Research Article

Multimodal EEG Biomarkers of Cognitive Decline: BrainView-Based Detection of Early Alzheimer’s Disease

Authors: Annie TL Young, Slav Danev, Pedro Gutierrez Castrellon, Jonathan RT Lakey.

DOI: 10.33425/2641-4333.1084

---

Abstract

Background: Early detection of Alzheimer’s disease (AD) is critical for timely intervention, yet current diagnostic tools can be costly and invasive. Quantitative electroencephalography (qEEG) and event-related potentials (ERPs) offer non- invasive, cost-effective alternatives with high temporal resolution.

Objective: This study evaluated the BrainView system, a multi-modal EEG platform integrating qEEG, source localization, ERP analysis, and machine learning, to assess its utility in detecting early cognitive decline.

Methods: Z-scored brain maps, sLORETA source localization, and ERP waveforms were analyzed in individuals with suspected early AD. Key frequency ratios and cortical activity patterns were compared against normative databases. Machine learning models, including XGBoost with LASSO regression, were trained to classify cognitive status.

Results: Elevated delta and theta activity, alongside reduced alpha and beta power, reflected typical EEG slowing associated with early AD. sLORETA localized these abnormalities to frontal and parietal regions. ERP analysis revealed attenuated P200 responses and atypically early, amplified P300 components, suggesting possible compensatory mechanisms. The XGBoost machine learning model achieved high classification accuracy (sensitivity = 0.88, specificity = 0.94), supporting the diagnostic value of EEG-based features.

Conclusion: The BrainView system effectively identifies neurophysiological patterns consistent with early AD. Combined with machine learning, EEG-based tools offer a scalable, non-invasive approach for early cognitive screening and monitoring. Further validation in larger, diverse cohorts is needed to support clinical integration.

View / Download PDF