deletions | additions       diff --git a/Initially_I_had_marked_a__.tex b/Initially_I_had_marked_a__.tex  index 43e4d6a..534424c 100644  --- a/Initially_I_had_marked_a__.tex  +++ b/Initially_I_had_marked_a__.tex 
...
Initially I had marked a few scans that had small amounts of dura classified as brain at the top of the head (that I was planning to edit later), which I had mistakenly excluded. We then included those scans to see if it affected the results (in particular, $CV_a$, as this is the most important for power calculations) and found that including these subjects, thankfully, did not change the results. We have now included the results from revised dataset in the manuscript. For the new QC procedure, I followed your advice to look at eTIV. This makes more sense because this is the one of the first steps in recon-all and a bad registration to the talairach template could lead to innacurate segmentations overall. We left out 3 scans in total, 1 scan from 1 subject from sites 4,6 and 11 that had very bad coregistration. From sites 4 and 6, we left out one scan of the same subject, and at site 11 it was a different subject. Other left out scans were due to time constraints (we didn't have time to scan all the subjects), one scan was left out because part of the head was not acquired (possibly the files were corrupted), and one scan was left out because it was an identical copy of the first instead of a rescan (perhaps a problem with transferring data). We've added a column in the acquisition tables to reflect how many actual scans were acquired at each site. site, and have added information about why the scans were left out in the Methods section.