Atrophy in bvFTD spans multiple large‐scale networks in prefrontal and temporal cortex

Abstract

AbstractBackgroundConverging evidence suggests that neurodegenerative disorders preferentially affect specific large‐scale neural networks 1. This is thought to explain common phenotypic presentations of these conditions. For example, atrophy patterns in Alzheimer’s disease target episodic memory circuits2. Examination of network‐specific neurodegeneration could lend insight into how these conditions spread within vulnerable cortical systems 3. Behavioral variant Frontotemporal Dementia (bvFTD) is most commonly linked to degeneration in the salience network, with clinical presentations of bvFTD explained as deficits in salience processing 4,5. However, phenotypic presentations of bvFTD are diverse6 and this syndrome is caused by varied proteinopathies and gene mutations7. As such, we hypothesized that bvFTD patients would exhibit atrophy patterns that cross canonical network boundaries within prefrontal and temporal cortex.MethodWe estimated cortical thickness using surface‐based analyses described in our previous studies2 in two independent sporadic bvFTD samples: a Massachusetts General Hospital sample (n=30) and a sample derived from the Frontotemporal Lobar Degeneration Neuroimaging Initiative (NIFD) (n=60; excluding cases contributed by MGH). Cortical thickness in these samples were then compared to that measured in two groups of age and gender matched controls (n=71 and n=133, respectively). In order to estimate individual network atrophy in bvFTD, we then compared the resultant cortical thickness maps to the topography of seven large‐scale cortical networks derived from the Yeo et al. 2011 parcellation8.ResultWe found highly similar patterns of cortical atrophy across the two independent samples. In both samples, cortical atrophy in bvFTD was not confined to the salience network, but also spanned rostral portions of the default‐mode network (including the hippocampal formation), the limbic (semantic appraisal) network, and rostral nodes of the frontoparietal control network. Moreover, regions exhibiting the most prominent atrophy included anterior temporal regions, not only the salience network.ConclusionThese findings argue against the hypothesis that bvFTD preferentially affects the salience network and instead underscores the distributed impact this syndrome has across prefrontal and anterior temporal cortical circuits. This heterogeneity could reflect the diverse neuropathological etiologies giving rise to the cognitive‐behavioral syndrome of bvFTD. Future efforts should be directed toward examining selectively‐vulnerable circuits in bvFTD syndromes associated with specific molecular pathologies.