The role of PPG in identification of mild cognitive impairment

Abstract

Early and reliable detection of cognitive impairment is crucial for optimized care of Alzheimer's disease. In our former publication, we derived features from gait signals and proposed a novel feature selection algorithm to identify mild cognitive impairment (MCI) aging. In this paper, we concentrate on applying the previously proposed algorithm on a different biosignal, photoplethysmography (PPG), to improve MCI classification. We also demonstrate data acquisition using a finger-tip wireless pulse oximeter and feature extraction from PPG. Our classification accuracy is 0.90 ± 0.01 with the dataset from 62 elderly participants (72.71 ± 10.63 years; 31 MCI and 31 control), which is a higher classification accuracy than only using the administered neuropsychological measures. This study verifies that PPG-derived parameters also have the potential to enhance the ability to accurately diagnosis cognitive impairment.