Welcome to the Neurological Clinical Informatics and Data Science Lab, where we explore the intricate world of neurological disorders with a focus on Alzheimer’s disease, dementia, and migraine disorders. Our multidisciplinary team harnesses the power of machine learning, artificial intelligence, and advanced statistical modeling to unravel the complexities of these conditions, aiming to improve diagnosis, prognosis, and treatment outcomes. By integrating clinical informatics and cutting-edge data science techniques, we strive to unlock new insights into the underlying mechanisms of neurological diseases and pave the way for personalized and precision medicine approaches.

Three side-by-side ROC plots labeled ‘CN vs MCI’, ‘CN vs Dementia’, and ‘MCI vs Dementia’. In each plot there are four curves: dotted blue for DA, solid blue for DAC, dotted red for DAI, and solid red for DACI. The legend lists the AUC for each (e.g., DA = 0.70, DAC = 0.74, DAI = 0.81, DACI = 0.84)

ROC curves for logistic-regression models classifying baseline clinical diagnosis, showing how adding inflammatory biomarkers (I) improves discrimination. Models build sequentially on demographics (D: age, sex, education), APOE4 status (A), and classic AD CSF biomarkers (C: Aβ, p-tau181). Blue lines = without inflammation; red lines = with inflammation. AUC values for each model are noted in the legend.
A 2×2 figure. Top-left: ROC plot for amyloid-β prediction with four colored curves (Model-1 in blue, Model-2 in red, Model-3 in green, Model-4 in purple) against the diagonal reference line. Top-right: ROC plot for tau prediction with five colored curves (Model-1 through Model-5). Bottom-left: bar chart showing percentage amyloid-β positivity on the y-axis for three risk-score tertiles on the x-axis (0–33, 34–66, 67–100), with four bars per tertile matching the four models and a dotted line at overall prevalence. Bottom-right: similar bar chart for tau positivity with five bars per tertile for each model and a dotted prevalence line
Top panels show ROC curves for four amyloid-β prediction models (left) and five tau prediction models (right) on held-out test sets. Bottom panels display the observed percentage of amyloid-β positivity (left) and tau positivity (right) within each risk-score tertile (0–33, 34–66, 67–100), with the overall biomarker prevalence indicated by the dotted horizontal line.