Fig. 5.

Summary of the statistical analysis on the anatomical and diffusion metrics. To avoid the statistically complex repeated measurement problem, the side difference variable P _HEMI = P _LEFT − P _RIGHT between the value of the parameter P in the left and right hemisphere was first submitted to an MRA with variables age and gender and their interaction term. If one of these variables showed a significant effect, a second MRA was conducted separately for P _LEFT and P _RIGHT in each hemisphere; otherwise a t test was performed on the side variable P _HEMI. If the t test demonstrated that P _HEMI was significantly different from zero, then as previously a second MRA was conducted separately for P _LEFT and P _RIGHT. Otherwise if P _HEMI was not significantly different from zero then a second MRA was performed on the average of P between the two hemisphere, (P _LEFT + P _RIGHT)/2