Fig. 4.

(a) Boxplots and histogram of cortical volume change comparing study durations and image resolutions using the Jacobian integration method. From left to right (year-one high-resolution, year-one low-resolution, year-two high-resolution, year-two low-resolution, year-three high-resolution, and year-three low-resolution), the average rates of cortical atrophy were − 0.56 ± 0.79, − 0.52 ± 0.87, − 0.48 ± 0.43, − 0.42 ± 0.47, − 0.43 ± 0.27, and − 0.41 ± 0.28% per year. (b) The sample size per arm required for varying treatment effect with fixed power of 80% and 0.05 significance level. The use of high resolution MRIs reduced the required sample size by 34% on average. The reduction was greater for short duration or small treatment effect. Longer study duration consistently decreased the required sample size (41% per year on average).