Introduction
Alzheimer’s disease (AD) is the leading cause of dementia, accounting for an estimated 60-80% of cases worldwide. Presently, there is no effective treatment for AD and only very limited medicines show the potential for delaying the progression of this neurodegenerative disease at its early stage. On the other hand, amnestic mild cognitive impairment (MCI) is characterized by cognitive decline greater than normal for a person’s age and education level without notably interfering with activities of daily life. It is well-accepted that MCI is a high-risk factor for the development of AD and reflects a prodromal dementia state of AD.
The gold standard of AD diagnosis is the Amyloid/Tau/Neurodegeneration (ATN) framework proposed by the National Institute of Aging and Alzheimer’s Association. In the ATN framework, the biological state of AD is classified through the identification of three biomarkers (i.e., amyloid, tau, and neurodegeneration) measured from cerebrospinal fluid (CSF) and positron emission tomography (PET) imaging.
Electroencephalography (EEG), a low-cost, non-invasive, and portable technique that directly measures neural activity with a high temporal resolution, has appeared as a potential tool for detecting neural biomarkers related to MCI and AD. Multiple lines of evidence have validated the possibility of using EEG to differentiate MCI and AD patients from healthy cohorts with diverse sensitivity and specificity.Moreover, while AD is the most common form of dementia, symptoms of preclinical and early AD mostly overlap with other types of dementia such as frontotemporal dementia (FTD), dementia with Lewy bodies (DLB), and vascular cognitive impairment (VCI).
Materials and methods
A total of 890 individuals were utilized in the study. The patients of MCI, probable AD, FTD, VCI, and DLB were diagnosed according to the respective clinical-based criteria. The CSF was collected through a lumbar puncture. All assessments of CSF biomarkers were estimated using an enzyme-linked immunosorbent assay (ELISA).
The EEG signal was recorded at 200 Hz from the participants in a 10-min-eye-closed resting state. The pipeline of the proposed classification and assessment framework is shown in Figure 1.Multiple types of EEG features were computed, including absolute power, relative power, Hjorth metrics (activity, mobility, and complexity), time-frequency property, sample entropy, and microstate measures.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 participant’s EEG biomarkers, demographic information (i.e., sex, age), CSF biomarkers, and APOE phenotype for assessing the disease progression and individual’s cognitive function.
Figure 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. Adapted from source
Results
The identified EEG biomarkers achieved over 70% accuracy in the three-level classification of HC (healthy controls), MCI, and AD. 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.
Conclusion
Increasing evidence indicates that EEG biomarkers are diagnostically significant and associated with the clinical progression of AD. In this study, we specified distinct neural biomarkers that were specifically linked to the CSF measures and cognitive function of AD patients. These neural biomarkers mainly included the power spectrum alterations of low-frequency oscillations at the occipital area and the altered signal complexity at the parietal and occipital regions. Finally, through a machine learning technique, we found that the combination of EEG biomarkers, CSF/APOE ε4 measures, and demographic information of patients was most effective in evaluating individual cognitive function and disease progression.