Graff-Radford, J. | bioRxiv Science

bioRxiv2026

Diffusion MRI based biomarkers reveal a Prolonged Pre Lesional Phase of Cerebral Small Vessel Disease

BackgroundWhite matter hyperintensities (WMH) are widely used to assess cerebral small vessel disease (SVD) but reflect late-stage injury. Diffusion MRI based biomarkers have been proposed to capture earlier SVD-related microstructural damage but their temporal progression relative to WMH and the risk factors associated with this progression have not been explored. MethodsWe analyzed longitudinal neuroimaging data from 2,047 participants from a population-based cohort study (aged 49-101 years, 47% female). Using multi-output nonlinear mixed-effects models, we characterized the temporal progression of WMH and four diffusion MRI based biomarkers: fractional anisotropy of the genu of the corpus callosum (Genu-FA), peak width of skeletonized mean diffusivity (PSMD), free water (FW), and Arteriolosclerosis-score (ARTS). Models incorporated participant-specific time shifts, correlations between biomarkers, and effects of risk factors (sex, education, APOE {varepsilon}4 status, and cardiometabolic conditions). ResultsARTS, Genu-FA, FW, and PSMD became abnormal in 50% of the study population 16, 12, 10, and 7 years before WMH, respectively. Global markers (ARTS, FW, PSMD, WMH) were correlated, indicating shared substrates of widespread white matter injury. Genu-FA, a vascular risk microstructural injury biomarker, was weakly coupled with WMH and had an earlier but more linear worsening across adulthood. Cardiometabolic conditions predicted earlier worsening of all biomarkers. Females showed earlier WMH, Genu-FA, and ARTS abnormalities whereas males exhibited earlier PSMD and FW abnormalities. ConclusionsDiffusion MRI based biomarkers capture microstructural injury at least a decade before appearance of WMH, revealing a prolonged phase of early SVD and highlighting their potential for SVD prevention.

neuroscience↗

bioRxiv2022

Synthesizing Images of Tau Pathology from Cross-modal Neuroimaging using Deep Learning

Given the prevalence of dementia and the development of pathology-specific disease modifying therapies, high-value biomarker strategies to inform medical decision making are critical. In-vivo tau positron emission tomography (PET) is an ideal target as a biomarker for Alzheimers disease diagnosis and treatment outcome measure. However, tau PET is not currently widely accessible to patients compared to other neuroimaging methods. In this study, we present a convolutional neural network (CNN) model that impute tau PET images from more widely-available cross-modality imaging inputs. Participants (n=1,192) with brain MRI, fluorodeoxyglucose (FDG) PET, amyloid PET, and tau PET were included. We found that a CNN model can impute tau PET images with high accuracy, the highest being for the FDG-based model followed by amyloid PET and MRI. In testing implications of AI-imputed tau PET, only the FDG-based model showed a significant improvement of performance in classifying tau positivity and diagnostic groups compared to the original input data, suggesting that application of the model could enhance the utility of the metabolic images. The interpretability experiment revealed that the FDG- and MRI-based models utilized the non-local input from physically remote ROIs to estimate the tau PET, but this was not the case for the PiB-based model. This implies that the model can learn the distinct biological relationship between FDG PET, MRI, and tau PET from the relationship between amyloid PET and tau PET. Our study suggests that extending neuroimagings use with artificial intelligence to predict protein specific pathologies has great potential to inform emerging care models.

bioengineering↗