Dear Alfonso and Conn users,
I recently submitted a paper about resting state fMRI using Conn and in particular Independent Component Analysis to compare patients and controls.
I have 2 questions of reviewers and I don't know how to improve these points:
-I chose different number of components (arbitrary cutpoints:10, 20, 30, 40...) to study the reliability of the results.
However, one of the reviewer stated that the dataset in the Buckner paper suggested a natural cut points in the data to analyze the DMN at 7 and 17 networks and asked "It might be better to use the data to figure how many components to model instead of setting arbitrary cutpoints."
Is it possible to do this in Conn?
- One reviewer stated that the tools for meaningfully addressing the known noise in rsFMRI are fairly limited in SPM based packages and asked how were respiratory, cardiac etc signal removed and why we did not used ICAfix. I replied that CONN uses a component-based noise reduction (CompCor) a method that performs principal component analysis to estimate the physiological noise from white matter and cerebrospinal fluid for each participant. Is CompCor enough to remove respiratory and cardiac signal or should we also use an ICA method?
Thank you so much for your help!
Sincerely yours.
Mickael
I recently submitted a paper about resting state fMRI using Conn and in particular Independent Component Analysis to compare patients and controls.
I have 2 questions of reviewers and I don't know how to improve these points:
-I chose different number of components (arbitrary cutpoints:10, 20, 30, 40...) to study the reliability of the results.
However, one of the reviewer stated that the dataset in the Buckner paper suggested a natural cut points in the data to analyze the DMN at 7 and 17 networks and asked "It might be better to use the data to figure how many components to model instead of setting arbitrary cutpoints."
Is it possible to do this in Conn?
- One reviewer stated that the tools for meaningfully addressing the known noise in rsFMRI are fairly limited in SPM based packages and asked how were respiratory, cardiac etc signal removed and why we did not used ICAfix. I replied that CONN uses a component-based noise reduction (CompCor) a method that performs principal component analysis to estimate the physiological noise from white matter and cerebrospinal fluid for each participant. Is CompCor enough to remove respiratory and cardiac signal or should we also use an ICA method?
Thank you so much for your help!
Sincerely yours.
Mickael