This article discusses a study that used machine learning (ML) algorithms to predict the existence of brain Aβ+ plaque among SCI or MCI with no less than 80% accuracy. The study used resting state eyes-closed EEG from SCI or MCI populations, each of which also underwent amyloid PET analysis for differential diagnosis of AD. The best classification model showed 90.9% sensitivity, 76.7% specificity and 82.9% accuracy when it was applied to the total of the SCD + MCI group. The study found that although a few previous studies have shown promising results in differentiating normal, MCI or Alzheimer’s dementia by applying various ML algorithms with EEG features, there is no PET-validated study for an EEG-based ML algorithm predicting the existence of brain Aβ+ plaque among SCI or MCI with no less than 80% accuracy.
Related Posts
