Researchers have unveiled an artificial intelligence (AI) model capable of analyzing electrocardiogram (ECG) data to identify signs of premature aging and cognitive decline. This innovative approach, detailed in a study to be presented at the American Stroke Association's International Stroke Conference 2025, leverages a deep neural network (DNN) to classify individuals based on their ECG-determined biological age, offering insights into their cognitive health.
The study, which examined data from over 63,000 UK Biobank participants, categorized subjects into normal aging, accelerated ECG-aging, and decelerated ECG-aging groups. Findings indicated a notable correlation between ECG-age and cognitive performance, with those in the decelerated ECG-aging group outperforming others on cognitive tests. This suggests that ECG-age, reflecting the heart's functional status and potentially overall organism health, could serve as a novel biomarker for cognitive health.
Bernard Ofosuhene, the study's lead author, highlighted the significance of ECG-age as a more nuanced measure of biological aging compared to chronological age. The research opens new avenues for cognitive health assessments, potentially utilizing ECG data from routine doctor visits or wearable devices. This could democratize access to cognitive evaluations, particularly in regions lacking specialized neuropsychiatric services.
Despite its promise, the study acknowledges limitations, including its focus on a specific age range and predominantly European ancestry participants, which may affect the findings' broader applicability. Future research directions include exploring gender differences and extending the study to more diverse populations.
While the AI model's ability to predict cognitive decline from ECG data is a groundbreaking development, researchers caution that further validation is necessary. The potential for early detection and intervention in cognitive decline, however, underscores the importance of this research in the fields of healthcare and artificial intelligence.
