help
help > RE: scrubbing denoising step
Jan 29, 2021 05:01 PM | Alfonso Nieto-Castanon - Boston University
RE: scrubbing denoising step
Hi Ilka,
In addition to what Andrew suggested, another thing that I would check is whether the 'realignment' parameters have been entered correctly into CONN (see this post https://www.nitrc.org/forum/message.php?msg_id=26493 for details). I mention this because the very large number of detected outliers in your data could just be a symptom of CONN mis-interpreting the amount of motion in your subject(s), which may happen if the format of the subject-motion files that you are entering into CONN does not quite follow CONN's conventions.
Best
Alfonso
Originally posted by boehm:
In addition to what Andrew suggested, another thing that I would check is whether the 'realignment' parameters have been entered correctly into CONN (see this post https://www.nitrc.org/forum/message.php?msg_id=26493 for details). I mention this because the very large number of detected outliers in your data could just be a symptom of CONN mis-interpreting the amount of motion in your subject(s), which may happen if the format of the subject-motion files that you are entering into CONN does not quite follow CONN's conventions.
Best
Alfonso
Originally posted by boehm:
Hi Alfonso,
in Order to be able to use scrubbing via the conn-Toolbox I only run the "functional outlier detection" in the preprocessing Pipeline as other preprocessing steps were processed outside of Conn.
I now have the Impression that the scrubbing covariat is not correct as:
- no normal Distribution of connectivity values after denoising
- warning overdetermined model
- 190P of the scrubbing covariate
When trying to solve my Problem I came across the post below. In the 19.c
Version of conn I cant't find the described 'covariate tools. subject-level aggregate' Option. Where do I find this Information in Version 19.c
I would also highly appreciate any other thoughts/ ideas about my Problem described above.
Thank you
Ilka
in Order to be able to use scrubbing via the conn-Toolbox I only run the "functional outlier detection" in the preprocessing Pipeline as other preprocessing steps were processed outside of Conn.
I now have the Impression that the scrubbing covariat is not correct as:
- no normal Distribution of connectivity values after denoising
- warning overdetermined model
- 190P of the scrubbing covariate
When trying to solve my Problem I came across the post below. In the 19.c
Version of conn I cant't find the described 'covariate tools. subject-level aggregate' Option. Where do I find this Information in Version 19.c
I would also highly appreciate any other thoughts/ ideas about my Problem described above.
Thank you
Ilka
Threaded View
| Title | Author | Date |
|---|---|---|
| Karolina Finc | Feb 17, 2016 | |
| Alfonso Nieto-Castanon | Feb 18, 2016 | |
| boehm | Jul 6, 2020 | |
| Alfonso Nieto-Castanon | Jan 29, 2021 | |
| Andrew Lynn | Jan 27, 2021 | |
| Karolina Finc | Feb 23, 2016 | |