Evidence for functional connectivity pattern changes in Frontotemporal dementia patients using connectome gradient mapping

Abstract

AbstractBackgroundFrontotemporal dementia (FTD) is a neurodegenerative syndrome with a broad range of clinical manifestations making its early detection and progression monitoring difficult. As we enter the first clinical trials, useful biomarkers are needed. It is believed that functional connectivity underlies cognitive function which makes it a potential meaningful biomarker of disease progression. Studies of functional connectivity in FTD have provided a range of findings suggesting changes in resting‐state networks. However, many of the analysis methods used are either far from the biological processes or restricted to a priori defined networks. Recent work has revealed the emergence of a brain network hierarchy during neurodevelopment, separating immediate environment processing from transmodal and integrative cortices. It is assumed that this organization supports higher order cognition.MethodWe included 36 behavioural FTD patients (bvFTD) and 50 controls. We applied a novel whole‐brain analysis approach, connectome gradient mapping, to resting‐state fMRI data. This approach extracts gradients representing measures of functional similarity. Results are presented as brain parcellations within connectivity space, where the distance between parcels describes the extent of segregation between networks (Fig A).ResultsbvFTD patients showed a disruption in macroscale hierarchy affecting integration and segregation of unimodal and transmodal networks. Specifically, in bvFTD patients, we observed a significant shift of cognitive/behavioural networks, such as the Limbic and Default‐mode, which are less segregated from the primary function networks, such as visual and auditory (Fig B).ConclusionUsing this novel method, we found that the Default‐mode and Limbic networks seem particularly vulnerable in FTD whereby they lose their evolutionarily derived characteristics of segregation. These networks underlie the complex cognitive and behavioural phenotypes observed in FTD. In the near future, we will evaluate whether this model could explain the main symptoms of patients, such as cognitive disinhibition (Stroop/Hayling tests), alterations in social cognition (miniSEA) and worse general cognitive function (MMSE). Overall, this method applied to early and/or presymptomatic stages will be useful for early disease identification and predicting treatment outcomes, making it of interest in future therapeutic studies for neurodegenerative diseases.