help > Edge files of BrainNet Viewer (Example Files)
Showing 1-6 of 6 posts
Sep 6, 2018 06:09 AM | Reddy Rani - International Institute of Information Technology
Edge files of BrainNet Viewer (Example Files)
- I have started exploring with Example files provided with BrainNet Viewer (BNV) tool. I can understand that edges were extracting either by a) region based, b) voxel-based. AAL edges are region based and the Pearson correlation coefficient (PCC) is used for edge extraction. The diagonal of the edge list (Matrix) is expected to be 1 in PCC, but the files provided with BNV are 0 (zero) diagonal values. Can I get help in understanding the same?
- Is DOS edge list is from a voxel-based network edges extraction technique? can I get a bit of elaborate workflow of edge extraction in DOS? With this edge list also I observed 0 (zero) values in diagonal. Is it a case, for the sake of measuring topological parameters of the network, the diagonal of the edge matrix is replaced with zero.
- I noticed the range of edge values differing, AAL and DOS are fro -1 to 1, but for HOA weighted matrix, the edges are ranging from 0 to 1. Why is this?
I am relatively new to the domain, please help me to understand. Any help and suggestions are highly appreciated.
Thanks In advance.
Sep 10, 2018 02:09 AM | Mingrui Xia
RE: Edge files of BrainNet Viewer (Example Files)
Hi,
1. Sure, the diagonal of the edge list should be 1 when using PCC to construct correlation matrix; however, the diagonal is not used in BNV. Thus, the diagonal was set to 0 manually.
2. The DOS edges were not extracted from voxel-based network, but similar to the AAL template. Please see Dosenbach et al., Science 2010 for more detail for this parcellation.
3. I couldn't remember if the HOA example is derived from fMRI data or DTI data, the values in which could be probabilities across a group of people. These example files are mainly provide an illustrating for how to use BNV with determined file format and data arrangement.
Best
Mingrui
Originally posted by Reddy Rani:
I am relatively new to the domain, please help me to understand. Any help and suggestions are highly appreciated.
Thanks In advance.
1. Sure, the diagonal of the edge list should be 1 when using PCC to construct correlation matrix; however, the diagonal is not used in BNV. Thus, the diagonal was set to 0 manually.
2. The DOS edges were not extracted from voxel-based network, but similar to the AAL template. Please see Dosenbach et al., Science 2010 for more detail for this parcellation.
3. I couldn't remember if the HOA example is derived from fMRI data or DTI data, the values in which could be probabilities across a group of people. These example files are mainly provide an illustrating for how to use BNV with determined file format and data arrangement.
Best
Mingrui
Originally posted by Reddy Rani:
- I have started exploring with Example files provided with BrainNet Viewer (BNV) tool. I can understand that edges were extracting either by a) region based, b) voxel-based. AAL edges are region based and the Pearson correlation coefficient (PCC) is used for edge extraction. The diagonal of the edge list (Matrix) is expected to be 1 in PCC, but the files provided with BNV are 0 (zero) diagonal values. Can I get help in understanding the same?
- Is DOS edge list is from a voxel-based network edges extraction technique? can I get a bit of elaborate workflow of edge extraction in DOS? With this edge list also I observed 0 (zero) values in diagonal. Is it a case, for the sake of measuring topological parameters of the network, the diagonal of the edge matrix is replaced with zero.
- I noticed the range of edge values differing, AAL and DOS are fro -1 to 1, but for HOA weighted matrix, the edges are ranging from 0 to 1. Why is this?
I am relatively new to the domain, please help me to understand. Any help and suggestions are highly appreciated.
Thanks In advance.
Sep 10, 2018 07:09 AM | Reddy Rani - International Institute of Information Technology
RE: Edge files of BrainNet Viewer (Example Files)
Originally posted by Reddy Rani:
I am relatively new to the domain, please help me to understand. Any help and suggestions are highly appreciated.
Thanks In advance.
- I have started exploring with Example files provided with BrainNet Viewer (BNV) tool. I can understand that edges were extracting either by a) region based, b) voxel-based. AAL edges are region based and the Pearson correlation coefficient (PCC) is used for edge extraction. The diagonal of the edge list (Matrix) is expected to be 1 in PCC, but the files provided with BNV are 0 (zero) diagonal values. Can I get help in understanding the same?
- Is DOS edge list is from a voxel-based network edges extraction technique? can I get a bit of elaborate workflow of edge extraction in DOS? With this edge list also I observed 0 (zero) values in diagonal. Is it a case, for the sake of measuring topological parameters of the network, the diagonal of the edge matrix is replaced with zero.
- I noticed the range of edge values differing, AAL and DOS are fro -1 to 1, but for HOA weighted matrix, the edges are ranging from 0 to 1. Why is this?
I am relatively new to the domain, please help me to understand. Any help and suggestions are highly appreciated.
Thanks In advance.
Sep 10, 2018 07:09 AM | Reddy Rani - International Institute of Information Technology
RE: Edge files of BrainNet Viewer (Example Files)
Thanks a lot for the reply. It is very helpful.
Aug 3, 2020 02:08 PM | Reddy Rani - International Institute of Information Technology
RE: Edge files of BrainNet Viewer (Example Files)
Hi,
As mentioned earlier, I noticed the range of edge values differing, AAL are between -1 to 1. I also noted that the correlation matrices were transferred into z-score matrices using Fisher's r-to-z transformation. The z transformed data range is -2 to 2. I need your help regarding the same.
Thanks in advance.
Regards,
Rani.
As mentioned earlier, I noticed the range of edge values differing, AAL are between -1 to 1. I also noted that the correlation matrices were transferred into z-score matrices using Fisher's r-to-z transformation. The z transformed data range is -2 to 2. I need your help regarding the same.
- When will/ how will we know, the data need to be z transformed.
- Can you help me to know, how to scale down the z transformed data to -1 to 1 range.
Thanks in advance.
Regards,
Rani.
Aug 5, 2020 04:08 AM | Mingrui Xia
RE: Edge files of BrainNet Viewer (Example Files)
Hi,
The Fisher's z transformation is usually performed when we need to apply statistical analysis on the connections.
The formula of Fisher's z transformation is Z = 0.5ln((1+r)/(1-r)). An inverse transform can turn the z values back to r.
Best,
Mingrui
Originally posted by Reddy Rani:
The Fisher's z transformation is usually performed when we need to apply statistical analysis on the connections.
The formula of Fisher's z transformation is Z = 0.5ln((1+r)/(1-r)). An inverse transform can turn the z values back to r.
Best,
Mingrui
Originally posted by Reddy Rani:
Hi,
As mentioned earlier, I noticed the range of edge values differing, AAL are between -1 to 1. I also noted that the correlation matrices were transferred into z-score matrices using Fisher's r-to-z transformation. The z transformed data range is -2 to 2. I need your help regarding the same.
Thanks in advance.
Regards,
Rani.
As mentioned earlier, I noticed the range of edge values differing, AAL are between -1 to 1. I also noted that the correlation matrices were transferred into z-score matrices using Fisher's r-to-z transformation. The z transformed data range is -2 to 2. I need your help regarding the same.
- When will/ how will we know, the data need to be z transformed.
- Can you help me to know, how to scale down the z transformed data to -1 to 1 range.
Thanks in advance.
Regards,
Rani.