Researchers at Massachusetts General Hospital and Brigham and Women's Hospital have developed a revolutionary artificial intelligence (AI) tool capable of predicting brain deterioration in patients years in advance, potentially enabling earlier treatment. This research, published in the Journal of Alzheimer's Disease, shows the AI tool uses electroencephalography (EEG) to analyze subtle changes in brain activity during sleep to make these predictions.
The research team followed 281 women aged 65 and older with normal cognitive function at the time of their initial sleep study for five years. At the second assessment, 96 participants had developed cognitive impairment. Researchers extracted brainwave patterns from EEG data, using AI to detect minute changes in gamma wave frequency during deep sleep. The AI tool successfully flagged 85% of individuals who ultimately experienced cognitive decline, achieving an overall accuracy of 77%.
Image Note: Image generated by AI, licensed through Midjourney.
Lead author Dr. Haha Yehe stated, "This could revolutionize how we approach dementia prevention." He added, "Through novel sophisticated analytical techniques, advanced information theory tools, and AI, we can detect subtle changes in sleep brainwave patterns years before symptoms appear, providing a valuable window for intervention."
According to the World Health Organization (WHO), approximately 55 million people worldwide live with dementia. Numerous studies show that changes in behavior and physiological function may occur before the onset of cognitive impairment. Therefore, identifying these early signals could offer a better window for effective intervention.
This AI-powered early detection system allows individuals at risk to make proactive lifestyle adjustments, such as exercise, mental stimulation, and a balanced diet, before cognitive impairment develops. These changes may help protect cognitive health and delay the onset of dementia.
However, researchers emphasize the need for larger, more diverse studies to validate and expand upon these findings.