Association of global gray matter volume and cortical thickness with cognitive function stratified by amyloid PET...

Structural magnetic resonance imaging (MRI) measures may relate to cognition differently depending on the underlying biological context. We examined the associations of global gray matter volume and cortical thickness with cognitive performance after stratification by amyloid positron emission tomography (PET) status.

This cross-sectional study included 113 participants (mean age 76.3 ± 7.8 years; 68.1% female) who underwent amyloid PET, structural MRI, and Mini-Mental State Examination (MMSE). Amyloid burden was quantified on the Centiloid scale, with positivity defined as > 25 (44 amyloid-negative, 69 amyloid-positive). Global gray matter volume normalized by total intracranial volume (GM/TIV) and global cortical thickness were derived with CAT12/SPM12. Multivariable linear regression was used to assess associations with the MMSE score after adjustment for age, sex, and amyloid burden, with prespecified stratification by amyloid status. Formal interaction analyses tested effect modification.

In the overall cohort, GM/TIV was independently associated with the MMSE score (β = 0.35, p = 0.007), whereas cortical thickness was not (β = 0.02, p = 0.844). In exploratory stratified analyses, GM/TIV was associated with the MMSE score in amyloid-positive participants (β = 0.48, p = 0.003) but not in amyloid-negative participants (β = 0.05, p = 0.835). Age was associated with the MMSE score in amyloid-negative participants (β = -0.33, p = 0.041) but not in amyloid-positive participants (β = 0.11, p = 0.377). Formal interaction analyses were not significant.

In this cohort, global GM/TIV was associated with cognitive performance more consistently than global cortical thickness. Exploratory stratified analyses suggested different association patterns according to amyloid PET status; however, formal interaction analyses were not significant. These findings should therefore be interpreted cautiously and validated in larger longitudinal studies.