Alzheimer’s disease (AD) is a neurodegenerative disease marked by cognitive decline and mild cognitive impairment (MCI), its antecedent. Diagnosis remains a complex process; by the time a patient presents with behavioral evidence of cognitive decline, disease progression is significant. BEAM (Biomarker-based Electrophysiology for Advanced Brain Monitoring), an Artificial Intelligence-based platform, interprets electroencephalogram (EEG) data and establishes biomarkers that can potentially allow for early detection of cognitive decline.

This project aims to determine if BEAM EEG biomarkers can be used as predictors of AD and which are the strongest. A retrospective chart review of 143 patients from Hawaii Pacific Neuroscience was performed and included at least one BEAM EEG report collected between June 2024 and June 2025. A generalized linear model (GLM) was fit on the biomarkers: peak alpha, eyes-open individualized theta-to-alpha ratio (ITAR), eyes-closed ITAR, auditory oddball (AO) N1 peak latency, AO P300 peak max latency, and max amplitude from 98 patients with AD or MCI.

Age, sex, body mass index, hypertension, hyperlipidemia, diabetes, and depression were controlled for. Odds ratios and 95% confidence intervals were analyzed to determine the impact each biomarker had on the likelihood of an AD versus MCI diagnosis. Of the six BEAM biomarkers analyzed, peak alpha was a significant predictor of AD. The results are consistent with peak alpha’s clinical association with cognitive decline at lower values, indicative of a slowing alpha frequency. However, the study is limited by sample size. Further studies could incorporate multiple BEAM EEGs per patient to address this.