AbstractBackgroundAlzheimer’s Disease (AD) is defined by the pathological accumulation of amyloid‐beta (Aß) and tau in the brain. Regional deposition of these proteins is associated with changes in frequency‐defined neurophysiological activity that can be detected using non‐invasive magnetoencephalography (MEG). Here we examined whether these neurophysiological alterations were associated with participants’ demographics and longitudinal changes in cognition. Additionally, we addressed the clinical utility of MEG to predict progression to mild cognitive impairment (MCI), as compared to other established AD imaging biomarkers.MethodWe used Positron Emission Tomography (PET) to measure the deposition of whole‐brain Aß ([18F]NAV4694) and tau ([18F] Flortaucipir), Magnetic Resonance Imaging (MRI) to measure hippocampal volume and resting‐state MEG to capture neurophysiological activity in a group of clinically unimpaired older adults with family history of AD (PREVENT‐AD cohort; n = 103). We implemented a multivariate partial least squares (PLS) analysis to test the association between imaging markers (i.e., neurophysiological activity and AD proteinopathy) and participants’ clinical profiles (i.e., demographic‐clinical variables and longitudinal cognition data). For the second analysis we used logistic regression models to assess the added value of neurophysiological activity to predict MCI progression compared to established MRI/PET imaging markers (n = 100; 14 MCI progressors).ResultThe PLS analysis identified significant latent variables that linked proteinopathy‐related neurophysiological activity slowing to older age, lower education, positive APOE e4 status and longitudinal deficits in cognition across multiple domains (Figure 1). The initial logistic regression model included demographic and clinical variables and had an AUC = 0.772, which increased to 0.794 after adding the MRI hippocampal volume. Incorporating MEG spectral power from temporal regions where tau accumulates resulted in a considerable increase in accuracy with an AUC = 0.873. The model including Aß and tau PET on top of MRI and MEG markers reached an AUC = 0.905 (Figure 2).ConclusionOur results show that the neurophysiological changes linked to AD pathology are associated with longitudinal cognitive impairment across multiple domains. MEG spectral power from early tau accumulating regions improved the accuracy for predicting MCI progression, contributing information beyond traditional demographic/clinical variables and structural MRI and representing a more accessible and less invasive alternative compared to PET imaging.