Comprehensive Summary
This paper, written by Luma Srour et al, discusses previous studies which use AI and neural networks for biological age estimation. Several techniques for biological age estimation were covered in the article, including blood and omic based aging clocks which focused on biomarkers in the blood to predict age, psychological aging clocks which used factors such as health and social data, and deep imaging aging clocks which looked at a person’s skin to determine age. Luma Srour et al discussed a group that used 4846 Chinese participants on blood based data and got a mean absolute error of 5.68 years through deep learning methods. Overall, results showed that AI and neural networks were able to capture the non-linear and complex relationship between each of these parameters and biological ages in an effective and accurate way.
Outcomes and Implications
Aging has been a great topic of interest in the medical field, and there have been drugs and medications prescribed for increased lifespan and anti-aging properties. In order to determine the effectiveness of these strategies, there needs to be an accurate way to determine biological aging in a diverse group of individuals. This paper speaks on the potential of AI and neural networks to create deep aging clocks. The authors note future directions associated with such neural network models, including a need for more ethnically diverse data, and the challenges with getting easily interpretable results.