Hi FreeSurfers

 

We are looking at cortical thickness after spinal cord injury in a pooled sample from three sites. All have 1x1x1 mm T1. Our current approach has been to exclude noisy data (16/159 participants), and to account for scanner differences using site as dummy variables.

 

We are considering using SynthSR  for two purposes:

1. reduce motion artefacts in (some) scans
2. better merge data between different scanner sites

 

If using SynthSR for noise reduction seems like a good idea, is it best to then run SynthSR on ALL datasets for consistency?

 

 

Best

Clas Linnman