Moreover, the analysis of human locomotion has already led to widely accepted health recommendations, such as the “10,000 steps per day” minimum activity advisory 4, which is now a basic health recommendation in many wellness applications. This tracking is already done without interfering with the daily routines of hundreds of millions of people all over the world. Instead, the recent introduction of low-power and compact sensors, based on micro-electromechanical systems (MEMS) has led to a new breed of the wearable and affordable devices providing unparalleled opportunities for the collecting and cloud-storing personal digitized activity records in a fully standardized and controlled way.
We note, that large-scale biochemical or genomic profiling is not impossible, but is still logistically difficult and expensive. A broader acceptance of the biomarkers of age, however, will depend on a better understanding of the observed correlations to the incidence of specific diseases, improved transferability of the models across populations and reduction of costs of the studies. This observation has led to the growing popularity of “biological clocks”, designed as linear predictors of chronological age from, e.g., DNA methylation 1, gene expression 2, plasma proteome 3 profiles, etc. Many physiological parameters demonstrate profound correlations with age. Altogether, our findings demonstrate the emerging potential of combined wearable sensors and deep learning technologies for applications involving continuous health risk monitoring and real-time feedback to patients and care providers. As a demonstration, we produced a separate deep CNN for mortality risks prediction that outperformed any of the biological age or a simple linear proportional hazards model. Consequently, we turned to the NHANES death register directly and introduced a novel way to train parametric proportional hazards models suitable for out-of-the-box implementation with any modern machine learning software. We found that the supervised approaches produce better chronological age estimations at the expense of a loss of the association between the aging acceleration and all-cause mortality. In this work, we used one-week long physical activity records from a 2003–2006 National Health and Nutrition Examination Survey (NHANES) to compare three increasingly accurate biological age models: the unsupervised Principal Components Analysis (PCA) score, a multivariate linear regression, and a state-of-the-art deep convolutional neural network (CNN). Naturally, linear combinations of physiological measures trained to estimate chronological age have recently emerged as a practical way to quantify aging in the form of biological age.
Age-related physiological changes in humans are linearly associated with age.
0 kommentar(er)
