Fig. 2.

Variables of importance for the separation between CTL versus AD. a subcortical volumes (normalized by intra cranial volume) b cortical gray matter volumes (normalized by intra cranial volume) c cortical thickness (raw data). Measures above zero have a larger value in AD compared to controls and measures below zero have a lower value in AD compared to controls. A measure with a high covariance is more likely to have an impact on group separation than a measure with a low covariance. Measures with jack knifed confidence intervals that include zero have low reliability