Hey Doug,

When using the bbregister feature to mask areas with b0 distortion, I find noticeably lower registration cost values (as revealed by the first number in the .mincost file that gets generated). When this number is calculated, does it take into account the registration in the areas affected by the mask? In other words, are the lower numbers evidence that masking distorted regions is significantly improving the registration accuracy, or just an artifact of masking out regions where the is bad registration from the cost calculation?

Thanks! Mike