Neonatal surface registration accuracy

1

Dear all.

I am registering my neonatal data to the dHCP cortical template and I would like to know how I could quantify the accuracy of the registration, is there any metric that I could obtain to measure this? Or the only way is to visualize all the subjects? For reference, I am using this script. Thanks in advance.

Best regards,

Manuel

2

You can optimise the registration parameters to achieve the best ratio of feature map correlation relative to areal distortion, whilst advisably also visually QCing. Note cortical folding patterns are so variable in the frontal lobe especially that perfect overlap is not achievable. Thus you should be checking correspondence of major folds in areas of the brain where folding patterns are known to be reasonably consistent - motor and sensory sulcus, temporal lobe gyri - occipital gyri.

3

Hi Emma,

Thanks for your response. To visually quantify this, I am propagating the label maps and looking if they have “sense” if the areas are well mapped in the template. Is this the correct map to visualize it? or do you suggest other maps/contrast? Thanks in advance.

Best regards,

Manuel