 |
 |
I've just emailed a response to this on an earlier thread. All the best, Mark On 9 May 2014, at 18:43, "Gullett, Joseph M" <[log in to unmask]<mailto:[log in to unmask]>> wrote: We recently encountered a problem one participant when using sienax. It appears the SNR or contrast ratio of the T1 image is poor, and we are unable to get a good BET result. When manually using BET, we get poor results even shifting the center point to accommodate the extra neck (and using –N command and adjusting the –f all over the place) and eyeballs that are not being stripped out. I can manually skull strip the data using the mask procedure in fslview and with fslmaths subtraction, but I do not know how to supply sienax with this self-created skull-stripped image. Is that possible? Also, given that the main use we are getting from sienax is the normalization ratio (to use with other ROI’s we are comparing), I am guessing this normalization ratio may not be reliable given the difficulties with BET due to low SNR contrast. Is this assumption correct? Is there a way to see an output of what the algorithm identified as skull and not-skull (such as when the bet procedure produces the .html file:// to inspect with a web browser)? Should I trust the normalization ratio given that the script uses the non-skull stripped image for normalization based on skull? Thanks in advance! ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Joseph Gullett, M.S. Doctoral Candidate Department of Clinical and Health Psychology University of Florida