Data-driven functional decline trajectories correspond to distinct brain perfusion imaging patterns

Sofia Michopoulou¹,², Dave Evenden², Angus Prosser², John Dickson³, JessicaTeeling², Christopher Kipps¹,²
¹University Hospital Southampton NHS Foundation Trust, ²University of Southampton, ³University College London Hospitals

 Study Background and Aim

• Cognitive decline is heterogeneous, limiting prognosis and care planning.
• Clinical tools such as CDR-SB track functional decline but provide limited biological insight.
• Perfusion SPECT provides a physiological marker of regional synaptic dysfunction, often before structural change.

Aim: To assess whether cerebral perfusion distinguishes individuals with different functional decline trajectories.

Methods

• Design: Brain Imaging in Dementia (BRAIID) longitudinal cohort (n = 232).
• Clustering: Longitudinal CDR-SB change grouped using curveRep() into four decline trajectories. See poster #2615 for on hierarchical clustering.
• Imaging: Baseline brain perfusion SPECT analysed using SPM to provide voxel-wise comparisons across different decline trajectories

Results:

• Data-driven clustering identified four distinct CDR-SB trajectories. Fast progressors showed the steepest decline: 4.02 units/year. Slow progressors declined 1 unit/year.
• Perfusion maps reveal that faster decline is associated with more extensive hypoperfusion in posterior cortical areas.

Conclusion and Clinical Impact

• Functional decline trajectories mapped onto distinct brain perfusion profiles in a real-world clinical cohort.
• Faster progression was associated with greater posterior hypoperfusion, particularly within the default mode network.
• Findings support network-based models of neurodegeneration, where dysfunction spreads across interconnected posterior hubs.
• Progression speed appears to reflect underlying biological heterogeneity, not only baseline clinical severity.
• Perfusion SPECT may support early risk stratification by identifying individuals at higher risk of rapid decline, improving patient stratification for therapy.

Contact sofia.michopoulou@uhs.nhs.uk

University of Southampton | University Hospital Southampton | www.ideac.org