Brain age prediction improves the early detection of Alzheimer’s disease in East Asian elderly

Abstract

AbstractIntroductionBrain age prediction is used to quantify the pathological and cognitive changes associated with brain aging. However, the predicted age derived from certain models can result in biased estimation and the concealment of inherent aged brain function.MethodsWe constructed a brain age prediction model for the East Asian elderly brain using the brain volume and cortical thickness features from cognitively normal (CN) brains. Furthermore, our model was used to estimate different diagnoses and to construct a classification model of mild cognitive impairment (MCI) conversion and Alzheimer’s disease (AD) conversion.ResultsOur model showed a strong association of the brain age difference (BAD) with three diagnosis groups. In addition, the classification models of MCI conversion and AD conversion showed acceptable and robust performances, respectively (area under the curve [AUC] = 0.66, AUC = 0.76).DiscussionWe believe that our model can be used to estimate the predicted status of an East Asian elderly brain. Moreover, the MCI conversion model has the potential to prevent severe cognitive impairment and can be used for the early detection of AD.