help > How is Spatial match to template correlation value calculated?
Showing 1-6 of 6 posts
Jun 25, 2019 01:06 PM | doora - ELTE
How is Spatial match to template correlation value calculated?
Hi!
To evaluate reproducibility, I compared two set of ICAs. I found information about how the Dice similarity coefficient (DSC) is calculated online, so I have an idea how that works, but there is another option in conn, calculation of the 'correlation' which yields an r value. But I found no information about how this is calculated in the manuals.
Is it e.g. calculated form thresholded whole-brain voxel z-scores of the functional maps? Can I read about this in some kind of documentation?
Thanks in advance
To evaluate reproducibility, I compared two set of ICAs. I found information about how the Dice similarity coefficient (DSC) is calculated online, so I have an idea how that works, but there is another option in conn, calculation of the 'correlation' which yields an r value. But I found no information about how this is calculated in the manuals.
Is it e.g. calculated form thresholded whole-brain voxel z-scores of the functional maps? Can I read about this in some kind of documentation?
Thanks in advance
Aug 4, 2020 01:08 PM | Tsvetoslav Ivanov
RE: How is Spatial match to template correlation value calculated?
Originally posted by doora:
Hi all,
I am also interested in finding out how the spatial correlation coefficients are calculated and what do they mean. Could somebody provide guidance?
Thanks,
Tsvet
Hi!
To evaluate reproducibility, I compared two set of ICAs. I found information about how the Dice similarity coefficient (DSC) is calculated online, so I have an idea how that works, but there is another option in conn, calculation of the 'correlation' which yields an r value. But I found no information about how this is calculated in the manuals.
Is it e.g. calculated form thresholded whole-brain voxel z-scores of the functional maps? Can I read about this in some kind of documentation?
Thanks in advance
To evaluate reproducibility, I compared two set of ICAs. I found information about how the Dice similarity coefficient (DSC) is calculated online, so I have an idea how that works, but there is another option in conn, calculation of the 'correlation' which yields an r value. But I found no information about how this is calculated in the manuals.
Is it e.g. calculated form thresholded whole-brain voxel z-scores of the functional maps? Can I read about this in some kind of documentation?
Thanks in advance
Hi all,
I am also interested in finding out how the spatial correlation coefficients are calculated and what do they mean. Could somebody provide guidance?
Thanks,
Tsvet
Aug 12, 2020 03:08 AM | Tsvetoslav Ivanov
RE: How is Spatial match to template correlation value calculated?
I found a paper by Grecius et. al. (2004) "Default-mode network
activity distinguishes Alzheimer's disease from healthy aging:
Evidence from functional MRI", which seems to provide an
explanation of spatial correlation:
"Next, a template of the default-mode network was used to select the "best-fit" of the remaining low-frequency components in each subject. To do this, we developed a nonlinear template-matching procedure that involved taking the average z score of voxels falling within the template minus the average z score of voxels outside the template and selecting the component in which this difference (the goodness of fit) was the greatest. z scores here reflect the degree to which a given voxel's time series correlates with the time series corresponding to the specific ICA component."
The paper has been cited close to 3500 times so I suppose it serves as a guide in the lierature. Is this the approach adopted by CONN? I have searched in the forum and could only find this question from last year, which was unaswered.
"Next, a template of the default-mode network was used to select the "best-fit" of the remaining low-frequency components in each subject. To do this, we developed a nonlinear template-matching procedure that involved taking the average z score of voxels falling within the template minus the average z score of voxels outside the template and selecting the component in which this difference (the goodness of fit) was the greatest. z scores here reflect the degree to which a given voxel's time series correlates with the time series corresponding to the specific ICA component."
The paper has been cited close to 3500 times so I suppose it serves as a guide in the lierature. Is this the approach adopted by CONN? I have searched in the forum and could only find this question from last year, which was unaswered.
Aug 19, 2021 03:08 PM | Yunmeng Zhu
RE: How is Spatial match to template correlation value calculated?
Originally posted by Tsvetoslav Ivanov:
Do you have any idea what is the exact criteria (e.g. a value of correlation coefficient) that CONN use to draw the red squares (the size of the red square)?
Thank you!
I found a paper by Grecius et. al. (2004)
"Default-mode network activity distinguishes Alzheimer's disease
from healthy aging: Evidence from functional MRI", which seems to
provide an explanation of spatial correlation:
"Next, a template of the default-mode network was used to select the "best-fit" of the remaining low-frequency components in each subject. To do this, we developed a nonlinear template-matching procedure that involved taking the average z score of voxels falling within the template minus the average z score of voxels outside the template and selecting the component in which this difference (the goodness of fit) was the greatest. z scores here reflect the degree to which a given voxel's time series correlates with the time series corresponding to the specific ICA component."
The paper has been cited close to 3500 times so I suppose it serves as a guide in the lierature. Is this the approach adopted by CONN? I have searched in the forum and could only find this question from last year, which was unaswered.
Hello,"Next, a template of the default-mode network was used to select the "best-fit" of the remaining low-frequency components in each subject. To do this, we developed a nonlinear template-matching procedure that involved taking the average z score of voxels falling within the template minus the average z score of voxels outside the template and selecting the component in which this difference (the goodness of fit) was the greatest. z scores here reflect the degree to which a given voxel's time series correlates with the time series corresponding to the specific ICA component."
The paper has been cited close to 3500 times so I suppose it serves as a guide in the lierature. Is this the approach adopted by CONN? I have searched in the forum and could only find this question from last year, which was unaswered.
Do you have any idea what is the exact criteria (e.g. a value of correlation coefficient) that CONN use to draw the red squares (the size of the red square)?
Thank you!
Aug 19, 2021 03:08 PM | Yunmeng Zhu
RE: How is Spatial match to template correlation value calculated?
hi,
I wonder if you got the answer?
Thank you!
I wonder if you got the answer?
Thank you!
Jun 30, 2022 05:06 PM | Federico Zanella
RE: How is Spatial match to template correlation value calculated?
Hi everyone,
has anyone received this information?
Thanks
has anyone received this information?
Thanks