Age estimation from brain MRI images using machine learning techniques and its applications

Statistical analysis using large-scale brain magnetic resonance (MR) image databases has observed that brain tissues have presented age-related morphological changes. This result indicates that the age of a subject can be estimated from his/her brain MR image by evaluating morphological changes in healthy aging. We have explored brain local features, which are useful for analyzing brain MR images. The brain local features are defined by volumes and cortical thickness of brain tissues parcellated into local regions defined by the automated anatomical labeling atlas. An age can be estimated by the machine learning approach with brain local features extracted from T1-weighted MR images. We consider using convolutional neural networks (CNNs) to extract brain features, where any medical knowledge is not required to define local features. We evaluate the performance of the proposed approaches using large-scale MR image databases. We also consider applying the proposed approaches to identify Alzheimer’s disease from brain MR images.