Brain Networks Route Neurodegeneration Patterns in Patients with Progressive Supranuclear Palsy

Abstract

Background: Progressive supranuclear palsy (PSP) is a neurodegenerative disease driven by 4-repeat τ pathology, which is thought to propagate across interconnected neurons.

Objectives: We hypothesized that interconnected brain regions exhibit correlated atrophy, and that atrophy propagates network-like from fast-declining epicenters to connected regions in PSP.

Methods: We combined resting-state functional magnetic resonance imaging (fMRI) connectomics with two independent 12-month longitudinal structural magnetic resonance imaging (MRI) datasets of PSP-Richardson syndrome (PSP-RS) patients (n_discovery/n_validation = 114/90). MRI-based gray matter volumes were assessed for 246 regions of the Brainnetome atlas and converted to w-scores indicating local atrophy (ie, volumes adjusted for age, sex, and intracranial volume based on regression models determined in a sample of 377 healthy amyloid- and τ-negative controls from the Alzheimer's Disease Neuroimaging Initiative [ADNI]). Annual volume changes were determined for each Brainnetome region of interest using longitudinal structural MRI. Resting-state fMRI from 69 ADNI healthy controls was used to determine a connectivity template.

Results: We observed pronounced atrophy and volume decline in the frontal lobe and subcortical regions bilaterally. Correlated atrophy and volume changes were found among interconnected brain regions, with regions with severe atrophy or rapid decline being strongly connected to similarly affected areas, whereas minimally affected regions were connected to less affected areas. Connectivity patterns of atrophy epicenters predicted patient level atrophy and volume decline.

Conclusions: Our findings show that key subcortical and frontal brain regions undergo atrophy in PSP-RS and that gray matter atrophy expands across interconnected brain regions, supporting the view that neurodegeneration patterns may follow the trans-neuronal τ propagation pattern in PSP-RS. © 2025 The Author(s). Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Keywords: PSP; functional connectivity; gray matter atrophy; imaging; tauopathies.

FIG. 1

Group‐average cross‐sectional (A&B) and longitudinal gray matter atrophy (C&D). Mean distribution of cross‐sectionally‐assessed gray matter atrophy was measured with w‐scores (lower scores = stronger volume reduction) for each sample and diagnostic group. [Color figure can be viewed at wileyonlinelibrary.com]

FIG. 2

Group‐average functional connectivity and covariance in gray matter atrophy matrices. Group‐average functional connectivity was computed based on resting‐state functional magnetic resonance imaging (fMRI) of 69 cognitively normal, amyloid‐PET and τ‐PET negative Alzheimer's Disease Neuroimaging Initiative (ADNI) participants (A). The covariance matrices are shown for the discovery cohort (B) and for the validation cohort (C). [Color figure can be viewed at wileyonlinelibrary.com]

FIG. 3

Associations between functional connectivity and atrophy covariance. The association between functional connectivity and covariance in w‐scores is illustrated in the scatter plot for the discovery cohort (A) and for the validation cohort (B). Note that connectivity is displayed as connectivity distance, where shorter distance indicates stronger connectivity. ROI‐based analyses between seed‐based connectivity and w‐scores were performed on subject‐level data, yielding a regression‐derived β‐value for the association between a given seed ROIs connectivity and w‐scores in the discovery (C) and validation cohort (D), showing that connectivity of regions with low w‐scores (ie strong atrophy) predicts brain wide atrophy patterns. The distribution of subject‐level β‐values between gray matter atrophy epicenter connectivity and w‐scores is shown in the box plots for the discovery cohort (E) and the validation cohort (F), respectively, illustrating that regions more strongly connected to a particular patients gray matter atrophy epicenter also show stronger atrophy. CI, confidence interval. [Color figure can be viewed at wileyonlinelibrary.com]

FIG. 4

Associations between functional connectivity and gray matter volume change in atrophy. The association between functional connectivity and rate of change of gray matter atrophy is illustrated in the scatter plot for the discovery cohort (A) and for the validation cohort (B). Note that connectivity is displayed as connectivity distance, where shorter distance indicates stronger connectivity. ROI‐based analyses between seed‐based connectivity and gray matter atrophy rates were performed on subject‐level data, yielding a regression‐derived β‐value for the association between a given seed ROIs connectivity and gray matter volume change in the discovery (C) and validation cohort (D), showing that connectivity of regions with fastest volume decline predicts brain wide volume decline. The distribution of subject‐level β‐values between gray matter volume decline epicenter connectivity and brain wide volume decline is shown in the box plots for the discovery cohort (E) and the validation cohort (F), respectively, illustrating that regions more strongly connected to a particular patients epicenter also show stronger volume decline. CI, confidence interval. [Color figure can be viewed at wileyonlinelibrary.com]