Resting-state EEG biomarkers of general and domain-specific cognitive impairment and plasma P-tau217 in patients with Alzheimer’s disease and mild cognitive impairment

Background Early detection of Alzheimer’s disease (AD)-related cognitive impairment remains challenging due to the lack of convenient, non-invasive screening tools. Objective To explore the multidimensional utility of quantitative electroencephalography (EEG) as a complementary tool for early detection of AD-related cognitive impairment. Methods Clinical data, electroencephalogram (EEG) records, cognitive scale scores and blood samples were collected. According to the established clinical diagnostic criteria, the participants were divided into three groups: cognitive normal group, mild cognitive impairment group (MCI) and Alzheimer’s disease (AD) dementia group. Traditional statistical methods and machine learning algorithms were used to analyze the correlation between 200 EEG indexes and the severity of cognitive impairment, domain-specific cognitive impairment and plasma P-tau217 concentration. Results A total of 360 participants (154 dementia patients with Alzheimer’s disease, 116 mild cognitive impairment patients and 90 healthy controls) were included. The area under the curve (AUC) of EEG in 360 participants is 0.70–0.87, which can effectively distinguish cognitive dysfunction related to Alzheimer’s disease. Among them, 102 participants improved the cognitive domain assessment scale, and the results showed that several EEG indexes were correlated with cognitive domain function score. The blood samples of 105 normal and MCI participants were tested for blood P-tau217 concentration. The relative power of right temporal lobe and right parietal lobe was negatively correlated with blood P-tau217 concentration, and the relative power of right parietal lobe was an independent influencing factor for the increase of blood P-tau217 level (AUC = 0.68). Conclusion Quantitative EEG indicators show potential as a complementary screening approach for AD-related cognitive impairment. Further validation in prospective, multicenter cohorts is needed before implementation in clinical practice.