Can measuring hippocampal atrophy with a fully automatic method be substantially less noisy than manual segmentation over both 1 and 3 years?

•Robust assessment of segmentation noise of atrophy measuring methods in Alzheimer's.•The noise of all methods violated the Gaussianity assumption of common statistics.•Using robust statistics FreeSurfer 5.3.0 had the same noise as manual segmentation.•MAPS-HBSI and FreeSurfer 6.0.0 were substantially less noisy than manual.•MAPS-HBSI and FreeSurfer 6.0.0 required less than two thirds of the subjects of manual.