help > Task-based functional connectivity: fALFF and seed-to-voxel analyses
Showing 1-2 of 2 posts
Sep 14, 2022 03:09 PM | Benxamin varela - Universidad de Santiago de Compostela
Task-based functional connectivity: fALFF and seed-to-voxel analyses
Dear Alfonso
I am conducting a task-based functional connectivity study with two groups (group A vs. group B). I performed a fALFF analysis and I have the following questions.
1º fALFF analyses are widely used in resting-state fMRI studies. Does It make sense to perform this type of analysis in a task-based functional connectivity study?
2º Considering that this is a task study we specify a high pass filter in the denoising (0.008-inf) (https://www.nitrc.org/forum/message.php?msg_id=11196). Is this filter appropriate for fALFF analysis in task-based functional connectivity studies?
3º We conducted these analyses as a pilot test and found that the group A shows an increase in fALFF values compared to the group B in an occipital cluster. As a post-hoc analysis, we performed a seed-to-voxel analysis using the significant cluster derived from the fALFF analysis. This post-hoc analysis showed an increase in the connectivity of this cluster with frontal regions for the group B compared to the group A. I don't understand how the group B may display, at the same time, a decreased fALFF values in the occipital cluster and a hiperconnectivity between that occipital cluster and frontal regions in the seed-to-voxel post-hoc analysis.
Any suggestions would be of great help.
Thank you very much in advance.
Best regards.
Benxa.
I am conducting a task-based functional connectivity study with two groups (group A vs. group B). I performed a fALFF analysis and I have the following questions.
1º fALFF analyses are widely used in resting-state fMRI studies. Does It make sense to perform this type of analysis in a task-based functional connectivity study?
2º Considering that this is a task study we specify a high pass filter in the denoising (0.008-inf) (https://www.nitrc.org/forum/message.php?msg_id=11196). Is this filter appropriate for fALFF analysis in task-based functional connectivity studies?
3º We conducted these analyses as a pilot test and found that the group A shows an increase in fALFF values compared to the group B in an occipital cluster. As a post-hoc analysis, we performed a seed-to-voxel analysis using the significant cluster derived from the fALFF analysis. This post-hoc analysis showed an increase in the connectivity of this cluster with frontal regions for the group B compared to the group A. I don't understand how the group B may display, at the same time, a decreased fALFF values in the occipital cluster and a hiperconnectivity between that occipital cluster and frontal regions in the seed-to-voxel post-hoc analysis.
Any suggestions would be of great help.
Thank you very much in advance.
Best regards.
Benxa.
Sep 26, 2022 11:09 PM | Alfonso Nieto-Castanon - Boston University
RE: Task-based functional connectivity: fALFF and seed-to-voxel analyses
Dear Benxa
That is a very good question, I do not know if there are other studies that may have looked at this but my impression is that it may make more sense to use ALFF rather than fALFF in the context of task designs. The reason (but I may be wrong here) is that unless your task design consist of very long blocks, and as you mention in (2), you would not want to low-pass filter the data in order to avoid mixing-up responses across blocks, but then I would expect that, when only using a high-pass filter as opposed to the standard band-pass filter, the resulting fALFF measure may perhaps overly depend on the amount of very-low-frequency noise in the data (e.g. for a TR of 2s, a typical fALFF measure would mostly depend on relative amount of variance within the [.01 .10] Hz band and the [.10 .25] Hz band, but when using only a high-pass filter instead fALFF will moslty depend on the relative amount of variance within the [.01 .25] Hz band and the [0 .01] Hz band). Because of this, I would probably suggest using ALFF instead, as this is simply going to measure the overall variance in the BOLD signal within each task/condition block, and may be more easily interpretable (?).
Hope this helps
Alfonso
Originally posted by Benxamin varela:
That is a very good question, I do not know if there are other studies that may have looked at this but my impression is that it may make more sense to use ALFF rather than fALFF in the context of task designs. The reason (but I may be wrong here) is that unless your task design consist of very long blocks, and as you mention in (2), you would not want to low-pass filter the data in order to avoid mixing-up responses across blocks, but then I would expect that, when only using a high-pass filter as opposed to the standard band-pass filter, the resulting fALFF measure may perhaps overly depend on the amount of very-low-frequency noise in the data (e.g. for a TR of 2s, a typical fALFF measure would mostly depend on relative amount of variance within the [.01 .10] Hz band and the [.10 .25] Hz band, but when using only a high-pass filter instead fALFF will moslty depend on the relative amount of variance within the [.01 .25] Hz band and the [0 .01] Hz band). Because of this, I would probably suggest using ALFF instead, as this is simply going to measure the overall variance in the BOLD signal within each task/condition block, and may be more easily interpretable (?).
Hope this helps
Alfonso
Originally posted by Benxamin varela:
Dear Alfonso
I am conducting a task-based functional connectivity study with two groups (group A vs. group B). I performed a fALFF analysis and I have the following questions.
1º fALFF analyses are widely used in resting-state fMRI studies. Does It make sense to perform this type of analysis in a task-based functional connectivity study?
2º Considering that this is a task study we specify a high pass filter in the denoising (0.008-inf) (https://www.nitrc.org/forum/message.php?msg_id=11196). Is this filter appropriate for fALFF analysis in task-based functional connectivity studies?
3º We conducted these analyses as a pilot test and found that the group A shows an increase in fALFF values compared to the group B in an occipital cluster. As a post-hoc analysis, we performed a seed-to-voxel analysis using the significant cluster derived from the fALFF analysis. This post-hoc analysis showed an increase in the connectivity of this cluster with frontal regions for the group B compared to the group A. I don't understand how the group B may display, at the same time, a decreased fALFF values in the occipital cluster and a hiperconnectivity between that occipital cluster and frontal regions in the seed-to-voxel post-hoc analysis.
Any suggestions would be of great help.
Thank you very much in advance.
Best regards.
Benxa.
I am conducting a task-based functional connectivity study with two groups (group A vs. group B). I performed a fALFF analysis and I have the following questions.
1º fALFF analyses are widely used in resting-state fMRI studies. Does It make sense to perform this type of analysis in a task-based functional connectivity study?
2º Considering that this is a task study we specify a high pass filter in the denoising (0.008-inf) (https://www.nitrc.org/forum/message.php?msg_id=11196). Is this filter appropriate for fALFF analysis in task-based functional connectivity studies?
3º We conducted these analyses as a pilot test and found that the group A shows an increase in fALFF values compared to the group B in an occipital cluster. As a post-hoc analysis, we performed a seed-to-voxel analysis using the significant cluster derived from the fALFF analysis. This post-hoc analysis showed an increase in the connectivity of this cluster with frontal regions for the group B compared to the group A. I don't understand how the group B may display, at the same time, a decreased fALFF values in the occipital cluster and a hiperconnectivity between that occipital cluster and frontal regions in the seed-to-voxel post-hoc analysis.
Any suggestions would be of great help.
Thank you very much in advance.
Best regards.
Benxa.