Increased ¹⁸F‐flortaucipir load correlates with changes in MEG functional connectivity and network topology, as well as oscillatory slowing

This is the authors' abstract. We don't add key points for this paper.

Background Recent fMRI studies have provided evidence that tau proteins may spread through the brain in stereotypical patterns following functional connections. However, fMRI is an indirect measure of neuronal activity, and has limited temporal resolution, contrary to magnetoencephalography(MEG). We aimed to investigate whether changes in neuronal activity, functional connectivity and network organization mirror regional tau retention in (preclinical) Alzheimer’s disease(AD), using MEG and in‐vivo 18 F‐flortaucipir (tau) PET. Methods Twenty participants with AD pathology according to the ATN‐criteria (CSF‐Aβ 42 and/or amyloid‐PET) were recruited from the Amsterdam Dementia Cohort, classed as subjective cognitive decline (SCD; n=10) and probable AD dementia (n=10), Table‐1. Participants underwent dynamic 130‐minutes 18 F‐flortaucipir PET, and 10‐minutes eyes‐closed, task‐free 306‐channel MEG recordings. Neuronal activity and 18 F‐flortaucipir non‐displaceable binding potential(BP ND ) were reconstructed for 52 regions‐of‐interest(ROIs) in an adjusted version of the probabilistic Hammers atlas. Associations between 18 F‐flortaucipir BP ND and MEG spectral measures (peak frequency, relative delta(0.5‐4Hz); theta(4‐8Hz); alpha(8‐13Hz); beta(13‐30Hz); gamma(30‐48Hz) power); connectivity (volume conduction‐corrected amplitude envelope correlation(AEC‐c) in delta, alpha and beta band); and network measures (minimum spanning tree betweenness centrality(BC)) were investigated using generalized estimating equation(GEE) models, for SCD and AD dementia combined. To investigate possible group effect, the GEE was repeated using only the AD dementia‐group. Result Increased regional 18 F‐flortaucipir binding correlated strongly with oscillatory slowing in both the total sample and AD‐group, reflected by negative correlations with peak frequency, relative alpha and beta power, and positive correlations with relative delta and theta power, most pronounced in parietal and occipital regions (Fig‐1A+B). Regional 18 F‐flortaucipir binding correlated with loss of connectivity for both groups mainly in the occipital regions (average β=‐0.677±0.16, β=‐0.660±0.23 resp.) for the beta band, Fig‐2A+B. The delta band showed an opposite effect in frontal and occipital regions. Higher 18 F‐flortaucipir binding was associated with lower centrality in temporal (delta) and subcortical regions (alpha, beta), and higher centrality in the parietal region (alpha), Fig‐2A+B. Conclusion We provide putative evidence that increased regional 18 F‐flortaucipir binding associates to reduced neuronal activity, as well as loss of connectivity in higher frequency bands, which could be replicated in a separate AD analysis. Furthermore, tau deposition seems to be related to damage to highly connected hub nodes.