Estimating the preclinical Alzheimer's disease course with multimodal data

Area of Science:

• Neuroscience
• Biostatistics
• Gerontology

Background:

• Establishing an individual's position along the Alzheimer's disease (AD) continuum in preclinical stages is challenging due to arbitrary baselines.
• Longitudinal data optimization can create a single metric to represent an individual's disease progression trajectory.

Purpose of the Study:

• To develop and validate a "cognitive time" (c-time) metric using longitudinal cognitive data.
• To assess the relationship between c-time and established markers of AD pathology, such as amyloid-beta (Aβ) deposition and neurodegeneration.

Main Methods:

• Developed a c-time metric via non-linear least-squares optimization of cognitive data from 316 participants.
• Analyzed associations between c-time and a time-to-amyloid-beta-positive (Aβ+) metric derived from longitudinal amyloid PET scans.
• Included demographic, brain volume, cortical thickness, tau PET, and cardiovascular risk factors in path analyses.

Main Results:

• C-time and time-to-Aβ+ were moderately positively correlated.
• Time-to-Aβ+ showed direct and indirect associations with c-time, mediated by regional tau PET, hippocampal volume, and cortical thickness.
• C-time demonstrated stronger associations with tau pathology, brain atrophy, and cortical thickness compared to time-to-Aβ+.

Conclusions:

• Optimizing longitudinal multimodal data can yield age-independent metrics indicating an individual's proximity to disease-related events.
• C-time offers a novel way to quantify an individual's position relative to cognitive decline onset.
• C-time provides valuable insights into the preclinical AD continuum, integrating cognitive, pathological, and structural brain changes.