Neural biomarker diagnosis and prediction to mild cognitive impairment and Alzheimer's disease using EEG technology

Abstract

Background: Electroencephalogram (EEG) has emerged as a non-invasive tool to detect the aberrant neuronal activity related to different stages of Alzheimer's disease (AD). However, the effectiveness of EEG in the precise diagnosis and assessment of AD and its preclinical stage, amnestic mild cognitive impairment (MCI), has yet to be fully elucidated. In this study, we aimed to identify key EEG biomarkers that are effective in distinguishing patients at the early stage of AD and monitoring the progression of AD.

Methods: A total of 890 participants, including 189 patients with MCI, 330 patients with AD, 125 patients with other dementias (frontotemporal dementia, dementia with Lewy bodies, and vascular cognitive impairment), and 246 healthy controls (HC) were enrolled. Biomarkers were extracted from resting-state EEG recordings for a three-level classification of HC, MCI, and AD. The optimal EEG biomarkers were then identified based on the classification performance. Random forest regression was used to train a series of models by combining participants' EEG biomarkers, demographic information (i.e., sex, age), CSF biomarkers, and APOE phenotype for assessing the disease progression and individual's cognitive function.

Results: The identified EEG biomarkers achieved over 70% accuracy in the three-level classification of HC, MCI, and AD. Among all six groups, the most prominent effects of AD-linked neurodegeneration on EEG metrics were localized at parieto-occipital regions. In the cross-validation predictive analyses, the optimal EEG features were more effective than the CSF + APOE biomarkers in predicting the age of onset and disease course, whereas the combination of EEG + CSF + APOE measures achieved the best performance for all targets of prediction.

Conclusions: Our study indicates that EEG can be used as a useful screening tool for the diagnosis and disease progression evaluation of MCI and AD.

Keywords: Alzheimer’s disease; Biomarker; Diagnosis; Electroencephalography; Mild cognitive impairment; Prediction.

Fig. 1

The schematic diagram of the classification a–d of HC/MCI/AD participants and assessment e–g of participants’ cognitive function and disease progression

Fig. 2

Distribution of the optimal feature set among all extracted EEG features in classification (indicated by red). Six types of EEG features were extracted, including a absolute PSD, b relative PSD, c Hjorth metrics (activity, mobility, and complexity), d time-frequency measures (STFT), e sample entropy, and f microstate measures (lifetime, occurrence rate, converting rate)

Fig. 3

The key EEG biomarkers at parieto-occipital regions effectively recognized distinct neural patterns among six groups. Selected EEG features included a absolute theta PSD at O2 (F = 42.46, p < 0.001), b relative theta PSD at O2 (F = 50.11, p < 0.001), c Hjorth mobility at O1 (F = 51.08, p < 0.001), d Hjorth mobility at O2 (F = 50.14, p < 0.001), and e Hjorth mobility at P4 (F = 47.09, p < 0.001). “*” indicates that there is a significant between-group difference (p<0.05, FDR corrected)

Fig. 4

The correlational analyses among brain measures (EEG features), cognitive decline (MMSE, MoCA), and CSF biomarkers. Numbers within the Ellipses represent the correlation coefficients of all x–y pairings

Fig. 5

Brain-cognition-CSF relationship in patients with MCI and AD. a Absolute theta PSD at O2 vs. Aβ42, b relative theta PSD at O2 vs. Aβ42, c Absolute theta PSD at O2 vs p-tau, d relative theta PSD at O2 vs. p-tau, e Hjorth mobility at O1 vs. MMSE, f Hjorth mobility at P4 vs. MMSE, g Hjorth mobility at O1 vs. MoCA, and h Hjorth mobility at P4 vs. MoCA

Fig. 6

Results of the prediction analyses using different combinations of features. The predictions of MMSE (a–c), MoCA (d–f), ADO (g–i), and COD (j–l) were obtained using EEG feature only (first column), CSF/APOE biomarkers (second column), hybrid features (EEG, CSF/APOE, sex, and age) as the predictors of the regression model, respectively. R² is the determination coefficient, and MAE is the mean absolute error