We developed a machine learning framework to quantify disease heterogeneity and progression in multiple system atrophy and Parkinson's disease using structural and diffusion MRI.
We applied the THETA score to investigate cognitive resilience mechanisms in individuals with heterogeneous global tau burden.
We evaluated ComBat and Deming regression for harmonizing white matter hyperintensity measurements across scanners and demonstrated that harmonization challenges persist.
We used diffusion imaging measures to distinguish between hypertensive arteriopathy and cerebral amyloid angiopathy contributions to small vessel disease.
We developed brain age prediction models designed to be invariant to MRI scanner changes for use in aging and dementia studies.
We developed an automatic classification system for head MR series to streamline neuroimaging data management.
We tested whether combining plasma biomarkers, genetics, cardiovascular risk, and lifestyle measures with MRI improves prediction of cognitive decline beyond any single data …
We introduce THETA (Tau Heterogeneity Evaluation in Alzheimer's Disease), a novel tau PET summary metric that captures the full spatial heterogeneity of tau deposition across the …
We applied AI methods to neuroimaging data to detect hypersomnia and characterize its neurobiological correlates.
We developed machine learning models integrating polygenic risk scores for enhanced prediction of Alzheimer's disease imaging and fluid biomarker endophenotypes.
We validated the THETA score using longitudinal PET data and histopathology, demonstrating its association with disease progression and neuropathological staging.
We used brain age models to quantify the impact of chronic medical conditions on dementia risk.
We examined vascular risk factors, gait measures, behavioral symptoms, and plasma biomarkers as indicators of vascular contributions to cognitive impairment and dementia.
We evaluated non-imaging measures for predicting variability in white matter changes relevant to vascular contributions to cognitive impairment and dementia.
We compared NODDI diffusion imaging with structural MRI for sensitivity to neurodegenerative changes.
We demonstrated that genetic risk scores enhance the diagnostic accuracy of plasma biomarkers for detecting brain amyloidosis.
We developed a novel method to encode two-way epistatic interactions between single nucleotide polymorphisms for machine learning applications in genetics.
We characterized white matter degeneration pathways associated with regional tau deposition in Alzheimer's disease.
We conducted a systematic comparison of cross-scanner harmonization methods for structural MRI and demonstrated that widely adopted techniques may not perform as reliably as the …
We developed the first deep learning application for detecting hip osteoarthritis from 2-D summation images derived from routine clinical CT scans.
We developed a filter using alkali-activated blast furnace slag for dye removal from wastewater.
We identified CT-based bone structural characteristics that discriminate patients with low-energy acetabular fractures from matched controls.
We identified structural risk factors for low-energy acetabular fractures in the elderly using computed tomography.