 |
 |
Hello, I have a concern about activations that are occurring outside of the brain in our data. What we are getting seems to be further outside the brain than might be normal. I attached an example of this in sample_image_Joyce(3), which represents a zstat image for a single contrast, thresholded from 3.1 to 7, for a single subject, averaged across runs. We are wondering whether this might be a BET or a registration issue. The BET issue may be that we have created functional masks that are too large compared to the structural images. The BET parameters we used for the structural images were -B -f 0.25 (whereas the default parameters for the structural BET in FSL appear to be bet2 -f 0.5). For the functional data, we kept the default BET parameters: "bet2 mean_func mask -f 0.3 -n -m". It also seems that we are achieving less than ideal alignment between the standard brain image and the functional data in our registrations. This could be the result of a BET issue, though we are wondering whether problems with registration alone could be causing our extra-brain activations. Any input or suggestions would be much appreciated. Thanks very much, Rachel