Print

---

Hi Hong,

 

1) After the Randomise process and when viewing the significant results with Fslview, the TBSS homepage only describes how we can color our significant coordinates.
But is there any way I can get the list of these significant coordinates at once, for example with one click of button, like SPM?
If I could see all the peak significant coordinates in a Table, I guess I won't have to search every Red-Yellow and Blue-Lightblue dots and worry about missing them.

----------------------

 If you find out how to do this step, please share with me.  I am new to this analysis too.

 

2) And is there any way I can find where these significant coordinates indicate as anatomical structures?
Currently I guess the only way I can find the name of those white matter structures is to buy and look into the Atlas of Human White Matter (Mori et al., 2005).

3) Some published articles describe that the authors have conducted sufficient number of permutations but not any further correction for multiple comparisions (e.g. TFCE).
Is it an appropriate way? Or, are the permutations and the TFCE correction (or FDR correction, or voxel-wise FWE correction) both needed?
------------------------

Permutations method is one of the non-parametric analysis methods. No correction is inappropriate.  Because you'll have high false positive rate. Look FSL online course about the difference between FDR, FWE.  I think if you have time, you can use both.  FWE v.s. FDR.

4) Should TFCE-uncorrected image and voxel-wise uncorrected image display the same results, since they are both uncorrected?
--------------------------

I am not sure if I understand this question. what's the exact command you use for those two image types?

 

Yingying

 

==============================================
Yingying Wang
Graduate Student
Biomedical Engineering
University of Cincinnati
Pediatric Neuroimaging Consortium Research Center
Cincinnati Children's Hospital Medical Center
MLC 5033, 3333 Burnet Avenue
Cincinnati, OH 45229-3039
O 513-636-3495
C 513-833-7448

>>> "Garin, Alba" <[log in to unmask]> 10/10/2011 3:06 AM >>>
Hello Hong,

fsl4.1-cluster is the command name, in some cases should work only with the command "cluster". It takes litlle time and shows you the group of voxels that are statistically significant (if you use "-t 0.95"). when i run the script did not open any FSL window.

Hope it helps.

Alba
________________________________________
De: FSL - FMRIB's Software Library [[log in to unmask]] En nombre de 홍순범 [[log in to unmask]]
Enviado el: domingo, 09 de octubre de 2011 16:15
Para: [log in to unmask]
Asunto: Re: [FSL] About TBSS results

Thank you for the reply, Alba.

Now I have tried "fsl4.1-cluster -i inputimage -t 0.95 --mm" but it's not working yet.
Is "fsl4.1-cluster" a command name?
In my case, when I try the script "fsl4.1-cluster -i inputimage -t 0.95 --mm", the FSL window opens, and nothing else happens.
Could you explain your answer in more detail?
Thank you very much.

Hong


2011/10/5 Garin, Alba <[log in to unmask]<mailto:[log in to unmask]>>
Hi.
I am newby in FSL but i can answer you to few questions because i used fsl lately.

1.- you can use
fsl4.1-cluster -i inputimage -t 0.95 --mm
it shows if you have any voxel (or cluster of voxels) below the threshold of significance (0.05) for you statistical analysis.
It gives you the results in table mode, displaying the x,y,z coordinates of the max value of each cluster, number of voxels per cluster. (coordinates are in mm ).

2.- you can use the fsl atlas, displaying them in Tools/Toolbars/atlas tools. you can click in atlases and check each atlas, or see structures and check if any of the structures you are insterested in is coloured.

Hope it helps.

Alba Garin
PhD Student
Applied Mechanics
Ceit - San Sebastian
Spain


________________________________________
De: FSL - FMRIB's Software Library [[log in to unmask]<mailto:[log in to unmask]>] En nombre de SUBSCRIBE FSL Anonymous [[log in to unmask]<mailto:[log in to unmask]>]
Enviado el: martes, 04 de octubre de 2011 14:01
Para: [log in to unmask]<mailto:[log in to unmask]>
Asunto: [FSL] About TBSS results

Hi, I'm now trying my first DTI analysis with FSL and TBSS. May I ask a few questions?

1) After the Randomise process and when viewing the significant results with Fslview, the TBSS homepage only describes how we can color our significant coordinates.
But is there any way I can get the list of these significant coordinates at once, for example with one click of button, like SPM?
If I could see all the peak significant coordinates in a Table, I guess I won't have to search every Red-Yellow and Blue-Lightblue dots and worry about missing them.

2) And is there any way I can find where these significant coordinates indicate as anatomical structures?
Currently I guess the only way I can find the name of those white matter structures is to buy and look into the Atlas of Human White Matter (Mori et al., 2005).

3) Some published articles describe that the authors have conducted sufficient number of permutations but not any further correction for multiple comparisions (e.g. TFCE).
Is it an appropriate way? Or, are the permutations and the TFCE correction (or FDR correction, or voxel-wise FWE correction) both needed?

4) Should TFCE-uncorrected image and voxel-wise uncorrected image display the same results, since they are both uncorrected?

Thank you very much.

Sincerely,

Hong