Publish Date:2024-04-09
The AI model has an area under the curve of 87% and a sensitivity of 91.7%.
China Medical University Hospital (CMUH) has developed 'BrainHealth', an AI dementia diagnosis system that can assess the severity in just one minute.
The AI model has an area under the curve (AUC) of 87% and a sensitivity of 91.7%.
An 81-year-old patient experienced increasing memory and communication difficulties. At CMUH, he underwent traditional evaluations, including blood tests, mental capacity assessments, and an MRI. The results showed brain atrophy and a lower-than-average Mini Mental State Examination (MMSE) score for his age.
These data from the tests, along with additional MRI findings, were input into the 'BrainHealth' system.
The AI predicted that the patient's brain was five years older than his actual age and indicated early-stage dementia, aligning with the conventional examination results.
The AI Brain Age Prediction System, a component of 'BrainHealth', uses MRI data from thousands of healthy individuals and dementia patients to create a normal distribution curve of brain regions.
Furthermore, BrainHealth's hereditary neurological disorder testing system detects Alzheimer's disease by using an AI model that compares brain waves and genetic characteristics against data from elderly individuals with and without neurological disorders.
The BrainHealth AI system also features a Neural Gene Discrimination System to allow complex assessments of dementia severity.
Recently, CMUH developed a new home-based system, Intelligent Detection of Respiratory Events through Automated Monitoring (iDREAM), for detecting sleep apnoea.
Featuring Quanta’s QOCA Portable ECG Monitoring Device, the iDREAM system demonstrated an accuracy rate of 95.8% for identifying severe obstructive sleep apnoea syndrome (OSAS).
Using AI, the device can accurately detect severe obstructive OSAS during sleep sessions at home.
Ref.: GlobalData