Additive effects of cerebrovascular disease functional connectome phenotype and plasma p-tau181 on longitudinal neurodegeneration and cognitive outcomes

Abstract

Introduction: We investigated the effects of multiple cerebrovascular disease (CeVD) neuroimaging markers on brain functional connectivity (FC), and how such CeVD-related FC changes interact with plasma phosphorylated tau (p-tau)181 (an Alzheimer's disease [AD] marker) to influence downstream neurodegeneration and cognitive changes.

Methods: Multivariate associations among four CeVD markers and whole-brain FC in 529 participants across the dementia spectrum were examined using partial least squares correlation. Interactive effects of CeVD-related FC patterns and p-tau181 on longitudinal gray matter volume (GMV) and cognitive changes were investigated using linear mixed-effects models.

Results: We identified a brain FC phenotype associated with high CeVD burden across all markers. Further, expression of this general CeVD-related FC phenotype and p-tau181 contributed additively, but not synergistically, to baseline and longitudinal GMV and cognitive changes.

Discussion: Our findings suggest that CeVD exerts global effects on the brain connectome and highlight the additive nature of AD and CeVD on neurodegeneration and cognition.

Highlights: Effects of multiple cerebrovascular disease (CeVD) markers on functional connectivity were studied. A global network phenotype linked to high burden across CeVD markers was identified. CeVD phenotype and plasma phosphorylated tau 181 contributed additively to downstream outcomes.

Keywords: Alzheimer's disease; atrophy; cerebrovascular disease; cognition; functional connectivity; longitudinal; phosphorylated tau 181; plasma biomarkers.

FIGURE 1

Study design schematic. Multivariate associations between residuals (after regressing out nuisance variables) of whole‐brain FC and multiple CeVD markers were examined using partial least squares correlation, an approach which seeks to identify a set of latent variables (i.e., linear combinations of the original variables) that maximizes the covariance between two sets of variables. The resultant FC (u) and CeVD (v) weights were then respectively back projected to their original residual values to obtain connectome (Xu) and CeVD (Yv) scores for each participant. Connectome scores, in particular, denote the extent to which each participant expressed the FC pattern that is maximally related to the CeVD markers. Main and interactive effects of connectome score and plasma p‐tau181 (an AD marker) on downstream outcomes, including GMV and cognitive/behavioral performance, were subsequently analyzed using linear mixed effects models. AD, Alzheimer's disease; CeVD, cerebrovascular disease; FC, functional connectivity; GMV, gray matter volume; p‐tau, phosphorylated tau.

FIGURE 2

Partial least squares correlation reveals a functional connectome phenotype that is linked to high burden across multiple CeVD markers and characterized by widespread connectivity changes. We identified one significant latent variable that explained 76.6% of the covariance between CeVD markers and FC. A, Matrix (left) displays the bootstrap ratios (z scores) of functional connections corresponding to this latent variable. Only significant connections (ratio > 2) are displayed, indicating the stable and reliable connections that contribute to the covariance between CeVD markers and FC. Bar chart (right) denotes the mean correlation between connectome scores of this latent variable and each of the CeVD markers, with error bars indicating 95% bootstrapped confidence intervals. The latent variable featured high CeVD burden across all markers, which was correspondingly associated with predominantly lower within‐network but higher between‐network FC in cortical regions, as well as lower subcortical connectivity to associative, transmodal networks but higher subcortical connectivity to sensorimotor, unimodal networks. B, Barplots display mean (± 1 standard error) CeVD and connectome scores for each diagnostic group. * indicate significant pairwise comparisons (P < 0.05). Significant differences in CeVD and connectome scores were observed across different diagnostic groups. C, Scatterplot displays the association between CeVD scores and connectome scores. ARWMC, age‐related white matter changes; CeVD, cerebrovascular disease; CIND, cognitive impairment no dementia; Cont, control network; DEM, dementia; DorsAttn, dorsal attention network; FC, functional connectivity; NCI, no cognitive impairment; SalVentAttn, salience/ventral attention network; SomMot, somatomotor network; TempPar, temporal parietal network; VisCent, central visual network; VisPeri, peripheral visual network.

FIGURE 3

CeVD‐related functional connectome phenotype and plasma p‐tau181 contribute additively to baseline and longitudinal changes in GMVs. A–D, Bar plots indicate T statistic values for the effects of (A) connectome score, (B) connectome score × time, (C) p‐tau181, and (D) p‐tau181 × time on GMVs across 10 networks. * indicates significant baseline effects (FDR‐adjusted p < 0.05) while ^# indicates significant longitudinal effects (FDR‐adjusted P < 0.05). Corresponding brain maps display T statistic values of networks whose GMVs showed significant effects. Lighter blue colors indicate more negative T statistic values. Additive, but not synergistic, effects of connectome score and plasma p‐tau181 on changes in GMVs at baseline and over time were observed. Higher connectome scores were associated with widespread lower GMVs across the brain at baseline, but were not associated with GMV changes over time. In contrast, higher p‐tau181 was associated with lower baseline GMVs in more localized brain regions, including the hippocampus and default network, but steeper longitudinal declines in GMVs over more widespread areas in the brain. E–F, The additive effects of connectome score and p‐tau181 are further illustrated by representative line plots (bands indicate 95% confidence intervals) depicting (E) significant connectome score (FDR‐adjusted P < 0.05; indicated by *) but non‐significant connectome score × time effects on control network GMV, and (F) significant p‐tau181 (FDR‐adjusted P < 0.05; indicated by *) and p‐tau181 × time effects (FDR‐adjusted P < 0.05; indicated by ^#) on hippocampal GMV. CeVD, cerebrovascular disease; FC, functional connectivity; FDR, false discovery rate; GMV, gray matter volume; p‐tau, phosphorylated tau.

FIGURE 4

CeVD‐related functional connectome phenotype and plasma p‐tau181 contribute additively to baseline and longitudinal changes in cognitive and behavioral performance. A–D, Bar plots indicate T statistic values for the effects of (A) connectome score, (B) connectome score × time, (C) p‐tau181, and (D) p‐tau181 × time on cognitive and behavioral scores. * indicates significant baseline effects (FDR‐adjusted P < 0.05) while ^# indicates significant longitudinal effects (FDR‐adjusted P < 0.05). Additive, but not synergistic, effects of connectome score and plasma p‐tau181 on baseline and longitudinal changes in cognitive and behavioral performance were observed. Higher connectome scores were associated with widespread lower baseline performance across most cognitive tests, as well as steeper longitudinal declines in MoCA and attention scores and steeper increases in CDR Sum of Boxes and total NPI scores. By comparison, higher p‐tau181 was associated with lower baseline scores in global cognition (MMSE, MoCA, and global) and memory (visual and verbal memory) tests, but widespread steeper longitudinal declines across all cognitive tests. E—F, The additive effects of connectome score and p‐tau181 are further illustrated by representative line plots (bands indicate 95% confidence intervals) depicting (E) significant connectome score (FDR‐adjusted P < 0.05; indicated by *) but non‐significant connectome score × time effects on visuomotor score, and (F) significant p‐tau181 (FDR‐adjusted P < 0.05; indicated by *) and p‐tau181 × time effects (FDR‐adjusted P < 0.05; indicated by ^#) on verbal memory score. CDR, Clinical Dementia Rating; CeVD, cerebrovascular disease; FC, functional connectivity; FDR, false discovery rate; MMSE, Mini‐Mental State Examination; MoCA, Montreal Cognitive Assessment; NPI, Neuropsychiatric Inventory; p‐tau, phosphorylated tau.