Researchers trained artificial intelligence model to measure people’s age from brain scans
Laboratory/Alamy
The abundance of 13 types of proteins in the blood appears to be a strong indicator of how quickly the brain is aging. This suggests that blood tests could one day help people track and even improve their brain health.
Most previous studies that looked at protein markers of brain aging in the blood involved fewer than 1,000 people, said Nicholas Seyfried of Emory University in Atlanta, Georgia. is not participating.
To understand the influence of these proteins more broadly, Wei-Shi Liu and colleagues at Fudan University in China surveyed approximately 11,000 adults (ranging in age from approximately 50 to 80 years at the time) who participated in the UK Biobank project. ) analyzed MRI brain scan data. of imaging.
Liu’s team trained an artificial intelligence model using data from 70% of the participants to determine features of brain images, such as the size of different brain regions and how different parts are connected to each other. The age of the participants was predicted based on When the model was applied to the remaining 30% of participants, its predictions were accurate to within 2.7 years of their actual age.
The researchers then used the model to predict the age of another group of about 4,700 people, with an average age of 63, who also underwent brain imaging for UK Biobank. The researchers calculated the difference between these participants’ actual ages and their AI-predicted ages, called the brain age gap. “The higher the age predicted by the AI compared to the actual age, the faster the brain ages,” Liu says.
The group also provided blood samples around the same time as their brains were imaged. Based on this, the research team identified eight proteins that appear to increase significantly and five proteins that decrease in abundance as the difference in brain age increases.
In an analysis of data from previous studies, researchers confirmed that these proteins are produced by brain cells and that their levels can influence the risk of dementia and stroke.
This suggests that blood tests for these proteins may reveal how quickly the brain ages. “These markers could be the canaries in the coal mine that say, ‘Hey, here we go, let’s start doing interventions to slow brain aging while there’s still plenty of time,'” Seyfried said.
But for this to be helpful, we need to know that these proteins can change with lifestyle changes. “You want to be able to say, ‘If I run this much, I’ll lose this much weight, and if I change my diet, I can correct those levels and bring them back into the normal range,'” Seyfried says.
Because the study was conducted primarily among wealthy white people, Seyfried said more research is needed to see if the results hold true for other populations with more diverse ethnicities and income levels.
The research team now hopes to conduct studies in animals to determine exactly how the 13 proteins affect the brain. For example, researchers might test whether disrupting levels of these proteins affects cognition or even the development of neurodegenerative conditions, Liu says. “In the coming decades, this could open up ways to target proteins to slow aging and disease.”
topic:
