Figure 6

Illustration of subjects raw performances in the boxplots titled 'Data' (the red lines indicate the median, the blue boxes extend from the upper to the lower quartile values, the whiskers show the most extreme points within 1.5 times the inter-quartile range and the red plus signs indicate outliers) and bootstrapped performances with linear adjustments (graphic lines titled 'Means and linear fits'; vertical lines show the standard deviation of the bootstrapped means and the red lines the linear fits). Distribution plots titled 'Non linear F values' show the distributions of F values measuring the distances between the bootstrapped data and the bootstrapped regression lines. Non central distribution suggests non linear effects; significant effects are marked with a star. For 'Data' and 'Means and linear fits' graphics, groups are ordered by increasing/decreasing mean values: (1) LCD/LLN, (2) LCD/HLN, (3) HCD/HLN, (4) HCD/LLN, (5) Controls. By contrast, the last (right end side) graphics show the bootstrapped data (blue circles) with smoothed interpolated data (in green – piecewise cubic hermite interpolating polynomial). Best scores (controls) reflected an optimal in the volume distribution: low cerebellar volumes on the left (groups 1 and 2) and high cerebellar volumes on the right (3 and 4).