Exploring the neuromagnetic signatures of cognitive decline from mild cognitive impairment to Alzheimer’s disease dementia

Abstract

AbstractIntroductionAlzheimer’s disease (AD) is the most common cause of dementia. Non-invasive, affordable, and largely available biomarkers that are able to identify patients at a prodromal stage of AD are becoming essential, especially in the context of new disease-modifying therapies. Mild cognitive impairment (MCI) is a critical stage preceding dementia, but not all MCI patients will progress to AD. This study explores the potential of non-invasive magnetoencephalography (MEG) to predict future cognitive decline from MCI to AD dementia.MethodsWe analyzed resting state MEG data from the BioFIND dataset including 117 MCI patients, of whom 64 progressed to AD dementia (AD progression) while 53 remained stable (stable MCI) using multivariate spectral analyses. The patients were followed-up between 2009 and 2018. Receiver operating characteristic curves obtained via logistic regression models were used to quantify separation of patients progressing to AD dementia from stable MCI.ResultsMEG beta power, particularly over parieto-occipital magnetometers, was significantly reduced in the AD progression group compared to stable MCI, indicative of future cognitive decline. Logistic regression models showed that MEG beta power outperformed conventional metrics like the Mini Mental Status Examination (MMSE) score and structural brain measures in predicting progression to AD dementia (AUC 0.81 vs 0.71 and AUC 0.81 vs 0.75, respectively). The combination of age, education, MMSE, MEG beta power and Hippocampal volume/Total grey matter ratio achieved a 0.83 AUC, 78% sensitivity and 76% specificity. Spectral covariance matrices analyzed with Riemannian methods exhibited significant differences between groups across a wider range of frequencies than spectral power.DiscussionThese findings highlight the potential of spectral power and covariance as robust non-invasive electrophysiological biomarkers to predict MCI progression that complement other diagnostic measures, including cognitive scores, structural magnetic resonance imaging (MRI) and biological biomarkers.