help
489 Subscribers 
help > Generate Scrubbing Covariate w/ Alternative FD metrics?
Showing 1-5 of 5 posts
May 7, 2020 08:05 PM | Paul Cernasov
Generate Scrubbing Covariate w/ Alternative FD metrics?
Hi Alfonso,
I'm stumped trying to create a new scrubbing covariate using the FD power metric (instead of ARTs default). I use the conn_convertl12l1covariate command (or the GUI) to create my QC_FDpower variable no problem. When I select this new variable as the input for a scrubbing variable it throws an error message (see attached). I looked at the help for the conn_convertl12l1covariate command and it indicates that scrubbing assumes the input contains both the BOLD signal change & the FD change as in QC_timeseries. How can I proceed?
Kindly,
Paul
I'm stumped trying to create a new scrubbing covariate using the FD power metric (instead of ARTs default). I use the conn_convertl12l1covariate command (or the GUI) to create my QC_FDpower variable no problem. When I select this new variable as the input for a scrubbing variable it throws an error message (see attached). I looked at the help for the conn_convertl12l1covariate command and it indicates that scrubbing assumes the input contains both the BOLD signal change & the FD change as in QC_timeseries. How can I proceed?
Kindly,
Paul
May 9, 2020 12:05 AM | Alfonso Nieto-Castanon - Boston University
RE: Generate Scrubbing Covariate w/ Alternative FD metrics?
Hi Paul,
Typically scrubbing expects 2 timeseries in the QC_timeseries input (framewise displacement and global-signal change) instead of just 1 timeseries as in your QC_FDpower input. In any way, it is perfectly fine to define an alternative scrubbing covariate based on thresholding only the QC_FDpower timeseries, you could do that, for example, using the syntax:
conn_convertl12l1covariate('scrubbing' , thr, [], 'QC_FDpower', 'scrubbingFD', []);
where thr is your desired FD threshold (see "doc conn_convertl12l1covariate" for details; the key is simply the [] part at the end which is going to suppress the creation of the default QC_motion and QC_GSchange second-level covariates)
Best
Alfonso
Originally posted by Paul Cernasov:
Typically scrubbing expects 2 timeseries in the QC_timeseries input (framewise displacement and global-signal change) instead of just 1 timeseries as in your QC_FDpower input. In any way, it is perfectly fine to define an alternative scrubbing covariate based on thresholding only the QC_FDpower timeseries, you could do that, for example, using the syntax:
conn_convertl12l1covariate('scrubbing' , thr, [], 'QC_FDpower', 'scrubbingFD', []);
where thr is your desired FD threshold (see "doc conn_convertl12l1covariate" for details; the key is simply the [] part at the end which is going to suppress the creation of the default QC_motion and QC_GSchange second-level covariates)
Best
Alfonso
Originally posted by Paul Cernasov:
Hi Alfonso,
I'm stumped trying to create a new scrubbing covariate using the FD power metric (instead of ARTs default). I use the conn_convertl12l1covariate command (or the GUI) to create my QC_FDpower variable no problem. When I select this new variable as the input for a scrubbing variable it throws an error message (see attached). I looked at the help for the conn_convertl12l1covariate command and it indicates that scrubbing assumes the input contains both the BOLD signal change & the FD change as in QC_timeseries. How can I proceed?
Kindly,
Paul
I'm stumped trying to create a new scrubbing covariate using the FD power metric (instead of ARTs default). I use the conn_convertl12l1covariate command (or the GUI) to create my QC_FDpower variable no problem. When I select this new variable as the input for a scrubbing variable it throws an error message (see attached). I looked at the help for the conn_convertl12l1covariate command and it indicates that scrubbing assumes the input contains both the BOLD signal change & the FD change as in QC_timeseries. How can I proceed?
Kindly,
Paul
May 11, 2020 07:05 PM | Paul Cernasov
RE: Generate Scrubbing Covariate w/ Alternative FD metrics?
Hi Alfonso,
Thank you for the reply and the explanation. I try entering your code below and I get the following error.
conn_convertl12l1covariate('scrubbing' , 0.2, [], 'QC_FDpower', 'scrubbingFD', []);
Error:
Index in position 2 is invalid. Array indices must be positive intergers or logical values.
Error in line 274
y5(nsub)=max(y5(nsub), max(abs(data(:,end-1)),[],1) ); %MaxGSchange
The code will only run if I enter a number in the second bracket, but the resulting scrubbing variable is always empty. Would you be able to advise?
Best,
Paul
Originally posted by Alfonso Nieto-Castanon:
Thank you for the reply and the explanation. I try entering your code below and I get the following error.
conn_convertl12l1covariate('scrubbing' , 0.2, [], 'QC_FDpower', 'scrubbingFD', []);
Error:
Index in position 2 is invalid. Array indices must be positive intergers or logical values.
Error in line 274
y5(nsub)=max(y5(nsub), max(abs(data(:,end-1)),[],1) ); %MaxGSchange
The code will only run if I enter a number in the second bracket, but the resulting scrubbing variable is always empty. Would you be able to advise?
Best,
Paul
Originally posted by Alfonso Nieto-Castanon:
Hi Paul,
Typically scrubbing expects 2 timeseries in the QC_timeseries input (framewise displacement and global-signal change) instead of just 1 timeseries as in your QC_FDpower input. In any way, it is perfectly fine to define an alternative scrubbing covariate based on thresholding only the QC_FDpower timeseries, you could do that, for example, using the syntax:
conn_convertl12l1covariate('scrubbing' , thr, [], 'QC_FDpower', 'scrubbingFD', []);
where thr is your desired FD threshold (see "doc conn_convertl12l1covariate" for details; the key is simply the [] part at the end which is going to suppress the creation of the default QC_motion and QC_GSchange second-level covariates)
Best
Alfonso
Typically scrubbing expects 2 timeseries in the QC_timeseries input (framewise displacement and global-signal change) instead of just 1 timeseries as in your QC_FDpower input. In any way, it is perfectly fine to define an alternative scrubbing covariate based on thresholding only the QC_FDpower timeseries, you could do that, for example, using the syntax:
conn_convertl12l1covariate('scrubbing' , thr, [], 'QC_FDpower', 'scrubbingFD', []);
where thr is your desired FD threshold (see "doc conn_convertl12l1covariate" for details; the key is simply the [] part at the end which is going to suppress the creation of the default QC_motion and QC_GSchange second-level covariates)
Best
Alfonso
May 13, 2020 03:05 PM | Alfonso Nieto-Castanon - Boston University
RE: Generate Scrubbing Covariate w/ Alternative FD metrics?
Hi Paul,
Sorry about that, my mistake, the correct syntax was:
conn_convertl12l1covariate('scrubbing' , 0.2, [], 'QC_FDpower', 'scrubbingFD', false);
(with "false" instead of "[ ]" at the end). If the resulting scrubbing variable is empty that would indicate that there are not scans with FD values above the chosen threshold (e.g. 0.2 above). Let me know if that seems to work correctly now.
Best
Alfonso
Originally posted by Paul Cernasov:
Sorry about that, my mistake, the correct syntax was:
conn_convertl12l1covariate('scrubbing' , 0.2, [], 'QC_FDpower', 'scrubbingFD', false);
(with "false" instead of "[ ]" at the end). If the resulting scrubbing variable is empty that would indicate that there are not scans with FD values above the chosen threshold (e.g. 0.2 above). Let me know if that seems to work correctly now.
Best
Alfonso
Originally posted by Paul Cernasov:
Hi Alfonso,
Thank you for the reply and the explanation. I try entering your code below and I get the following error.
conn_convertl12l1covariate('scrubbing' , 0.2, [], 'QC_FDpower', 'scrubbingFD', []);
Error:
Index in position 2 is invalid. Array indices must be positive intergers or logical values.
Error in line 274
y5(nsub)=max(y5(nsub), max(abs(data(:,end-1)),[],1) ); %MaxGSchange
The code will only run if I enter a number in the second bracket, but the resulting scrubbing variable is always empty. Would you be able to advise?
Best,
Paul
Originally posted by Alfonso Nieto-Castanon:
Thank you for the reply and the explanation. I try entering your code below and I get the following error.
conn_convertl12l1covariate('scrubbing' , 0.2, [], 'QC_FDpower', 'scrubbingFD', []);
Error:
Index in position 2 is invalid. Array indices must be positive intergers or logical values.
Error in line 274
y5(nsub)=max(y5(nsub), max(abs(data(:,end-1)),[],1) ); %MaxGSchange
The code will only run if I enter a number in the second bracket, but the resulting scrubbing variable is always empty. Would you be able to advise?
Best,
Paul
Originally posted by Alfonso Nieto-Castanon:
Hi Paul,
Typically scrubbing expects 2 timeseries in the QC_timeseries input (framewise displacement and global-signal change) instead of just 1 timeseries as in your QC_FDpower input. In any way, it is perfectly fine to define an alternative scrubbing covariate based on thresholding only the QC_FDpower timeseries, you could do that, for example, using the syntax:
conn_convertl12l1covariate('scrubbing' , thr, [], 'QC_FDpower', 'scrubbingFD', []);
where thr is your desired FD threshold (see "doc conn_convertl12l1covariate" for details; the key is simply the [] part at the end which is going to suppress the creation of the default QC_motion and QC_GSchange second-level covariates)
Best
Alfonso
Typically scrubbing expects 2 timeseries in the QC_timeseries input (framewise displacement and global-signal change) instead of just 1 timeseries as in your QC_FDpower input. In any way, it is perfectly fine to define an alternative scrubbing covariate based on thresholding only the QC_FDpower timeseries, you could do that, for example, using the syntax:
conn_convertl12l1covariate('scrubbing' , thr, [], 'QC_FDpower', 'scrubbingFD', []);
where thr is your desired FD threshold (see "doc conn_convertl12l1covariate" for details; the key is simply the [] part at the end which is going to suppress the creation of the default QC_motion and QC_GSchange second-level covariates)
Best
Alfonso
May 13, 2020 10:05 PM | Paul Cernasov
RE: Generate Scrubbing Covariate w/ Alternative FD metrics?
Alfonso,
That syntax works. Thank you so much.
I hope it's not cliche but CONN Toolbox is really the bomb and makes grad school kinda fun!
Cheers,
Paul
That syntax works. Thank you so much.
I hope it's not cliche but CONN Toolbox is really the bomb and makes grad school kinda fun!
Cheers,
Paul