When does brain aging accelerate? Dangers of quadratic fits in cross-sectional studies

• Published in: NeuroImage (Orlando, Fla.) Vol. 50; no. 4; pp. 1376 - 1383
• Main Authors: Fjell, Anders. M., Walhovd, Kristine B., Westlye, Lars T., Østby, Ylva, Tamnes, Christian K., Jernigan, Terry L., Gamst, Anthony, Dale, Anders M.
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.05.2010; Elsevier Limited
• Subjects: Adolescent; Adult; Age; Aged; Aged, 80 and over; Aging; Aging - pathology; Brain - pathology; Brain research; Child; Cognition & reasoning; Cross-Sectional Studies; Family medical history; Female; FreeSurfer; Hippocampus; Hippocampus - pathology; Humans; Image Processing, Computer-Assisted - methods; Local smoothing; Magnetic Resonance Imaging; Male; Middle Aged; Models, Neurological; Nonparametric; Organ Size; Polynomial models; Quadratic functions; Smoothing spline; Studies; Young Adult; Nonparametric; Magnetic resonance imaging; Quadratic functions; Smoothing spline; Aging; Polynomial models; FreeSurfer; Hippocampus; Local smoothing
• ISSN: 1053-8119; 1095-9572
• DOI: 10.1016/j.neuroimage.2010.01.061

Many brain structures show a complex, non-linear pattern of maturation and age-related change. Often, quadratic models (β0+β1age+β2age2+ε) are used to describe such relationships. Here, we demonstrate that the fitting of quadratic models is substantially affected by seemingly irrelevant factors, such as the age-range sampled. Hippocampal volume was measured in 434 healthy participants between 8 and 85 years of age, and quadratic models were fit to subsets of the sample with different age-ranges. It was found that as the bottom of the age-range increased, the age at which volumes appeared to peak was moved upwards and the estimated decline in the last part of the age-span became larger. Thus, whether children were included or not affected the estimated decline between 60 and 85 years. We conclude that caution should be exerted in inferring age-trajectories from global fit models, e.g. the quadratic model. A nonparametric local smoothing technique (the smoothing spline) was found to be more robust to the effects of different starting ages. The results were replicated in an independent sample of 309 participants.