Premium
Amyloid‐dependent association of grey matter network disruptions with phospho‐tau in preclinical Alzheimer’s disease
Author(s) -
Lorenzini Luigi,
Ingala Silvia,
Wottschel Viktor,
Wink Alle Meije,
Mutsaerts Henri JMM,
Haller Sven,
Blennow Kaj,
Schwarz Adam J.,
Gispert Juan Domingo,
Chetelat Gael,
Waldman Adam,
Visser Pieter Jelle,
Tijms Betty M,
Barkhof Frederik
Publication year - 2020
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1002/alz.044739
Subject(s) - dementia , grey matter , betweenness centrality , disease , alzheimer's disease , neuroscience , association (psychology) , psychology , amyloid (mycology) , population , medicine , pathology , centrality , magnetic resonance imaging , mathematics , psychotherapist , white matter , radiology , environmental health , combinatorics
Background Structural Grey Matter Networks (GMN) show progressive disruptions in Alzheimer Disease (AD), starting from early, preclinical stages, and are related to amyloid aggregation. However, GMN associations with tau biomarkers remain unclear. Here, we investigated the relationship between GMN and phospho‐tau (p‐Tau) levels in CSF and whether this relationship was dependent on amyloid beta (Aβ), in a non‐demented population. Method We selected 423 individuals from the first data release of the European Prevention of Alzheimer Disease (EPAD, v500.0) that had MRI and CSF markers available; 84% were cognitively unimpaired (CDR=0) and 16% had CDR=0.5 (Table 1). Abnormal Aβ was defined by levels of <1025 pg/mL, and abnormal p‐Tau by levels of >24 pg/mL in the CSF. Single‐subject networks were extracted from 3D‐T1w structural MRI using a previously described method (Tijms BM et al., 2012), and for each individual we computed whole brain betweenness centrality, degree, density, clustering, modularity, path length and small world properties. Association of network metrics with Aβ and p‐Tau was investigated using linear models including Aβ, p‐Tau+ and Aβ‐p‐Tau interaction, both dichotomously (model 1) and continuously (model 2), adjusted for age, sex and Clinical Dementia Rating (CDR) scores. Results In total 154 (36%) of individuals had Aβ+ (Table 1). Abnormal Aβ (both dichotomously and countinuously) was associated with lower degree, density, clustering and small‐world network properties, independently of tau (p<0.001). Higher p‐Tau levels were related to a reduction in clustering, modularity and small‐worldness values in the Aβ+ group only (p<0.05). Conclusion In cognitively unimpaired subjects, grey matter networks show alterations in amyloid‐positive subjects, independently of tau, and within these individuals stronger GMN disruptions are related to higher p‐Tau levels. Our results indicate a relationship between molecular markers of AD and atrophy patterns at the single‐subject level, specifically going towards a more random and less clustered network topology, and suggest a possible role of GMN measures in identifying early stages of the pathology.