Purpose: To investigate whether the method of applying image masks can alter quantifiable measures determined from whole-brain MTR calculations.
Materials and methods: Thirty-five T1/MT image pairs were obtained from five normal volunteers. For each pair a mask was used to specify the regions to be analyzed. Using these regions, a histogram was used to calculate seven global MTR metrics. This process was performed three ways: 1) using a unique mask for each T1/MT pair, 2) sharing a single mask for each subject and registering all intrasubject images to the image corresponding to their mask, and 3) sharing a single mask for each subject and transforming that mask into alignment with each of their original T1/MT image pairs.
Results: With respect to the first method, the latter two methods caused small but significant differences in several parameters.
Conclusion: The method of applying image masks can affect whole-brain MTR values.