Hi Michael and Chris,
Thanks a lot for the help, this works well. However, I am not absolutely sure I need to use a binary matrix for both NBS and the calculation of graph metrics. I still want to have a symmetric, weighted, thresholded matrix. The threshold should only help to remove spurious connections.
Furthermore, what do I need to do with the diagonal - are 0s fine or make 1s more sense.
Many thanks again.
André
________________________________
Von: FSL - FMRIB's Software Library [[log in to unmask]]" im Auftrag von "Watson, Christopher [[log in to unmask]]
Gesendet: Dienstag, 25. August 2015 22:41
An: [log in to unmask]
Betreff: Re: [FSL] fdt_network_matrix - no symmetry?
For the computer and user alike! I forgot it was that simple.
________________________________
From: FSL - FMRIB's Software Library <[log in to unmask]> on behalf of Harms, Michael <[log in to unmask]>
Sent: Tuesday, August 25, 2015 4:21 PM
To: [log in to unmask]
Subject: Re: [FSL] fdt_network_matrix - no symmetry?
Matlab is most efficient when you can avoid 'for' loops.
Athresh = (A > thresh);
cheers,
--
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: [log in to unmask]
From: <Watson>, Christopher <[log in to unmask]<mailto:[log in to unmask]>>
Reply-To: FSL - FMRIB's Software Library <[log in to unmask]<mailto:[log in to unmask]>>
Date: Tuesday, August 25, 2015 12:18 PM
To: FSL - FMRIB's Software Library <[log in to unmask]<mailto:[log in to unmask]>>
Subject: Re: [FSL] fdt_network_matrix - no symmetry?
My Matlab is a bit rusty so this is probably not the best way, but it works (and is certainly fast enough):
N = size(A, 1);
Athresh = zeros(N, N);
for i=1:N,
for j=1:N,
if A(i, j) > thresh, Athresh(i, j) = 1; else Athresh(i, j) = 0; end
end
end
Where A is the original matrix, and thresh is whatever threshold you choose.
I wouldn't publish results with no thresholding; in my opinion, you have to choose *some* threshold above which you are willing to believe the interregional connection is real.
________________________________
From: FSL - FMRIB's Software Library <[log in to unmask]<mailto:[log in to unmask]>> on behalf of André Schmidt <[log in to unmask]<mailto:[log in to unmask]>>
Sent: Tuesday, August 25, 2015 11:35 AM
To: [log in to unmask]<mailto:[log in to unmask]>
Subject: [FSL] AW: [FSL] fdt_network_matrix - no symmetry?
Thanks Chris. This is what I have done.
I guess I have to threshold the matrix somehow, although there is also much evidence published without thresholding. Do you have any suggestion for the threshold? According to a nice paper (Cao et al 2013, JNS), a threshold range between 0.01 and 0.1 at intervals of 0.0025 is may appropriate? How can I do this in matlab?
Many thanks again.
André
________________________________
Von: FSL - FMRIB's Software Library [[log in to unmask]<mailto:[log in to unmask]>]" im Auftrag von "Watson, Christopher [[log in to unmask]<mailto:[log in to unmask]>]
Gesendet: Dienstag, 25. August 2015 17:20
An: [log in to unmask]<mailto:[log in to unmask]>
Betreff: Re: [FSL] fdt_network_matrix - no symmetry?
I don' think the order matters much, but dividing by waytotal first will save you a step or 2.
________________________________
From: FSL - FMRIB's Software Library <[log in to unmask]<mailto:[log in to unmask]>> on behalf of André Schmidt <[log in to unmask]<mailto:[log in to unmask]>>
Sent: Tuesday, August 25, 2015 8:24 AM
To: [log in to unmask]<mailto:[log in to unmask]>
Subject: [FSL] AW: [FSL] fdt_network_matrix - no symmetry?
Hi Moises,
I guess I have figured out how to calculate the transpose matrix and the average of the two matrices.
Just a quick follow up question: I also like to normalize the matrix with the waytotal values. Should I do this before these calculations?
Also thanks for checking the recently sent seed mask, your help is highly appreciated.
Best wishes
André
________________________________
Von: FSL - FMRIB's Software Library [[log in to unmask]<mailto:[log in to unmask]>]" im Auftrag von "Moises Hernandez [[log in to unmask]<mailto:[log in to unmask]>]
Gesendet: Dienstag, 25. August 2015 13:10
An: [log in to unmask]<mailto:[log in to unmask]>
Betreff: Re: [FSL] fdt_network_matrix - no symmetry?
Hi André,
They are not symmetric because you are running different experiments from each seeds-ROI.
First from left_thalamus to the rest.
Then from left_caudate to the rest.
etc...
So, you will get different results for left_thalamus->left_caudate (experiment 1)
and left_caudate->left_thalamus (experiment 2).
What you can do is to calculate the transpose of the Matrix and then average both Matrices.
Moises.
On 24 August 2015 at 16:57, André Schmidt <[log in to unmask]<mailto:[log in to unmask]>> wrote:
Dear fsl experts,
I have performed probtrackx2 on my DTI data and now like to compare the obtained fdt_network_matrix across two groups using network-based statistics. To do this, the connectivity matrices must be symmetric about the diagonal. However, I realized that this is not the case for the fdt network matrices (see attached an example for the matrix file and seed file). Why is the number of streamlines from region i to j not the same as from j to i? Can I symmetrize the matrix somehow?
Thank you very much for your explanations and help
Best wishes
André
________________________________
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.
