Atilla Kiraly, Google, 2018, Predicting Brain Age Using Structural Neuroimaging and Deep Learning – Google Research
<p>December 30, 2019, 2:12 PM</p>
<p> </p>
<p>Early detection of aging-related diseases requires a model of the underlying biological aging process. In this paper, we develop a <strong>brain-age predictor by using structural magnetic resonance imaging (SMRI</strong>) and deep learning and evaluate the predicted brain age as a marker of brain- aging. Our approach does not require any domain knowledge in that it uses a transfer-learning paradigm and has been validated on real SMRI data collected from elderly subjects. We developed two different predictive models by using <strong>convolutional neural network (CNN) based regression and bucket classification to predict brain ages from SMRI images.</strong> Our models achieved root mean squared errors (RMSE) of 5.54 and 6.44 (years) in predicting brain ages of cognitively normal subjects. Further analysis showed that there is a substantial difference between the predicted brain ages of cognitively impaired subjects and normal subjects within the same chronological age group.</p>