 top pre/re-prints and 3 letters of recommendation to: [log in to unmask] We will give priority to applications received by October 15th but will continue the search until the position is filled. Diversity Mission: Rutgers/Newark is the most ethnically diverse research-oriented campus in America (16 consecutive years, US News & World Report). Applicants may include a brief statement in their cover letter explaining how their membership in our department will advance the University commitment to diversity. All qualified applicants will receive consideration for employment without regard to race, color, religion, sex, sexual orientation, gender identity or expression, national origin, disability, protected veteran status or any other classification protected by law. Rutgers-Newark is an Affirmative Action/Equal Opportunity Employer and actively encourages applications from minorities, women, and other underrepresented groups.
>
> ------------------------------
>
> Date: Thu, 2 Jul 2015 22:29:55 +0100
> From: Will Graves <[log in to unmask]>
> Subject: Research Assistant/Lab Manager position
>
> A research assistant/lab manager position is available in the Language Behavior and Brain Imaging Lab (http://lbbil.rutgers.edu/) at Rutgers University in Newark, New Jersey. Much of our research is devoted to the cognitive neuroscience of reading, with potential application to reading disorders. Other aspects of brain and language studied in the lab include concept formation and speech production. Research is performed using a variety of techniques such as functional magnetic resonance imaging (fMRI), diffusion tensor imaging (DTI), behavioral responses, gene-brain correlations, and magnetoencephalography (MEG).
>
> Responsibilities will include data collection from human research participants in both a purely behavioral and functional brain imaging setting, contacting and scheduling research participants, managing institutional review board (IRB) protocols, and data analysis.
>
> Requirements for a successful applicant include spoken and written proficiency in English, a minimum of a bachelor-level degree (e.g., BA or BS), preferably in psychology, neuroscience, computer science, engineering, biology, or a related field, and willingness to make a 2-year commitment. Preference will be given to applicants who have experience in cognitive neuroscience research with human participants, are proficient with the linux computing environment, have used experiment delivery and data acquisition software such as E-prime, and can program in a scripting language such as Matlab or Python.
>
> Rutgers is the state university of New Jersey, and its Newark campus is in the state’s largest city. Newark is undergoing a renaissance of its own and is only minutes from Manhattan by train. Applications will be reviewed as they are received, with a deadline of September 15th. Please email a resume or CV and contact information for 3 references to [log in to unmask]
>
> ------------------------------
>
> End of FSL Digest - 1 Jul 2015 to 2 Jul 2015 (#2015-2)
> ******************************************************
|
