Re: [HCP-Users] MELODIC denoising vs. released ICA-FIX datasets

[Stephen Smith]

Hi - it sounds like maybe it's working fine within the limits of slight 
differences in mathematical precision between the C++ vs matlab parts of the 
processing - so the main question would be - have you looked in a viewer at the 
difference image - e.g. are the voxels with large differences isolated or eg at 
the edge of the brain?

Cheers.



> On 28 Sep 2016, at 02:49, Glasser, Matthew  wrote:
> 
> I think this is more of a question for the FSL list, but I don’t know fsl_glm 
> well enough to say if what you are doing is equivalent or not.
> 
> Peace,
> 
> Matt.
> 
> From: "Ely, Benjamin" >
> Date: Tuesday, September 27, 2016 at 7:20 PM
> To: Matt Glasser >, "Burgess, 
> Gregory" >, 
> "HCP-Users@humanconnectome.org " 
> >
> Subject: Re: [HCP-Users] MELODIC denoising vs. released ICA-FIX datasets
> 
> Hi Matt and Greg,
> 
> Thanks for the feedback! I've looked at the various fix .m files from the 
> current release; based on fix_3_clean.m, I tried the following for a single 
> resting-state run:
> 
> # highpass filter; sigma of 1000.08 = FWHM of 2355 per Smith et al 2013 
> NeuroImage, also consistent with comments in the fix_3_clean.m script
> fslmaths rfMRI_REST1_LR.nii.gz -bptf 1000.08 -1 REST1LR_bp
> 
> # format movement parameters (manually corrected header after paste step, not 
> shown)
> Text2Vest Movement_Regressors.txt Movement_Regressors.mat
> Text2Vest Movement_Regressors.txt Movement_Regressors_dt.mat
> paste Movement_Regressors.mat Movement_Regressors_dt.mat > 
> Movement_Regressors_all.mat
> 
> # regress movement parameters out of timeseries and re-add mean
> fsl_glm -i REST1LR_bp.nii.gz -d Movement_Regressors_all.mat 
> --out_res=REST1LR_bp_mc_demeaned.nii.gz --demean 
> fslmaths REST1LR_bp.nii.gz -Tmean REST1LR_bp_mean
> fslmaths REST1LR_bp_mc_demeaned.nii.gz -add REST1LR_bp_mean.nii.gz 
> REST1LR_bp_mc
> 
> # regress movement parameters out of melodic mix
> 
> fsl_glm -i filtered_func_data.ica/melodic_mix -d Movement_Regressors_all.mat 
> --out_res=melodic_mix_mc --demean
> 
> # regress unique variance from bad components (taken from .fix file) out of 
> timeseries
> 
> fsl_regfilt -i REST1LR_bp_mc.nii.gz -d melodic_mix_mc -o 
> REST1LR_bp_mc_softICA -f "1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 
> 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 28, 29, 30, 31, 32, 33, 34, 36, 
> 38, 39, 40, 41, 42, 43, 44, 45, 46, 49, 50, 51, 52, 53, 54, 55, 56, 57, 61, 
> 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 81, 82, 83, 84" 
> 
> # compare against HCP's released FIX-denoised file
> fslmaths rfMRI_REST1_LR_hp2000_clean -sub REST1LR_bp_mc_softICA 
> diff_REST1LR_bp_mc_softICA
> 
> Visual inspection and fslstats indicate reasonably good agreement between my 
> denoised file and the HCP's denoised file; the mean difference is about 0.84 
> units (compared to a mean signal intensity of around 10,000), and the 
> "robust" range of the difference is about +/- 72 units. More worryingly, 
> though, the maximum difference is around 2000 units, and around 6000 voxels 
> show differences greater than 500 units, so I'm not sure machine precision 
> can account for the differences.
> 
> Does the above denoising scheme seem consistent with what FIX is doing? I 
> plan to use FIX going forward, rather than trying to replicate it using the 
> FSL command-line, but I'd like to understand any discrepancies between the 
> two. 
> 
> Thanks again,
> -Ely
> 
>  
> 
> The materials in this message are private and may contain Protected 
> Healthcare Information or other information of a sensitive nature. If you are 
> not the intended recipient, be advised that any unauthorized use, disclosure, 
> copying or the taking of any action in reliance on the contents of this 
> information is strictly prohibited. If you have received this email in error, 
> please immediately notify the sender via telephone or return mail.
> 
> ___
> HCP-Users mailing list
> HCP-Users@humanconnectome.org 
> http://lists.humanconnectome.org/mailman/listinfo/hcp-users 
> 

---
Stephen M. Smith, Professor of Biomedical Engineering
Head of Analysis,  Oxford University FMRIB Centre

FMRIB, JR Hospital, Headington, Oxford  OX3 9DU, UK
+44 (0) 1865 222726  (fax 222717)
st...@fmrib.ox.ac.ukhttp://www.fmrib.ox.ac.uk/~steve 

---

Stop the cultural destruction of Tibet 

Re: [HCP-Users] MELODIC denoising vs. released ICA-FIX datasets

[Glasser, Matthew]

I think this is more of a question for the FSL list, but I don't know fsl_glm 
well enough to say if what you are doing is equivalent or not.

Peace,

Matt.

From: "Ely, Benjamin" >
Date: Tuesday, September 27, 2016 at 7:20 PM
To: Matt Glasser >, "Burgess, 
Gregory" >, 
"HCP-Users@humanconnectome.org" 
>
Subject: Re: [HCP-Users] MELODIC denoising vs. released ICA-FIX datasets

Hi Matt and Greg,

Thanks for the feedback! I've looked at the various fix .m files from the 
current release; based on fix_3_clean.m, I tried the following for a single 
resting-state run:


# highpass filter; sigma of 1000.08 = FWHM of 2355 per Smith et al 2013 
NeuroImage, also consistent with comments in the fix_3_clean.m script

fslmaths rfMRI_REST1_LR.nii.gz -bptf 1000.08 -1 REST1LR_bp

# format movement parameters (manually corrected header after paste step, not 
shown)
Text2Vest Movement_Regressors.txt Movement_Regressors.mat
Text2Vest Movement_Regressors.txt Movement_Regressors_dt.mat

paste Movement_Regressors.mat Movement_Regressors_dt.mat > 
Movement_Regressors_all.mat

# regress movement parameters out of timeseries and re-add mean

fsl_glm -i REST1LR_bp.nii.gz -d Movement_Regressors_all.mat 
--out_res=REST1LR_bp_mc_demeaned.nii.gz --demean

fslmaths REST1LR_bp.nii.gz -Tmean REST1LR_bp_mean

fslmaths REST1LR_bp_mc_demeaned.nii.gz -add REST1LR_bp_mean.nii.gz REST1LR_bp_mc


# regress movement parameters out of melodic mix

fsl_glm -i filtered_func_data.ica/melodic_mix -d Movement_Regressors_all.mat 
--out_res=melodic_mix_mc --demean


# regress unique variance from bad components (taken from .fix file) out of 
timeseries

fsl_regfilt -i REST1LR_bp_mc.nii.gz -d melodic_mix_mc -o REST1LR_bp_mc_softICA 
-f "1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 
22, 23, 24, 25, 26, 28, 29, 30, 31, 32, 33, 34, 36, 38, 39, 40, 41, 42, 43, 44, 
45, 46, 49, 50, 51, 52, 53, 54, 55, 56, 57, 61, 65, 66, 67, 68, 69, 70, 71, 72, 
73, 74, 75, 76, 77, 78, 79, 81, 82, 83, 84"


# compare against HCP's released FIX-denoised file

fslmaths rfMRI_REST1_LR_hp2000_clean -sub REST1LR_bp_mc_softICA 
diff_REST1LR_bp_mc_softICA


Visual inspection and fslstats indicate reasonably good agreement between my 
denoised file and the HCP's denoised file; the mean difference is about 0.84 
units (compared to a mean signal intensity of around 10,000), and the "robust" 
range of the difference is about +/- 72 units. More worryingly, though, the 
maximum difference is around 2000 units, and around 6000 voxels show 
differences greater than 500 units, so I'm not sure machine precision can 
account for the differences.


Does the above denoising scheme seem consistent with what FIX is doing? I plan 
to use FIX going forward, rather than trying to replicate it using the FSL 
command-line, but I'd like to understand any discrepancies between the two.


Thanks again,

-Ely


The materials in this message are private and may contain Protected Healthcare 
Information or other information of a sensitive nature. If you are not the 
intended recipient, be advised that any unauthorized use, disclosure, copying 
or the taking of any action in reliance on the contents of this information is 
strictly prohibited. If you have received this email in error, please 
immediately notify the sender via telephone or return mail.

___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Re: [HCP-Users] MELODIC denoising vs. released ICA-FIX datasets

[Ely, Benjamin]

Hi Matt and Greg,

Thanks for the feedback! I've looked at the various fix .m files from the 
current release; based on fix_3_clean.m, I tried the following for a single 
resting-state run:


# highpass filter; sigma of 1000.08 = FWHM of 2355 per Smith et al 2013 
NeuroImage, also consistent with comments in the fix_3_clean.m script

fslmaths rfMRI_REST1_LR.nii.gz -bptf 1000.08 -1 REST1LR_bp

# format movement parameters (manually corrected header after paste step, not 
shown)
Text2Vest Movement_Regressors.txt Movement_Regressors.mat
Text2Vest Movement_Regressors.txt Movement_Regressors_dt.mat

paste Movement_Regressors.mat Movement_Regressors_dt.mat > 
Movement_Regressors_all.mat

# regress movement parameters out of timeseries and re-add mean

fsl_glm -i REST1LR_bp.nii.gz -d Movement_Regressors_all.mat 
--out_res=REST1LR_bp_mc_demeaned.nii.gz --demean

fslmaths REST1LR_bp.nii.gz -Tmean REST1LR_bp_mean

fslmaths REST1LR_bp_mc_demeaned.nii.gz -add REST1LR_bp_mean.nii.gz REST1LR_bp_mc


# regress movement parameters out of melodic mix

fsl_glm -i filtered_func_data.ica/melodic_mix -d Movement_Regressors_all.mat 
--out_res=melodic_mix_mc --demean


# regress unique variance from bad components (taken from .fix file) out of 
timeseries

fsl_regfilt -i REST1LR_bp_mc.nii.gz -d melodic_mix_mc -o REST1LR_bp_mc_softICA 
-f "1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 
22, 23, 24, 25, 26, 28, 29, 30, 31, 32, 33, 34, 36, 38, 39, 40, 41, 42, 43, 44, 
45, 46, 49, 50, 51, 52, 53, 54, 55, 56, 57, 61, 65, 66, 67, 68, 69, 70, 71, 72, 
73, 74, 75, 76, 77, 78, 79, 81, 82, 83, 84"


# compare against HCP's released FIX-denoised file

fslmaths rfMRI_REST1_LR_hp2000_clean -sub REST1LR_bp_mc_softICA 
diff_REST1LR_bp_mc_softICA


Visual inspection and fslstats indicate reasonably good agreement between my 
denoised file and the HCP's denoised file; the mean difference is about 0.84 
units (compared to a mean signal intensity of around 10,000), and the "robust" 
range of the difference is about +/- 72 units. More worryingly, though, the 
maximum difference is around 2000 units, and around 6000 voxels show 
differences greater than 500 units, so I'm not sure machine precision can 
account for the differences.


Does the above denoising scheme seem consistent with what FIX is doing? I plan 
to use FIX going forward, rather than trying to replicate it using the FSL 
command-line, but I'd like to understand any discrepancies between the two.


Thanks again,

-Ely

___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Re: [HCP-Users] ROIs and Betas from Cifti Data

[Timothy Coalson]

Actually, -cifti-math will output a dlabel file with an improper label
table, which will look empty (it is not a proper dlabel file, but it does
have the ROI you want in it).  You should then use -cifti-change-mapping
with -scalar on that output to make it into a viewable ROI file (you could
also do this on the restricted dlabel file before -cifti-math).

Basically, there are several ways to go about the task, and you took a
different path than I expected.  By starting with -cifti-label-to-roi, you
can avoid the need to manually edit a text file for the purpose of
extracting the labels you want (and it also avoids doing math on label
files, which has some rough edges).

Tim


On Tue, Sep 27, 2016 at 5:29 PM, Timothy Coalson  wrote:

> If you already have a dlabel file that excludes the things you aren't
> interested in, you don't need to bother with the -cifti-reduce I
> described.  Since "unlabeled" is generally represented as 0, you can use
> -cifti-math with 'x > 0' to generate a combined binary ROI from that
> restricted dlabel file.
>
> Tim
>
>
> On Tue, Sep 27, 2016 at 5:26 PM, Michael F.W. Dreyfuss <
> mid2...@med.cornell.edu> wrote:
>
>> The command I was using was :
>>
>> wb_command -cifti-reduce Brodmann_Bilateral_IFG.dlabel.nii MAX
>> Brodmann_Bilateral_IFG_Merged.dscalar.nii
>>
>> Where Brodmann_Bilateral_IFG.dlabel.nii contains two labels.
>>
>> Again, this command does not seem to the production of a label file
>> directly.
>>
>> Thanks
>>
>> On Sep 27, 2016, at 6:20 PM, Michael F.W. Dreyfuss <
>> mid2...@med.cornell.edu> wrote:
>>
>> I’m sorry, but this is still not clear to me. It seems this function is
>> meant to perform an operation to produce a scalar file, but what is
>> produced when I use MAX for example is just a scalar file where the value
>> for the key each labeled region had in the label file is not the value in
>> that scalar file. so BA44 was 65, BA45 was 66, and now there are two
>> regions with values 65 and 66.
>>
>> What I am trying to do is to merge regions together into one region as a
>> label file to use as a single mask. Is this possible with this function?
>>
>> I’m sorry if I’m missing something here.
>>
>> Thank you,
>> Michael
>>
>>
>> On Sep 27, 2016, at 5:57 PM, Timothy Coalson  wrote:
>>
>> http://www.humanconnectome.org/software/workbench-command.
>> php?function=-cifti-reduce
>> 
>>
>> It is a simple reduction across columns (by default, anyway) in the
>> file.  For these kinds of files, that means it isn't reducing across
>> space.  MAX is one of the options it can calculate, and for binary ROIs,
>> that is equivalent to boolean OR, which is what we want.
>>
>> The commands are generally written to perform small, specific actions,
>> such that they could be suitable for multiple tasks.
>>
>> Tim
>>
>>
>> On Tue, Sep 27, 2016 at 4:49 PM, Michael F.W. Dreyfuss <
>> mid2...@med.cornell.edu> wrote:
>>
>>> How would this work with -cifti-reduce? If I have a cifti with BA44 and
>>> BA45 for example and I want those two regions to be merged into one, how
>>> would -cifti-reduce do that? Wouldn’t it be more suited to calculating
>>> something (i.e. mean, mode, etc.) within each of those parcels?
>>>
>>> Thank you
>>>
>>> On Sep 27, 2016, at 5:40 PM, Timothy Coalson  wrote:
>>>
>>> wb_shortcuts
>>>
>>>
>>>
>>
>>
>>
>

___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Re: [HCP-Users] ROIs and Betas from Cifti Data

[Timothy Coalson]

If you already have a dlabel file that excludes the things you aren't
interested in, you don't need to bother with the -cifti-reduce I
described.  Since "unlabeled" is generally represented as 0, you can use
-cifti-math with 'x > 0' to generate a combined binary ROI from that
restricted dlabel file.

Tim


On Tue, Sep 27, 2016 at 5:26 PM, Michael F.W. Dreyfuss <
mid2...@med.cornell.edu> wrote:

> The command I was using was :
>
> wb_command -cifti-reduce Brodmann_Bilateral_IFG.dlabel.nii MAX
> Brodmann_Bilateral_IFG_Merged.dscalar.nii
>
> Where Brodmann_Bilateral_IFG.dlabel.nii contains two labels.
>
> Again, this command does not seem to the production of a label file
> directly.
>
> Thanks
>
> On Sep 27, 2016, at 6:20 PM, Michael F.W. Dreyfuss <
> mid2...@med.cornell.edu> wrote:
>
> I’m sorry, but this is still not clear to me. It seems this function is
> meant to perform an operation to produce a scalar file, but what is
> produced when I use MAX for example is just a scalar file where the value
> for the key each labeled region had in the label file is not the value in
> that scalar file. so BA44 was 65, BA45 was 66, and now there are two
> regions with values 65 and 66.
>
> What I am trying to do is to merge regions together into one region as a
> label file to use as a single mask. Is this possible with this function?
>
> I’m sorry if I’m missing something here.
>
> Thank you,
> Michael
>
>
> On Sep 27, 2016, at 5:57 PM, Timothy Coalson  wrote:
>
> http://www.humanconnectome.org/software/workbench-
> command.php?function=-cifti-reduce
> 
>
> It is a simple reduction across columns (by default, anyway) in the file.
> For these kinds of files, that means it isn't reducing across space.  MAX
> is one of the options it can calculate, and for binary ROIs, that is
> equivalent to boolean OR, which is what we want.
>
> The commands are generally written to perform small, specific actions,
> such that they could be suitable for multiple tasks.
>
> Tim
>
>
> On Tue, Sep 27, 2016 at 4:49 PM, Michael F.W. Dreyfuss <
> mid2...@med.cornell.edu> wrote:
>
>> How would this work with -cifti-reduce? If I have a cifti with BA44 and
>> BA45 for example and I want those two regions to be merged into one, how
>> would -cifti-reduce do that? Wouldn’t it be more suited to calculating
>> something (i.e. mean, mode, etc.) within each of those parcels?
>>
>> Thank you
>>
>> On Sep 27, 2016, at 5:40 PM, Timothy Coalson  wrote:
>>
>> wb_shortcuts
>>
>>
>>
>
>
>

___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Re: [HCP-Users] ROIs and Betas from Cifti Data

[Timothy Coalson]

To me, a mask is a binary ROI (1 where the stuff you want is, 0
elsewhere).  This is quite different from a label file (lots of different
integers).  I did not mean for you to run -cifti-reduce on a dlabel file.

I was suggesting that you use -cifti-label-to-roi to make a binary ROI from
each label you are interested in, use -cifti-merge (or wb_shortcuts
-cifti-concatenate) to put all those binary ROIs as separate maps in a
single dscalar file, then use -cifti-reduce on that multi-ROI dscalar file
to get a single map with the combined binary ROI.

To then separate the ROI per-hemisphere, you can use -cifti-separate (on
the binary ROI dscalar file).

Tim


On Tue, Sep 27, 2016 at 5:20 PM, Michael F.W. Dreyfuss <
mid2...@med.cornell.edu> wrote:

> I’m sorry, but this is still not clear to me. It seems this function is
> meant to perform an operation to produce a scalar file, but what is
> produced when I use MAX for example is just a scalar file where the value
> for the key each labeled region had in the label file is not the value in
> that scalar file. so BA44 was 65, BA45 was 66, and now there are two
> regions with values 65 and 66.
>
> What I am trying to do is to merge regions together into one region as a
> label file to use as a single mask. Is this possible with this function?
>
> I’m sorry if I’m missing something here.
>
> Thank you,
> Michael
>
>
>
> On Sep 27, 2016, at 5:57 PM, Timothy Coalson  wrote:
>
> http://www.humanconnectome.org/software/workbench-
> command.php?function=-cifti-reduce
> 
>
> It is a simple reduction across columns (by default, anyway) in the file.
> For these kinds of files, that means it isn't reducing across space.  MAX
> is one of the options it can calculate, and for binary ROIs, that is
> equivalent to boolean OR, which is what we want.
>
> The commands are generally written to perform small, specific actions,
> such that they could be suitable for multiple tasks.
>
> Tim
>
>
> On Tue, Sep 27, 2016 at 4:49 PM, Michael F.W. Dreyfuss <
> mid2...@med.cornell.edu> wrote:
>
>> How would this work with -cifti-reduce? If I have a cifti with BA44 and
>> BA45 for example and I want those two regions to be merged into one, how
>> would -cifti-reduce do that? Wouldn’t it be more suited to calculating
>> something (i.e. mean, mode, etc.) within each of those parcels?
>>
>> Thank you
>>
>> On Sep 27, 2016, at 5:40 PM, Timothy Coalson  wrote:
>>
>> wb_shortcuts
>>
>>
>>
>
>

___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Re: [HCP-Users] ROIs and Betas from Cifti Data

[Michael F.W. Dreyfuss]

The command I was using was :

wb_command -cifti-reduce Brodmann_Bilateral_IFG.dlabel.nii MAX 
Brodmann_Bilateral_IFG_Merged.dscalar.nii

Where Brodmann_Bilateral_IFG.dlabel.nii contains two labels.

Again, this command does not seem to the production of a label file directly.

Thanks

On Sep 27, 2016, at 6:20 PM, Michael F.W. Dreyfuss 
> wrote:

I’m sorry, but this is still not clear to me. It seems this function is meant 
to perform an operation to produce a scalar file, but what is produced when I 
use MAX for example is just a scalar file where the value for the key each 
labeled region had in the label file is not the value in that scalar file. so 
BA44 was 65, BA45 was 66, and now there are two regions with values 65 and 66.

What I am trying to do is to merge regions together into one region as a label 
file to use as a single mask. Is this possible with this function?

I’m sorry if I’m missing something here.

Thank you,
Michael


On Sep 27, 2016, at 5:57 PM, Timothy Coalson 
> wrote:

http://www.humanconnectome.org/software/workbench-command.php?function=-cifti-reduce

It is a simple reduction across columns (by default, anyway) in the file.  For 
these kinds of files, that means it isn't reducing across space.  MAX is one of 
the options it can calculate, and for binary ROIs, that is equivalent to 
boolean OR, which is what we want.

The commands are generally written to perform small, specific actions, such 
that they could be suitable for multiple tasks.

Tim


On Tue, Sep 27, 2016 at 4:49 PM, Michael F.W. Dreyfuss 
> wrote:
How would this work with -cifti-reduce? If I have a cifti with BA44 and BA45 
for example and I want those two regions to be merged into one, how would 
-cifti-reduce do that? Wouldn’t it be more suited to calculating something 
(i.e. mean, mode, etc.) within each of those parcels?

Thank you

On Sep 27, 2016, at 5:40 PM, Timothy Coalson 
> wrote:

wb_shortcuts





___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Re: [HCP-Users] ROIs and Betas from Cifti Data

[Timothy Coalson]

The easier way is to use -cifti-merge (or wb_shortcuts -cifti-concatenate
for simpler syntax), then -cifti-reduce.  It is possible to do with
-cifti-math, but not as convenient.

Tim


On Tue, Sep 27, 2016 at 4:37 PM, Michael F.W. Dreyfuss <
mid2...@med.cornell.edu> wrote:

> Thank you, that works. Then if I want multiple parcels to be collapsed
> into a single ROI to make a mask spanning several ROIs that are not
> considered separate keys, which command would do that.
>
> Thank you,
> Michael
>
>
> On Sep 27, 2016, at 5:26 PM, Timothy Coalson  wrote:
>
> The easiest way to deal with this is probably to use -cifti-separate to
> get them as gifti label files, which are each single-hemisphere.  The use
> of the same key/name for both hemispheres was an unfortunate choice, which
> we intended to avoid in the future.
>
> Tim
>
>
> On Tue, Sep 27, 2016 at 4:15 PM, Michael F.W. Dreyfuss <
> mid2...@med.cornell.edu> wrote:
>
>> Thank you,
>>
>>
>> The problem was that the computer I was using before had an older version
>> of workbench. It works on my current computer.
>>
>>
>> As far as using
>>
>> -cifti-label-to-roi
>>
>>
>> How would that work with files like Human.Brodmann09.32k_fs_LR.dlabel.nii
>> where the ROIs in both hemispheres are under the same key? If I want to
>> look at BA44 in the right hemisphere only, for example, how can I separate
>> it from BA44 in the left hemisphere to look at them separately?
>>
>> Thank you,
>> Michael
>>
>>
>> --
>> *From:* Timothy Coalson 
>> *Sent:* Tuesday, September 27, 2016 4:10:15 PM
>> *To:* Michael F.W. Dreyfuss
>> *Cc:* Burgess, Gregory; hcp-users@humanconnectome.org
>> *Subject:* Re: [HCP-Users] ROIs and Betas from Cifti Data
>>
>> The surface structures in -cifti-separate are much coarser, they
>> represent the cifti file organization, not the parcellation areas.  You
>> want CORTEX_LEFT and CORTEX_RIGHT for that command to work.  However, the
>> outputs are gifti files, not cifti files - it is a format conversion
>> command, and probably not what you want to do.
>>
>> Instead of separating to gifti files, you should use -cifti-label-to-roi
>> on the dlabel file to get the areas by name as separate ROIs (as dscalar
>> files), and then use -cifti-merge and -cifti-reduce to combine them back
>> into one larger ROI.
>>
>> Tim
>>
>>
>> On Tue, Sep 27, 2016 at 10:36 AM, Michael F.W. Dreyfuss <
>> mid2...@med.cornell.edu> wrote:
>>
>>> I can get separate the volumetric part of cifti files with a command
>>> like:
>>>
>>> wb_command -cifti-separate 
>>> ToyNogo_fdr_palm/ToyNogo_results_merged_tstat.dscalar.nii
>>> COLUMN -volume ACCUMBENS_RIGHT R_Acc_Beta.nii.gz -roi R_NAcc_ROI.nii.gz
>>>
>>> But when I try that on surface data I get the error:
>>> wb_command -cifti-separate ToyNogo_GlasserParcellation_FL
>>> OBS1_fdr_palm/ToyNogo_results_dat_tstat.pscalar.nii COLUMN -label
>>> R_IFSp_ROI ToyNogo_R_IFSp_Beta.pscalar.ni
>>> i
>>> -roi R_IFSp_ROI.plabel.nii
>>>
>>> ERROR: unrecognized structure type
>>>
>>> The problem seems to be in recognizing the structures name (i.e.
>>> R_IFSp_ROI). Do you know how I can reference a specific structure or
>>> structures with a command from a label file, or is there another comparable
>>> command you would suggest for isolating structures from a label file?
>>>
>>> Thank you,
>>> Michael
>>>
>>> On Sep 27, 2016, at 10:41 AM, Burgess, Gregory 
>>> wrote:
>>>
>>> HCP did not have a task that was geared toward response inhibition.
>>> Furthermore, although it’s alluring to believe that a single parcel will
>>> encapsulate all of response inhibition, it’s doubtful.
>>>
>>> Why not select a set of parcels in and near IFG, and correct for
>>> multiple comparisons? Use a meta-analysis as your guide (
>>> https://urldefense.proofpoint.com/v2/url?u=http-3A__neurosy
>>> nth.org_analyses_terms_response-2520inhibition_=DQIGaQ=
>>> lb62iw4YL4RFalcE2hQUQealT9-RXrryqt9KZX2qu2s=rPclmYysc_
>>> z1plf99IoNsmxWf1JolkKMmL6bXnYFSwg=Pb-c1m96L6xixMU-j16J-
>>> GkC6tGV5lAdGNKy-Tj1ht8=DxHCajoqQSf0nIWSDbUHyB6J7F2eax1NswvaNpprHaU=
>>> ). Your statistical power should still benefit from the parcellated
>>> analysis and reduced number of multiple comparisons, relative to a
>>> whole-brain analysis.
>>>
>>> --Greg
>>>
>>> 
>>> Greg Burgess, Ph.D.
>>> Staff Scientist, Human Connectome Project
>>> Washington University School of Medicine
>>> Department of Psychiatry
>>> Phone: 314-362-7864
>>> Email: gburg...@wustl.edu
>>>
>>> On Sep 27, 2016, at 9:26 AM, Michael F.W. Dreyfuss <
>>> mid2...@med.cornell.edu> wrote:
>>>
>>> I agree, that’s why I was 

Re: [HCP-Users] ROIs and Betas from Cifti Data

[Michael F.W. Dreyfuss]

Thank you, that works. Then if I want multiple parcels to be collapsed into a 
single ROI to make a mask spanning several ROIs that are not considered 
separate keys, which command would do that.

Thank you,
Michael

On Sep 27, 2016, at 5:26 PM, Timothy Coalson 
> wrote:

The easiest way to deal with this is probably to use -cifti-separate to get 
them as gifti label files, which are each single-hemisphere.  The use of the 
same key/name for both hemispheres was an unfortunate choice, which we intended 
to avoid in the future.

Tim


On Tue, Sep 27, 2016 at 4:15 PM, Michael F.W. Dreyfuss 
> wrote:

Thank you,


The problem was that the computer I was using before had an older version of 
workbench. It works on my current computer.


As far as using

-cifti-label-to-roi

How would that work with files like Human.Brodmann09.32k_fs_LR.dlabel.nii where 
the ROIs in both hemispheres are under the same key? If I want to look at BA44 
in the right hemisphere only, for example, how can I separate it from BA44 in 
the left hemisphere to look at them separately?

Thank you,
Michael



From: Timothy Coalson >
Sent: Tuesday, September 27, 2016 4:10:15 PM
To: Michael F.W. Dreyfuss
Cc: Burgess, Gregory; 
hcp-users@humanconnectome.org
Subject: Re: [HCP-Users] ROIs and Betas from Cifti Data

The surface structures in -cifti-separate are much coarser, they represent the 
cifti file organization, not the parcellation areas.  You want CORTEX_LEFT and 
CORTEX_RIGHT for that command to work.  However, the outputs are gifti files, 
not cifti files - it is a format conversion command, and probably not what you 
want to do.

Instead of separating to gifti files, you should use -cifti-label-to-roi on the 
dlabel file to get the areas by name as separate ROIs (as dscalar files), and 
then use -cifti-merge and -cifti-reduce to combine them back into one larger 
ROI.

Tim


On Tue, Sep 27, 2016 at 10:36 AM, Michael F.W. Dreyfuss 
> wrote:
I can get separate the volumetric part of cifti files with a command like:

wb_command -cifti-separate 
ToyNogo_fdr_palm/ToyNogo_results_merged_tstat.dscalar.nii COLUMN -volume 
ACCUMBENS_RIGHT R_Acc_Beta.nii.gz -roi R_NAcc_ROI.nii.gz

But when I try that on surface data I get the error:
wb_command -cifti-separate 
ToyNogo_GlasserParcellation_FLOBS1_fdr_palm/ToyNogo_results_dat_tstat.pscalar.nii
 COLUMN -label R_IFSp_ROI 
ToyNogo_R_IFSp_Beta.pscalar.nii
 -roi R_IFSp_ROI.plabel.nii

ERROR: unrecognized structure type

The problem seems to be in recognizing the structures name (i.e. R_IFSp_ROI). 
Do you know how I can reference a specific structure or structures with a 
command from a label file, or is there another comparable command you would 
suggest for isolating structures from a label file?

Thank you,
Michael

On Sep 27, 2016, at 10:41 AM, Burgess, Gregory 
> wrote:

HCP did not have a task that was geared toward response inhibition. 
Furthermore, although it’s alluring to believe that a single parcel will 
encapsulate all of response inhibition, it’s doubtful.

Why not select a set of parcels in and near IFG, and correct for multiple 
comparisons? Use a meta-analysis as your guide 
(https://urldefense.proofpoint.com/v2/url?u=http-3A__neurosynth.org_analyses_terms_response-2520inhibition_=DQIGaQ=lb62iw4YL4RFalcE2hQUQealT9-RXrryqt9KZX2qu2s=rPclmYysc_z1plf99IoNsmxWf1JolkKMmL6bXnYFSwg=Pb-c1m96L6xixMU-j16J-GkC6tGV5lAdGNKy-Tj1ht8=DxHCajoqQSf0nIWSDbUHyB6J7F2eax1NswvaNpprHaU=
 ). Your statistical power should still benefit from the parcellated analysis 
and reduced number of multiple comparisons, relative to a whole-brain analysis.

--Greg


Greg Burgess, Ph.D.
Staff Scientist, Human Connectome Project
Washington University School of Medicine
Department of Psychiatry
Phone: 314-362-7864
Email: gburg...@wustl.edu

On Sep 27, 2016, at 9:26 AM, Michael F.W. Dreyfuss 
> wrote:

I agree, that’s why I was checking in to see if there was a sub parcel you had 
identified as being involved in response inhibition from your tasks, such as 
flanker. There is a lot of background of IFG being involved in response 
inhibition, particularly on go/nogo tasks,, so I was wondering if you had any 
information on specifically where within your parcellation that may be most 
relevant.

On Sep 27, 2016, at 

Re: [HCP-Users] ROIs and Betas from Cifti Data

[Timothy Coalson]

The easiest way to deal with this is probably to use -cifti-separate to get
them as gifti label files, which are each single-hemisphere.  The use of
the same key/name for both hemispheres was an unfortunate choice, which we
intended to avoid in the future.

Tim


On Tue, Sep 27, 2016 at 4:15 PM, Michael F.W. Dreyfuss <
mid2...@med.cornell.edu> wrote:

> Thank you,
>
>
> The problem was that the computer I was using before had an older version
> of workbench. It works on my current computer.
>
>
> As far as using
>
> -cifti-label-to-roi
>
> How would that work with files like Human.Brodmann09.32k_fs_LR.dlabel.nii
> where the ROIs in both hemispheres are under the same key? If I want to
> look at BA44 in the right hemisphere only, for example, how can I separate
> it from BA44 in the left hemisphere to look at them separately?
>
> Thank you,
> Michael
>
>
> --
> *From:* Timothy Coalson 
> *Sent:* Tuesday, September 27, 2016 4:10:15 PM
> *To:* Michael F.W. Dreyfuss
> *Cc:* Burgess, Gregory; hcp-users@humanconnectome.org
> *Subject:* Re: [HCP-Users] ROIs and Betas from Cifti Data
>
> The surface structures in -cifti-separate are much coarser, they represent
> the cifti file organization, not the parcellation areas.  You want
> CORTEX_LEFT and CORTEX_RIGHT for that command to work.  However, the
> outputs are gifti files, not cifti files - it is a format conversion
> command, and probably not what you want to do.
>
> Instead of separating to gifti files, you should use -cifti-label-to-roi
> on the dlabel file to get the areas by name as separate ROIs (as dscalar
> files), and then use -cifti-merge and -cifti-reduce to combine them back
> into one larger ROI.
>
> Tim
>
>
> On Tue, Sep 27, 2016 at 10:36 AM, Michael F.W. Dreyfuss <
> mid2...@med.cornell.edu> wrote:
>
>> I can get separate the volumetric part of cifti files with a command
>> like:
>>
>> wb_command -cifti-separate 
>> ToyNogo_fdr_palm/ToyNogo_results_merged_tstat.dscalar.nii
>> COLUMN -volume ACCUMBENS_RIGHT R_Acc_Beta.nii.gz -roi R_NAcc_ROI.nii.gz
>>
>> But when I try that on surface data I get the error:
>> wb_command -cifti-separate ToyNogo_GlasserParcellation_FL
>> OBS1_fdr_palm/ToyNogo_results_dat_tstat.pscalar.nii COLUMN -label
>> R_IFSp_ROI ToyNogo_R_IFSp_Beta.pscalar.nii -roi R_IFSp_ROI.plabel.nii
>>
>> ERROR: unrecognized structure type
>>
>> The problem seems to be in recognizing the structures name (i.e.
>> R_IFSp_ROI). Do you know how I can reference a specific structure or
>> structures with a command from a label file, or is there another comparable
>> command you would suggest for isolating structures from a label file?
>>
>> Thank you,
>> Michael
>>
>> On Sep 27, 2016, at 10:41 AM, Burgess, Gregory 
>> wrote:
>>
>> HCP did not have a task that was geared toward response inhibition.
>> Furthermore, although it’s alluring to believe that a single parcel will
>> encapsulate all of response inhibition, it’s doubtful.
>>
>> Why not select a set of parcels in and near IFG, and correct for multiple
>> comparisons? Use a meta-analysis as your guide (
>> https://urldefense.proofpoint.com/v2/url?u=http-3A__
>> neurosynth.org_analyses_terms_response-2520inhibition_=
>> DQIGaQ=lb62iw4YL4RFalcE2hQUQealT9-RXrryqt9KZX2qu2s=
>> rPclmYysc_z1plf99IoNsmxWf1JolkKMmL6bXnYFSwg=Pb-
>> c1m96L6xixMU-j16J-GkC6tGV5lAdGNKy-Tj1ht8=DxHCajoqQSf0nIWSD
>> bUHyB6J7F2eax1NswvaNpprHaU= ). Your statistical power should still
>> benefit from the parcellated analysis and reduced number of multiple
>> comparisons, relative to a whole-brain analysis.
>>
>> --Greg
>>
>> 
>> Greg Burgess, Ph.D.
>> Staff Scientist, Human Connectome Project
>> Washington University School of Medicine
>> Department of Psychiatry
>> Phone: 314-362-7864
>> Email: gburg...@wustl.edu
>>
>> On Sep 27, 2016, at 9:26 AM, Michael F.W. Dreyfuss <
>> mid2...@med.cornell.edu> wrote:
>>
>> I agree, that’s why I was checking in to see if there was a sub parcel
>> you had identified as being involved in response inhibition from your
>> tasks, such as flanker. There is a lot of background of IFG being involved
>> in response inhibition, particularly on go/nogo tasks,, so I was wondering
>> if you had any information on specifically where within your parcellation
>> that may be most relevant.
>>
>> On Sep 27, 2016, at 9:57 AM, Harms, Michael  wrote:
>>
>> Just a reminder to be careful here to avoid issues of
>> circularity/double-dipping.  You indicated that you had a priori hypotheses
>> about IFG involvement, but that doesn’t allow you to then select a
>> particular IFG parcel based on your task activation map.  Ideally, you
>> would have selected your parcel(s) for analysis prior to ever
>> computing/viewing your activation map (unless what you showed was a map
>> from an independent, unrelated set of subjects).
>>
>>
>> 

Re: [HCP-Users] ROIs and Betas from Cifti Data

[Michael F.W. Dreyfuss]

Thank you,


The problem was that the computer I was using before had an older version of 
workbench. It works on my current computer.


As far as using

-cifti-label-to-roi

How would that work with files like Human.Brodmann09.32k_fs_LR.dlabel.nii where 
the ROIs in both hemispheres are under the same key? If I want to look at BA44 
in the right hemisphere only, for example, how can I separate it from BA44 in 
the left hemisphere to look at them separately?

Thank you,
Michael



From: Timothy Coalson 
Sent: Tuesday, September 27, 2016 4:10:15 PM
To: Michael F.W. Dreyfuss
Cc: Burgess, Gregory; hcp-users@humanconnectome.org
Subject: Re: [HCP-Users] ROIs and Betas from Cifti Data

The surface structures in -cifti-separate are much coarser, they represent the 
cifti file organization, not the parcellation areas.  You want CORTEX_LEFT and 
CORTEX_RIGHT for that command to work.  However, the outputs are gifti files, 
not cifti files - it is a format conversion command, and probably not what you 
want to do.

Instead of separating to gifti files, you should use -cifti-label-to-roi on the 
dlabel file to get the areas by name as separate ROIs (as dscalar files), and 
then use -cifti-merge and -cifti-reduce to combine them back into one larger 
ROI.

Tim


On Tue, Sep 27, 2016 at 10:36 AM, Michael F.W. Dreyfuss 
> wrote:
I can get separate the volumetric part of cifti files with a command like:

wb_command -cifti-separate 
ToyNogo_fdr_palm/ToyNogo_results_merged_tstat.dscalar.nii COLUMN -volume 
ACCUMBENS_RIGHT R_Acc_Beta.nii.gz -roi R_NAcc_ROI.nii.gz

But when I try that on surface data I get the error:
wb_command -cifti-separate 
ToyNogo_GlasserParcellation_FLOBS1_fdr_palm/ToyNogo_results_dat_tstat.pscalar.nii
 COLUMN -label R_IFSp_ROI ToyNogo_R_IFSp_Beta.pscalar.nii -roi 
R_IFSp_ROI.plabel.nii

ERROR: unrecognized structure type

The problem seems to be in recognizing the structures name (i.e. R_IFSp_ROI). 
Do you know how I can reference a specific structure or structures with a 
command from a label file, or is there another comparable command you would 
suggest for isolating structures from a label file?

Thank you,
Michael

On Sep 27, 2016, at 10:41 AM, Burgess, Gregory 
> wrote:

HCP did not have a task that was geared toward response inhibition. 
Furthermore, although it’s alluring to believe that a single parcel will 
encapsulate all of response inhibition, it’s doubtful.

Why not select a set of parcels in and near IFG, and correct for multiple 
comparisons? Use a meta-analysis as your guide 
(https://urldefense.proofpoint.com/v2/url?u=http-3A__neurosynth.org_analyses_terms_response-2520inhibition_=DQIGaQ=lb62iw4YL4RFalcE2hQUQealT9-RXrryqt9KZX2qu2s=rPclmYysc_z1plf99IoNsmxWf1JolkKMmL6bXnYFSwg=Pb-c1m96L6xixMU-j16J-GkC6tGV5lAdGNKy-Tj1ht8=DxHCajoqQSf0nIWSDbUHyB6J7F2eax1NswvaNpprHaU=
 ). Your statistical power should still benefit from the parcellated analysis 
and reduced number of multiple comparisons, relative to a whole-brain analysis.

--Greg


Greg Burgess, Ph.D.
Staff Scientist, Human Connectome Project
Washington University School of Medicine
Department of Psychiatry
Phone: 314-362-7864
Email: gburg...@wustl.edu

On Sep 27, 2016, at 9:26 AM, Michael F.W. Dreyfuss 
> wrote:

I agree, that’s why I was checking in to see if there was a sub parcel you had 
identified as being involved in response inhibition from your tasks, such as 
flanker. There is a lot of background of IFG being involved in response 
inhibition, particularly on go/nogo tasks,, so I was wondering if you had any 
information on specifically where within your parcellation that may be most 
relevant.

On Sep 27, 2016, at 9:57 AM, Harms, Michael 
> wrote:

Just a reminder to be careful here to avoid issues of 
circularity/double-dipping.  You indicated that you had a priori hypotheses 
about IFG involvement, but that doesn’t allow you to then select a particular 
IFG parcel based on your task activation map.  Ideally, you would have selected 
your parcel(s) for analysis prior to ever computing/viewing your activation map 
(unless what you showed was a map from an independent, unrelated set of 
subjects).


___
HCP-Users mailing list
HCP-Users@humanconnectome.org
https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.humanconnectome.org_mailman_listinfo_hcp-2Dusers=DQIGaQ=lb62iw4YL4RFalcE2hQUQealT9-RXrryqt9KZX2qu2s=rPclmYysc_z1plf99IoNsmxWf1JolkKMmL6bXnYFSwg=Pb-c1m96L6xixMU-j16J-GkC6tGV5lAdGNKy-Tj1ht8=Kd8JGIfyDD_MWOgH24WZFrQNLqIT2e2KIaFCaLTw3PM=




--Greg

Re: [HCP-Users] Label Import with -discard-others

[Timothy Coalson]

For that, you want -cifti-separate, splitting into hemispheres (resulting
in left and right label.gii files in this case) is exactly what it does.
See my response in the other thread for pulling individual areas out of a
parcellation and combining their ROIs.

Tim


On Tue, Sep 27, 2016 at 2:44 PM, Michael F.W. Dreyfuss <
mid2...@med.cornell.edu> wrote:

> In a related question, if I want to treat hemispheres separately, how can
> I create a simple hemispheric label file, or split a file like
> Human.Brodmann09.32k_fs_LR.dlabel.nii into two hemispheres to treat the
> Brodmann areas of each area separately?
>
>
> Essentially, all of my questions are directed to find out how I can easily
> separate and combine labels from existing files.
>
>
> Thank you,
>
> Michael
> --
> *From:* Michael F.W. Dreyfuss
> *Sent:* Tuesday, September 27, 2016 2:19:06 PM
> *To:* hcp-users@humanconnectome.org
> *Subject:* Label Import with -discard-others
>
>
> Hello, I have been trying to make a file containing the parcellations for 
> BA44 and BA45 using the Broadmann using the following command:
>
>
> wb_command -cifti-label-import Human.Brodmann09.32k_fs_LR.dlabel.nii 
> BrodmannLabels.txt Brodmann_IFG.dlabel.nii -discard-others
>
>
> Where BrodmannLabels.txt contains:
>
> BA44
>
> 66 100 255 100 255
>
> BA45
>
> 65 255 100 100 255
>
> It seems to recognize these labels and keeps their names on the file if I do 
> not use -discard-others, but when I use -discard-others the output is totally 
> blank, rather than just containing the selected parcels.
>
>
> Do you have a suggestions for how to do what I am trying to do?
>
>
> thank you,
>
> Michael
>
> ___
> HCP-Users mailing list
> HCP-Users@humanconnectome.org
> http://lists.humanconnectome.org/mailman/listinfo/hcp-users
>

___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Re: [HCP-Users] ROIs and Betas from Cifti Data

[Timothy Coalson]

The surface structures in -cifti-separate are much coarser, they represent
the cifti file organization, not the parcellation areas.  You want
CORTEX_LEFT and CORTEX_RIGHT for that command to work.  However, the
outputs are gifti files, not cifti files - it is a format conversion
command, and probably not what you want to do.

Instead of separating to gifti files, you should use -cifti-label-to-roi on
the dlabel file to get the areas by name as separate ROIs (as dscalar
files), and then use -cifti-merge and -cifti-reduce to combine them back
into one larger ROI.

Tim


On Tue, Sep 27, 2016 at 10:36 AM, Michael F.W. Dreyfuss <
mid2...@med.cornell.edu> wrote:

> I can get separate the volumetric part of cifti files with a command like:
>
> wb_command -cifti-separate ToyNogo_fdr_palm/ToyNogo_
> results_merged_tstat.dscalar.nii COLUMN -volume ACCUMBENS_RIGHT
> R_Acc_Beta.nii.gz -roi R_NAcc_ROI.nii.gz
>
> But when I try that on surface data I get the error:
> wb_command -cifti-separate ToyNogo_GlasserParcellation_
> FLOBS1_fdr_palm/ToyNogo_results_dat_tstat.pscalar.nii COLUMN -label
> R_IFSp_ROI ToyNogo_R_IFSp_Beta.pscalar.nii -roi R_IFSp_ROI.plabel.nii
>
> ERROR: unrecognized structure type
>
> The problem seems to be in recognizing the structures name (i.e.
> R_IFSp_ROI). Do you know how I can reference a specific structure or
> structures with a command from a label file, or is there another comparable
> command you would suggest for isolating structures from a label file?
>
> Thank you,
> Michael
>
> On Sep 27, 2016, at 10:41 AM, Burgess, Gregory  wrote:
>
> HCP did not have a task that was geared toward response inhibition.
> Furthermore, although it’s alluring to believe that a single parcel will
> encapsulate all of response inhibition, it’s doubtful.
>
> Why not select a set of parcels in and near IFG, and correct for multiple
> comparisons? Use a meta-analysis as your guide (https://urldefense.
> proofpoint.com/v2/url?u=http-3A__neurosynth.org_analyses_
> terms_response-2520inhibition_=DQIGaQ=lb62iw4YL4RFalcE2hQUQealT9-
> RXrryqt9KZX2qu2s=rPclmYysc_z1plf99IoNsmxWf1JolkKMmL6bXnYF
> Swg=Pb-c1m96L6xixMU-j16J-GkC6tGV5lAdGNKy-Tj1ht8=
> DxHCajoqQSf0nIWSDbUHyB6J7F2eax1NswvaNpprHaU= ). Your statistical power
> should still benefit from the parcellated analysis and reduced number of
> multiple comparisons, relative to a whole-brain analysis.
>
> --Greg
>
> 
> Greg Burgess, Ph.D.
> Staff Scientist, Human Connectome Project
> Washington University School of Medicine
> Department of Psychiatry
> Phone: 314-362-7864
> Email: gburg...@wustl.edu
>
> On Sep 27, 2016, at 9:26 AM, Michael F.W. Dreyfuss <
> mid2...@med.cornell.edu> wrote:
>
> I agree, that’s why I was checking in to see if there was a sub parcel you
> had identified as being involved in response inhibition from your tasks,
> such as flanker. There is a lot of background of IFG being involved in
> response inhibition, particularly on go/nogo tasks,, so I was wondering if
> you had any information on specifically where within your parcellation that
> may be most relevant.
>
> On Sep 27, 2016, at 9:57 AM, Harms, Michael  wrote:
>
> Just a reminder to be careful here to avoid issues of
> circularity/double-dipping.  You indicated that you had a priori hypotheses
> about IFG involvement, but that doesn’t allow you to then select a
> particular IFG parcel based on your task activation map.  Ideally, you
> would have selected your parcel(s) for analysis prior to ever
> computing/viewing your activation map (unless what you showed was a map
> from an independent, unrelated set of subjects).
>
>
> ___
> HCP-Users mailing list
> HCP-Users@humanconnectome.org
> https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.
> humanconnectome.org_mailman_listinfo_hcp-2Dusers=DQIGaQ&
> c=lb62iw4YL4RFalcE2hQUQealT9-RXrryqt9KZX2qu2s=rPclmYysc_
> z1plf99IoNsmxWf1JolkKMmL6bXnYFSwg=Pb-c1m96L6xixMU-j16J-
> GkC6tGV5lAdGNKy-Tj1ht8=Kd8JGIfyDD_MWOgH24WZFrQNLqIT2e2KIaFCaLTw3PM=
>
>
>
>
> --Greg
>
> 
> Greg Burgess, Ph.D.
> Staff Scientist, Human Connectome Project
> Washington University School of Medicine
> Department of Psychiatry
> Phone: 314-362-7864
> Email: gburg...@wustl.edu
>
>
> 
> The materials in this message are private and may contain Protected
> Healthcare Information or other information of a sensitive nature. If you
> are not the intended recipient, be advised that any unauthorized use,
> disclosure, copying or the taking of any action in reliance on the contents
> of this information is strictly prohibited. If you have received this email
> in error, please immediately notify the sender via telephone or return mail.
>
>
> ___
> HCP-Users mailing list
> 

Re: [HCP-Users] Label Import with -discard-others

[Michael F.W. Dreyfuss]

In a related question, if I want to treat hemispheres separately, how can I 
create a simple hemispheric label file, or split a file like 
Human.Brodmann09.32k_fs_LR.dlabel.nii into two hemispheres to treat the 
Brodmann areas of each area separately?


Essentially, all of my questions are directed to find out how I can easily 
separate and combine labels from existing files.


Thank you,

Michael


From: Michael F.W. Dreyfuss
Sent: Tuesday, September 27, 2016 2:19:06 PM
To: hcp-users@humanconnectome.org
Subject: Label Import with -discard-others


Hello, I have been trying to make a file containing the parcellations for BA44 
and BA45 using the Broadmann using the following command:


wb_command -cifti-label-import Human.Brodmann09.32k_fs_LR.dlabel.nii 
BrodmannLabels.txt Brodmann_IFG.dlabel.nii -discard-others


Where BrodmannLabels.txt contains:

BA44

66 100 255 100 255

BA45

65 255 100 100 255


It seems to recognize these labels and keeps their names on the file if I do 
not use -discard-others, but when I use -discard-others the output is totally 
blank, rather than just containing the selected parcels.


Do you have a suggestions for how to do what I am trying to do?


thank you,

Michael

___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

[HCP-Users] Label Import with -discard-others

[Michael F.W. Dreyfuss]

Hello, I have been trying to make a file containing the parcellations for BA44 
and BA45 using the Broadmann using the following command:


wb_command -cifti-label-import Human.Brodmann09.32k_fs_LR.dlabel.nii 
BrodmannLabels.txt Brodmann_IFG.dlabel.nii -discard-others


Where BrodmannLabels.txt contains:

BA44

66 100 255 100 255

BA45

65 255 100 100 255


It seems to recognize these labels and keeps their names on the file if I do 
not use -discard-others, but when I use -discard-others the output is totally 
blank, rather than just containing the selected parcels.


Do you have a suggestions for how to do what I am trying to do?


thank you,

Michael

___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Re: [HCP-Users] ROIs and Betas from Cifti Data

[Michael F.W. Dreyfuss]

I can get separate the volumetric part of cifti files with a command like:

wb_command -cifti-separate 
ToyNogo_fdr_palm/ToyNogo_results_merged_tstat.dscalar.nii COLUMN -volume 
ACCUMBENS_RIGHT R_Acc_Beta.nii.gz -roi R_NAcc_ROI.nii.gz

But when I try that on surface data I get the error:
wb_command -cifti-separate 
ToyNogo_GlasserParcellation_FLOBS1_fdr_palm/ToyNogo_results_dat_tstat.pscalar.nii
 COLUMN -label R_IFSp_ROI ToyNogo_R_IFSp_Beta.pscalar.nii -roi 
R_IFSp_ROI.plabel.nii

ERROR: unrecognized structure type

The problem seems to be in recognizing the structures name (i.e. R_IFSp_ROI). 
Do you know how I can reference a specific structure or structures with a 
command from a label file, or is there another comparable command you would 
suggest for isolating structures from a label file?

Thank you,
Michael

On Sep 27, 2016, at 10:41 AM, Burgess, Gregory 
> wrote:

HCP did not have a task that was geared toward response inhibition. 
Furthermore, although it’s alluring to believe that a single parcel will 
encapsulate all of response inhibition, it’s doubtful.

Why not select a set of parcels in and near IFG, and correct for multiple 
comparisons? Use a meta-analysis as your guide 
(https://urldefense.proofpoint.com/v2/url?u=http-3A__neurosynth.org_analyses_terms_response-2520inhibition_=DQIGaQ=lb62iw4YL4RFalcE2hQUQealT9-RXrryqt9KZX2qu2s=rPclmYysc_z1plf99IoNsmxWf1JolkKMmL6bXnYFSwg=Pb-c1m96L6xixMU-j16J-GkC6tGV5lAdGNKy-Tj1ht8=DxHCajoqQSf0nIWSDbUHyB6J7F2eax1NswvaNpprHaU=
 ). Your statistical power should still benefit from the parcellated analysis 
and reduced number of multiple comparisons, relative to a whole-brain analysis.

--Greg


Greg Burgess, Ph.D.
Staff Scientist, Human Connectome Project
Washington University School of Medicine
Department of Psychiatry
Phone: 314-362-7864
Email: gburg...@wustl.edu

On Sep 27, 2016, at 9:26 AM, Michael F.W. Dreyfuss 
> wrote:

I agree, that’s why I was checking in to see if there was a sub parcel you had 
identified as being involved in response inhibition from your tasks, such as 
flanker. There is a lot of background of IFG being involved in response 
inhibition, particularly on go/nogo tasks,, so I was wondering if you had any 
information on specifically where within your parcellation that may be most 
relevant.

On Sep 27, 2016, at 9:57 AM, Harms, Michael 
> wrote:

Just a reminder to be careful here to avoid issues of 
circularity/double-dipping.  You indicated that you had a priori hypotheses 
about IFG involvement, but that doesn’t allow you to then select a particular 
IFG parcel based on your task activation map.  Ideally, you would have selected 
your parcel(s) for analysis prior to ever computing/viewing your activation map 
(unless what you showed was a map from an independent, unrelated set of 
subjects).


___
HCP-Users mailing list
HCP-Users@humanconnectome.org
https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.humanconnectome.org_mailman_listinfo_hcp-2Dusers=DQIGaQ=lb62iw4YL4RFalcE2hQUQealT9-RXrryqt9KZX2qu2s=rPclmYysc_z1plf99IoNsmxWf1JolkKMmL6bXnYFSwg=Pb-c1m96L6xixMU-j16J-GkC6tGV5lAdGNKy-Tj1ht8=Kd8JGIfyDD_MWOgH24WZFrQNLqIT2e2KIaFCaLTw3PM=




--Greg


Greg Burgess, Ph.D.
Staff Scientist, Human Connectome Project
Washington University School of Medicine
Department of Psychiatry
Phone: 314-362-7864
Email: gburg...@wustl.edu



The materials in this message are private and may contain Protected Healthcare 
Information or other information of a sensitive nature. If you are not the 
intended recipient, be advised that any unauthorized use, disclosure, copying 
or the taking of any action in reliance on the contents of this information is 
strictly prohibited. If you have received this email in error, please 
immediately notify the sender via telephone or return mail.


___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Re: [HCP-Users] ROIs and Betas from Cifti Data

[Burgess, Gregory]

HCP did not have a task that was geared toward response inhibition. 
Furthermore, although it’s alluring to believe that a single parcel will 
encapsulate all of response inhibition, it’s doubtful.

Why not select a set of parcels in and near IFG, and correct for multiple 
comparisons? Use a meta-analysis as your guide 
(http://neurosynth.org/analyses/terms/response%20inhibition/). Your statistical 
power should still benefit from the parcellated analysis and reduced number of 
multiple comparisons, relative to a whole-brain analysis.

--Greg


Greg Burgess, Ph.D.
Staff Scientist, Human Connectome Project
Washington University School of Medicine
Department of Psychiatry
Phone: 314-362-7864
Email: gburg...@wustl.edu

> On Sep 27, 2016, at 9:26 AM, Michael F.W. Dreyfuss  
> wrote:
>
> I agree, that’s why I was checking in to see if there was a sub parcel you 
> had identified as being involved in response inhibition from your tasks, such 
> as flanker. There is a lot of background of IFG being involved in response 
> inhibition, particularly on go/nogo tasks,, so I was wondering if you had any 
> information on specifically where within your parcellation that may be most 
> relevant.
>
>> On Sep 27, 2016, at 9:57 AM, Harms, Michael  wrote:
>>
>> Just a reminder to be careful here to avoid issues of 
>> circularity/double-dipping.  You indicated that you had a priori hypotheses 
>> about IFG involvement, but that doesn’t allow you to then select a 
>> particular IFG parcel based on your task activation map.  Ideally, you would 
>> have selected your parcel(s) for analysis prior to ever computing/viewing 
>> your activation map (unless what you showed was a map from an independent, 
>> unrelated set of subjects).
>>
>
> ___
> HCP-Users mailing list
> HCP-Users@humanconnectome.org
> http://lists.humanconnectome.org/mailman/listinfo/hcp-users
>



--Greg


Greg Burgess, Ph.D.
Staff Scientist, Human Connectome Project
Washington University School of Medicine
Department of Psychiatry
Phone: 314-362-7864
Email: gburg...@wustl.edu



The materials in this message are private and may contain Protected Healthcare 
Information or other information of a sensitive nature. If you are not the 
intended recipient, be advised that any unauthorized use, disclosure, copying 
or the taking of any action in reliance on the contents of this information is 
strictly prohibited. If you have received this email in error, please 
immediately notify the sender via telephone or return mail.

___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Re: [HCP-Users] ROIs and Betas from Cifti Data

[Michael F.W. Dreyfuss]

I agree, that’s why I was checking in to see if there was a sub parcel you had 
identified as being involved in response inhibition from your tasks, such as 
flanker. There is a lot of background of IFG being involved in response 
inhibition, particularly on go/nogo tasks,, so I was wondering if you had any 
information on specifically where within your parcellation that may be most 
relevant.

On Sep 27, 2016, at 9:57 AM, Harms, Michael 
> wrote:

Just a reminder to be careful here to avoid issues of 
circularity/double-dipping.  You indicated that you had a priori hypotheses 
about IFG involvement, but that doesn’t allow you to then select a particular 
IFG parcel based on your task activation map.  Ideally, you would have selected 
your parcel(s) for analysis prior to ever computing/viewing your activation map 
(unless what you showed was a map from an independent, unrelated set of 
subjects).



___
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users

Re: [HCP-Users] ROIs and Betas from Cifti Data

[Harms, Michael]

Hi Michael,
Just a reminder to be careful here to avoid issues of circularity/double-dipping.  You indicated that you had a priori hypotheses about IFG involvement, but that doesn’t allow you to then select a particular IFG parcel based on your task activation map.
  Ideally, you would have selected your parcel(s) for analysis prior to ever computing/viewing your activation map (unless what you showed was a map from an independent, unrelated set of subjects).


cheers,
-MH




-- 
Michael Harms, Ph.D.

---
Conte Center for the Neuroscience of Mental Disorders
Washington University School of Medicine
Department of Psychiatry, Box 8134
660 South Euclid Ave. 
Tel: 314-747-6173
St. Louis, MO  63110 
Email: mha...@wustl.edu







From:  on behalf of "Glasser, Matthew" 
Date: Tuesday, September 27, 2016 at 4:52 AM
To: "Michael F.W. Dreyfuss" , "hcp-users@humanconnectome.org" ,
 NEUROSCIENCE tim 
Subject: Re: [HCP-Users] ROIs and Betas from Cifti Data





I clicked in a corresponding location and your activation appears to fall within area IFSp.  The inferior frontal gyrus contains multiple areas, as does the inferior frontal sulcus.  We have only directly commented on the function of the areas in relation
 to the 7 HCP acquired tasks (See the supplementary neuroanatomical results).


Peace,


Matt.




From: "Michael F.W. Dreyfuss" 
Date: Monday, September 26, 2016 at 5:59 PM
To: Matt Glasser , "hcp-users@humanconnectome.org" ,
 Timothy Coalson 
Subject: Re: [HCP-Users] ROIs and Betas from Cifti Data







Yes, but none of these regions is exactly what's typically called IFG, so to limit comparisons I was wondering if you knew of a reason that one of them should be considered a locus of response inhibition based on your data that would warrant using that region
 as an a priori ROI to look for neural correlates of differences in response inhibition.


From: Glasser, Matthew 
Sent: Monday, September 26, 2016 6:22:50 PM
To: Michael F.W. Dreyfuss; 
hcp-users@humanconnectome.org; NEUROSCIENCE tim
Subject: Re: [HCP-Users] ROIs and Betas from Cifti Data
 


Can’t you just click which area is under that cluster?  I didn’t realize you already had the data registered.


Peace,


Matt.




From: "Michael F.W. Dreyfuss" 
Date: Monday, September 26, 2016 at 4:09 PM
To: Matt Glasser , "hcp-users@humanconnectome.org" ,
 Timothy Coalson 
Subject: Re: [HCP-Users] ROIs and Betas from Cifti Data







Thank you,


I've attached an image with a dot on the cluster I'm interested in. I could send you the dscalar file too, but prefer not to distribute to the whole listserv. 


Since we have a priori hypotheses about the IFG being involved in this task (it is a go/nogo task), I want test for an effect specifically in that region similar to the one shown that I am seeing. In volumetric analysis I would use an anatomically or functionally
 defined mask and either A) look for a region within that mask that survives cluster thresholding to localize where in that region is involved or B) test if average betas are different within that region between conditions. I was wondering if one of the parcels
 from your parcellation was something you know to be where response inhibition on tasks like go/nogo may be localized as a reason for using that parcellation or set of parcellations. Otherwise I would be interested in something like a combined mask of BA44
 and BA45 as representing IFG.


Also, is it possible to do something like TFCE or fdr correction within a region like this within palm as that would reduce the area I am looking over for an effect and reduce the cluster size needed compared to a whole brain analysis?


Thank you,
Michael


From: Glasser, Matthew 
Sent: Monday, September 26, 2016 2:18:26 PM
To: Michael F.W. Dreyfuss; NEUROSCIENCE tim
Cc: hcp-users@humanconnectome.org
Subject: Re: [HCP-Users] ROIs and Betas from Cifti Data
 


Is there a surface-based map of this task contrast somewhere that I could look at?  Without that I would just be guessing…


Peace,


Matt.




From: "Michael F.W. Dreyfuss" 
Date: Monday, September 26, 2016 at 1:13 PM
To: Matt Glasser , Timothy Coalson 
Cc: "hcp-users@humanconnectome.org" 
Subject: Re: [HCP-Users] ROIs and Betas from Cifti Data







Hi Matt, I'm interested in using part of your parcellation to look at response inhibition on a go/nogo