Background and Purpose. Brain atrophy is a proposed marker of disease progression in multiple sclerosis (MS). Many magnetic resonance imaging-based methods of atrophy quantification exist, but their relative sensitivity and precision is unclear. Our aim was to compare atrophy rates from the brain boundary shift integral (BBSI), structural image evaluation, using normalization of atrophy (SIENA) (both registration-based methods) and segmented brain volume difference, in patients with clinically isolated syndromes (CIS), relapsing remitting MS (RRMS), and controls. Methods. Thirty-seven CIS patients, 30 with early RRMS and 16 controls had T1-weighted volumetric imaging at baseline and 1 year. Brain atrophy rates were determined using segmented brain volume difference, BBSI, and SIENA. Results. BBSI and SIENA were more precise than subtraction of segmented brain volumes and were more sensitive distinguishing RRMS subjects from controls. A strong correlation was observed between BBSI and SIENA. Atrophy rates were greater in CIS and RRMS subjects than controls (RRMS P < .001). With all methods, significantly greater atrophy rates were observed in CIS patients who developed clinically definite MS relative to subjects who did not. Conclusion. Registration-based techniques are more precise and sensitive than segmentation-based methods in measuring brain atrophy, with BBSI and SIENA providing comparable results.