Premium
Cholinergic network disruption in AD subtypes: A study using graph theory
Author(s) -
GonzalezBurgos Lissett,
Grothe Michel,
Teipel Stefan J.,
Westman Eric,
Ferreira Daniel
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.043178
Subject(s) - basal forebrain , neuroscience , precuneus , cholinergic , temporal lobe , cholinergic neuron , thalamus , atrophy , basal ganglia , hippocampal formation , psychology , biology , pathology , medicine , functional magnetic resonance imaging , central nervous system , epilepsy
Background Cholinergic dysfunction is central in Alzheimer’s disease (AD). Recent advances in neuroimaging techniques and availability of larger postmortem datasets have opened new opportunities to understand the role of the cholinergic network in AD pathogenesis. However, there is pathophysiological heterogeneity within AD. AD subtypes with different distribution of neurofibrilary tangles, regional atrophy, and network disruption have been revealed in neuropathological and structural magnetic resonance imaging (sMRI) studies. We aimed to identify the cholinergic network using graph theory on sMRI data, and investigate whether differential cholinergic network disruption underlies the subtypes of AD. Method We applied graph theoretical analysis to structural MRI data from 75 Ab‐positive AD patients and 64 Ab‐negative controls from ADNI‐1. AD patients were classified into three subtypes according to Murray et al., (2011): typical AD, limbic‐predominant and hippocampal‐sparing. We built brain structural networks from 68 cortical regions and 14 subcortical gray matter structures. Network disruption was investigated through modular analyses by applying the Newman algorithm on weighted graphs. Result We identified 43 AD patients with a brain atrophy pattern consistent with typical AD; 17 patients with a limbic‐predominant pattern; and 15 patients with a hippocampal‐sparing pattern. We identified the cholinergic network in healthy controls, which included the basal forebrain, medial temporal lobe, precuneus, lateral parietal and occipital cortex, basal ganglia and thalamus. This network was completely fragmented in AD patients, but disruption differed according to subtype. The basal forebrain retained correlations only with basal ganglia and medial temporal lobe regions in typical AD; while in hippocampal‐sparing and limbic‐predominant subtypes these correlations also extended to new regions in frontal and lateral temporal cortices. Conclusion Graph theory on sMRI is able to identify a cholinergic network similar to that previously described in postmortem and in vivo diffusion imaging studies. AD subtypes have a distinct signature of cholinergic network disruption largely associated to their atrophy patterns. It is thus possible that pathology targeting the cholinergic system is one of the factors contributing to the emergence of these distinct subtypes. Our results may help to further understand pathophysiological heterogeneity in AD, with potential applications to research, clinical work and drug decision.