Skip to main content
. 2018 Jun 11;24(2):266–281. doi: 10.1038/s41380-018-0098-1

Fig. 2.

Fig. 2

Brain-predicted age and brain-PAD. a The results of using Gaussian processes regression to predict chronological age using structural neuroimaging data in a sample of 2001 healthy individuals, aged 18–90 years, based on ten-fold cross validation (mean absolute error = 4.93 years, r = 0.94). b Same as a, with the brain-predicted age values for participants from the Lothian Birth Cohort 1936 overlaid in red. This demonstrates that, despite the narrow age range at time of scanning (72–74 years), these N = 669 individuals had a wide variability in brain-predicted ages. c Illustration of how brain-predicted age difference (brain-PAD) scores are calculated, highlighting the individuals from the Lothian Birth Cohort 1936 with the lowest and highest brain-predicted ages. Brain-PAD is the difference between brain-predicted age and chronological age for an individual. Positive brain-PAD suggests that the individual’s brain appears older than their chronological age, whereas a negative brain-PAD suggests that their brain appears younger