Abstract
Background
Brain amyloid (AB) and neurodegeneration (assessed using PET and MRI) are key AD pathology biomarkers, particularly during transitions from cognitively unimpaired (CU) to cognitively impaired (CI) states. Importantly, the role of brain white matter microstructural measures in these changes, such as those derived from neurite orientation dispersion and density imaging (NODDI), are poorly understood. We explored roles of cognitive status, brain amyloid pathology, and neurodegeneration on NODDI Free Water (FW), a potential neuroinflammation biomarker.
Method
Participants enrolled in the Wake Forest Alzheimer’s Disease Research Center (ADRC) Clinical Core cohort (N=74, Table 1) underwent neuropsychological assessment, T1/NODDI MRI, and PiB PET. Participants were adjudicated as cognitively impaired (MCI or AD) or unimpaired using NIA‐AA criteria. T1 MRI were processed (FreeSurfer v5.3) to generate regions of interest (ROIs). An “AD‐signature” cortical thickness value was calculated, and thresholded to classify participants as neurodegeneration‐positive (N+) or ‐negative (N‐). A set of AD‐sensitive cortical ROIs was used to extract mean PET signal, which was thresholded to adjudicate AB positivity (AB‐/AB+). NODDI was processed to yield parameter maps using AMICO; mean isotropic volume fraction (FW) was calculated over a set of white matter tract ROIs. A two‐way ANOVA/ANCOVA (covariates: age, sex, education) assessed effects of neurodegeneration and AB on FW (N+/N‐ x AB+/AB‐), and a two‐way ANOVA/ANCOVA assessed effects of cognitive impairment and AB on FW (CU/CI x AB+/AB‐).
Result
We first examined effects of neurodegeneration and AB on NODDI FW. We found a significant main effect of neurodegeneration (p=.042), such that N+ had higher FW; this was marginal after adjustment (p=.07). We also observed a significant main effect of AB (p<.001; adjusted p=.012), with AB+ having higher FW. There was no significant interaction. We next explored how cognitive status and AB interacted on FW. We found a significant main effect of cognitive impairment, with CI having higher FW (p=.018; adjusted p=.013). We also found an interaction between diagnosis and AB (p=.037; adjusted p=.044) such that among AB+ individuals, presence of CI related to increased FW (Figure 1).
Conclusion
NODDI FW increases were significantly affected by the presence of neurodegeneration, AB positivity, and cognitive impairment.