Abstract
BACKGROUND: Neurovascular biomarkers have the potential to enhance early diagnosis of Alzheimer's disease (AD) and AD-related dementias (ADRD), as cerebrovascular alterations often precede neurodegeneration. However, their clinical application remains challenging due to insufficient specificity, heterogeneity, and technical limitations. METHODS: Here, we report that vessel- and cortical layer-specific parameters exhibit promising diagnostic sensitivity for neurovascular impairments in an AD/ADRD mouse model, apolipoprotein E (APOE) 4 knock-in (KI), compared to APOE3-KI at 12 months of age. Using two in vivo imaging modalities, 3D capillary-resolution optical Doppler tomography and laser speckle contrast imaging, we measured 36 morphological and functional vascular parameters and evaluated their diagnostic performance using a machine-learning Support Vector Machine classifier. RESULTS: APOE4 mice showed significant alterations including reduced venular and arterial cerebral blood flow velocities and diameters, increased vascular tortuosity, layer-dependent decreases in vascular density, and impaired cerebrovascular reactivity. Venule- and microcirculation-related parameters and dynamic vasoactivity to brain stimuli demonstrated high diagnostic accuracy (~ 90%). CONCLUSION: Together, these findings provide in vivo evidence for early, cortical layer-specific neurovascular dysfunction caused by APOE4 that increases the susceptibility to dementia and highlight the potential of combining neurovascular biomarkers from optical imaging with AI-based classifier for identification of increased AD/ADRD risk.