Scientists harness AI to predict dementia 15 years before symptoms appear
In the battle against dementia, Chinese scientists have harnessed the power of artificial intelligence (AI) and a vast databank to predict the onset of the disease up to 15 years before symptoms appear.
This research, rooted in the analysis of blood samples from over 50,000 individuals, has identified key plasma biomarkers that signal an increased risk of developing various forms of dementia, including Alzheimer’s disease.
The study’s findings, detailed in a recent publication in the peer-reviewed journal Nature Aging, represent a significant stride toward early detection and intervention in dementia care.
Harnessing AI for early detection
The research team, led by Professor Yu Jintai of Fudan University-affiliated Huashan Hospital, utilized the comprehensive UK Biobank cohort, which included data from over 50,000 people aged 40-69.
These participants contributed blood samples and answered questions about their health and lifestyle. The researchers tracked their health for an average of 14 years, and sadly, over 1,400 developed dementia during that time.
The researchers then used proteomics to analyze thousands of proteins in the participants’ blood plasma (the liquid part of blood). They aimed to identify specific proteins whose levels might change in people who later develop dementia.
To analyze this massive dataset, the researchers employed an AI algorithm called the “light gradient boosting machine.” Using this, the researchers could sift through the data, identifying proteins whose expression levels began to deviate from normal up to a decade before the clinical manifestation of dementia.
“The use of artificial intelligence was ‘one of the key factors for the success of this research,” said Professor Yu Jintai. This machine learning tool proved crucial in the study, offering robust pattern recognition and predictive capabilities that facilitated the efficient analysis of the vast amount of data.
Beyond diagnosis
Blood tests for diagnosing dementia have gained popularity due to their ability to detect the disease from a single drop of blood in symptomatic patients.
This research aimed beyond merely diagnosing dementia in people who already showed symptoms. Instead, the goal was to predict the onset of the disease even before any noticeable symptoms occurred.
This early detection could be crucial for starting treatment sooner, potentially slowing down the progression of dementia. By incorporating protein levels with other information like age and sex, the researchers built a predictive model that could estimate an individual’s risk of developing dementia.
By integrating this protein data with demographic information, such as age and sex, the researchers developed an optimal predictive model. The model successfully identified key plasma biomarkers associated with an increased risk of developing dementia, including Alzheimer’s disease.
This model also offers a non-invasive, cost-effective, and easily accessible method for screening at-risk individuals.
Global implications and future research
While the research primarily focused on a cohort with over 90% white participants, limiting its global applicability, efforts are underway to conduct similar studies on Chinese cohorts.
This will enable the team to compare findings across different populations and enhance understanding of dementia’s biomarkers.
The potential implications of this research are vast, offering hope for developing new treatments and intervention strategies.
As the global population ages, the incidence of dementia is expected to rise, making the need for early detection and intervention methods more critical than ever.
This study represents a significant step forward, laying the groundwork for future advancements in the fight against dementia.
Source: Interesting Engineering
How 40 Hz Light and Sound Waves Could Halt Alzheimer’s
Scientists harness AI to predict dementia 15 years before symptoms appear