<div dir="ltr"><div><div><div><div><div><div><div>Hi Camino Experts,<br></div>  I am new to the software and have one or two questions related to errors I received in processing.<br><br></div>1) selectshells - in the web page (<a href="http://camino.cs.ucl.ac.uk/index.php?n=Man.Selectshells">http://camino.cs.ucl.ac.uk/index.php?n=Man.Selectshells</a>) there is a flag named -unweightedb which specifies shells representing B=0 which are not actually at that value (i.e. HCP data where b=5 is the b=0 condition). I tried the following command with the latest binary:<br><br>selectshells -inputfile tensor.nii -schemefile dwi.scheme -maxbval 1.5E09 -outputroot test -minbval 1.1E07 -unweightedb 1E07<br><br></div>In this case b< 10 (1E7) should be the max b=0 condition, and 11 (1.1E07) <= b <= 1500 (1.5E9) should be the range of values included in the subshell. I keep getting an unable to parse -unweightedb error no matter how I try it. Is this option OBSOLETE?<br><br><br>If I use -minbval with the default value of 0 , what happens in the case where my minimum is b=5? Will it automatically detect that b=5 is the lowest number and assume this is the unweighted option since -unweightedb does not work?<br><br><br></div>2) Restore - In some of the Camino user posts the HCP b=5 would create errors since a B=0 is not detected<br><br></div>a) Is this issue resolved or is the solution of manually changing scheme files where b=5 to b=0 the way it works?<br></div><div>--If it is resolved, can the calculations handle b=5 or does the program take the lowest b value and assume it is b=0 internally or is the scheme file modified? I am curious how this was handled.<br></div><br>b) Can Restore algorithm make more accurate calculations using all 3 shells (1k2k3k) or is < 1500 still the preferred route? My guess is that it is NOT able to use all three shells, but I wanted to verify.<br><br></div><div><div><div><div><div>c) As for the noise estimate I took the sigma of the background [1-head mask (large to avoid ghosting/inside head) ]. This works well for Siemens data, however from my understanding Philips may do some sort of masking/noise suppression outside of the brain. Has this been an issue in the past and if so have you tried another region to "estimate noise" inside the brain (i.e gap between brain and skull etc) for Philips data in the past?<br><br><br></div><div>Thanks,<br></div><div>Ajay<br></div></div></div></div></div></div>