Fig. 1

Generating a reference system for rating an individual’s cortical thickness data with respect to a control population. In methods 3 and 4, each cortical thickness map from a population sample (a) was divided into 68 distinct brain regions (borders are indicated as black lines in b). Given that the different brain regions have different means and standard deviations (c), this approach is biologically more plausible than generating one common reference system for all brain regions (as was tested here in methods 1 and 2). Based on these null distributions (see d for an example), the observed values for an individual can be rated within the control population (see red line in e) and statistically significant cortical thinning (i.e. atrophy) can be assessed