Modeling regional vulnerability to Alzheimer pathology

Area of Science:

• Neurology
• Biostatistics
• Gerontology

Background:

• Alzheimer's disease (AD) pathology progresses spatially, but quantifying interindividual variation in this progression remains challenging.
• Latent growth curve (LGC) models are typically used for temporal changes, not spatial distributions of neuropathology.

Purpose of the Study:

• To apply LGC models to the spatial distribution of Alzheimer's disease (AD) pathology.
• To quantify interindividual variation in regional vulnerability to AD lesions.
• To investigate the relationship between spatial pathology gradients and clinical outcomes.

Main Methods:

• Utilized autopsy data from 435 participants in the Honolulu-Asia Aging Study.
• Applied LGC techniques to analyze the spatial distribution of neurofibrillary tangles (NFT) and neuritic plaques (NP).
• Examined associations between spatial change gradients (dNFT, dNP), Braak stage, cognition, and apolipoprotein E (APOE) e4 status.

Main Results:

• The spatial change gradient of neuritic plaques (dNP) was weakly and inversely associated with that of neurofibrillary tangles (dNFT).
• Both dNFT and dNP were significantly and inversely correlated with Braak stage.
• dNFT explained 61% of the variance in Braak stage, was linked to cognitive decline, while dNP was associated with APOE e4 burden.

Conclusions:

• This study introduces the novel application of LGC models to spatially-ordered neuropathological data.
• Quantified interindividual differences in regional vulnerability to AD pathology.
• Demonstrated distinct associations of NFT and NP spatial gradients with Braak stage, cognitive change, and APOE e4 status.