help > non-square adj. matrices
Showing 1-2 of 2 posts
Mar 27, 2017 02:03 PM | Toni Abe
non-square adj. matrices
Hello,
First of all, that you very much Andrew for this contribution and for making this public.
What I am trying to do is test the edges between one (or a few ROIs) and all other ROIs in the brain/my atlas (e.g. test edges between 1 ROI and all others: 90-1=89 in the case of AAL). To do that, I figured my adjacency matrix will have to include only entries that are relevant to that (in the case of my example, a 1x89 matrix), and so I was wondering if NBS is able to handle non-square matrices?
Thank you.
Toni
First of all, that you very much Andrew for this contribution and for making this public.
What I am trying to do is test the edges between one (or a few ROIs) and all other ROIs in the brain/my atlas (e.g. test edges between 1 ROI and all others: 90-1=89 in the case of AAL). To do that, I figured my adjacency matrix will have to include only entries that are relevant to that (in the case of my example, a 1x89 matrix), and so I was wondering if NBS is able to handle non-square matrices?
Thank you.
Toni
Mar 27, 2017 11:03 PM | Andrew Zalesky
RE: non-square adj. matrices
Hi Toni,
If you are only interested in testing the elements of a 1 x 89 matrix, you simply fill all the other entries of an 89 x 89 matrix with zeros, except for the first column and first row. Note that matrix elements that are zero for every subject are completely ignored by the software.
For example, suppose you are interested in an 1 x 3 matrix of with the values [ 1 2 3 ], then you would embed this as follows:
[0 1 2 3
1 0 0 0
2 0 0 0
3 0 0 0]
In other words, just set to zero all the connections that you want to exclude from statistical testing a priori.
Andrew
Originally posted by Toni Abe:
If you are only interested in testing the elements of a 1 x 89 matrix, you simply fill all the other entries of an 89 x 89 matrix with zeros, except for the first column and first row. Note that matrix elements that are zero for every subject are completely ignored by the software.
For example, suppose you are interested in an 1 x 3 matrix of with the values [ 1 2 3 ], then you would embed this as follows:
[0 1 2 3
1 0 0 0
2 0 0 0
3 0 0 0]
In other words, just set to zero all the connections that you want to exclude from statistical testing a priori.
Andrew
Originally posted by Toni Abe:
Hello,
First of all, that you very much Andrew for this contribution and for making this public.
What I am trying to do is test the edges between one (or a few ROIs) and all other ROIs in the brain/my atlas (e.g. test edges between 1 ROI and all others: 90-1=89 in the case of AAL). To do that, I figured my adjacency matrix will have to include only entries that are relevant to that (in the case of my example, a 1x89 matrix), and so I was wondering if NBS is able to handle non-square matrices?
Thank you.
Toni
First of all, that you very much Andrew for this contribution and for making this public.
What I am trying to do is test the edges between one (or a few ROIs) and all other ROIs in the brain/my atlas (e.g. test edges between 1 ROI and all others: 90-1=89 in the case of AAL). To do that, I figured my adjacency matrix will have to include only entries that are relevant to that (in the case of my example, a 1x89 matrix), and so I was wondering if NBS is able to handle non-square matrices?
Thank you.
Toni