Hi,
I have some BOLD data from a fast event-related experiment with 3 experimental conditions, 1, 2 and 3. A reviewer wants me to isolate the signal for condition 1. Under the assumptions of the GLM, the signal is a linear combination of A1+B2+C3+Error, so it seems that I need to construct a design matrix that contains only the onsets for B and C, and then run selxavg3-sess with the —svres argument given. The residuals should thus be the best estimate available of condition 1 + error. So in a nutshell, I want the best approximation of the preprocessed BOLD time series, as though conditions 2 and 3 never happened, but without any other transformation of the BOLD data.
I have a student working on this, but he’s running into a problem of the residuals no longer being in anatomical space, or at least the same anatomical space he started in (mri_segstats is reporting a dimension mismatch, which is why I emphasized above I don’t want any other transformations applied to the data). The data were originally processed in fsaverage space. Barring my student accidentally switching to the subject’s native surface space, is there any reason what I’ve described above shouldn’t work?
/********************************************** * Chris McNorgan * Assistant Professor * Department of Psychology * University at Buffalo, * The State University of New York * http://ccnlab.buffalo.edu/ http://ccnlab.buffalo.edu/ * Office: 716.645.0236 * Lab: 716.645.0222 **********************************************/
External Email - Use Caution
What you have described should work. The residuals will be in whatever space the analysis was made in (mkanalysis-sess). What ROI are you trying to get an average over? What are your mkanalysis-sess and mri_segstats commands?
On 05/22/2018 10:26 PM, Mcnorgan, Christopher wrote:
Hi,
I have some BOLD data from a fast event-related experiment with 3 experimental conditions, 1, 2 and 3. A reviewer wants me to isolate the signal for condition 1. Under the assumptions of the GLM, the signal is a linear combination of A1+B2+C3+Error, so it seems that I need to construct a design matrix that contains only the onsets for B and C, and then run selxavg3-sess with the —svres argument given. The residuals should thus be the best estimate available of condition 1 + error. So in a nutshell, I want the best approximation of the preprocessed BOLD time series, as though conditions 2 and 3 never happened, but without any other transformation of the BOLD data.
I have a student working on this, but he’s running into a problem of the residuals no longer being in anatomical space, or at least the same anatomical space he started in (mri_segstats is reporting a dimension mismatch, which is why I emphasized above I don’t want any other transformations applied to the data). The data were originally processed in fsaverage space. Barring my student accidentally switching to the subject’s native surface space, is there any reason what I’ve described above shouldn’t work?
/**********************************************
- Chris McNorgan
- Assistant Professor
- Department of Psychology
- University at Buffalo,
- The State University of New York
- http://ccnlab.buffalo.edu/
- Office: 716.645.0236
- Lab: 716.645.0222
**********************************************/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Douglas N. Greve -
Mcnorgan, Christopher