clinical research science + quantitative methods

Advancing clinical research through rigorous quantitative science.

I am Reuben Addison, a clinical research scientist with advanced training in study design, Bayesian statistics, biomedical data science, and movement science. My work focuses on generating evidence that can stand up to scientific scrutiny and still be useful to clinical, regulatory, and research teams making real decisions.

Explore selected work Start a conversation
Ph.D. + M.S. Clinical science + data science
Bayesian + AI/ML Evidence to prediction
Clinical Research Trials, evidence, outcomes
01 / About

Clinical research strengthened by rigorous quantitative science.

My work centers on clinical and translational research, with quantitative methods used to strengthen study design, evidence generation, interpretation, and decision-making.

I came into quantitative science through clinical and movement-disorders research. That still shapes how I work. I start with the question a clinician, investigator, or study team needs answered, then build the study design, statistical framework, and computational workflow around that question.

Over the years that has meant designing and analyzing Parkinson's disease studies, building EEG and neuroimaging pipelines, modeling longitudinal treatment and behavioral outcomes, and more recently developing Bayesian clinical-trial and biomedical AI systems. I am especially interested in research where uncertainty, provenance, reproducibility, and interpretation directly affect the strength of the clinical evidence.

I currently lead the Motor Control Lab at DePauw University and work with the Chinese American Biopharmaceutical Society's data-science program. I completed Georgia Tech's M.S. in Analytics, Computational Data Analytics Track (Data Science) in 2026, building on a Ph.D. in Kinesiology focused on motor behavior and neuroscience.

02 / What I do

From evidence synthesis to models that can survive scrutiny.

01 // TRIALS

Clinical Trial Design + Evidence Generation

I build Bayesian and group-sequential trial designs, evaluate power and predictive assurance, test prior sensitivity, and connect design assumptions back to the evidence that supports them.

Bayesian design · simulation · assurance · evidence synthesis
02 // INFERENCE

Bayesian + Longitudinal Modeling

I use multilevel, mixed-effects, and treatment-response models to carry uncertainty through the analysis rather than burying it behind a single estimate.

PyMC · R · mixed effects · repeated measures · treatment effects
03 // AI

Biomedical Machine Learning

I develop CNN, Transformer, ensemble, and interpretable ML pipelines for clinical, physiological, and movement data, with validation designed around the scientific use case.

PyTorch · Transformers · CNNs · XGBoost · interpretability
04 // SIGNALS

Movement + Physiological Signals

My research background includes Parkinson's disease, EEG, fMRI, TMS, ECG, motion capture, and markerless pose estimation. I build the preprocessing and quality-control steps that make downstream analysis defensible.

EEG · ECG · fMRI · TMS · biomechanics · computer vision
03 / Selected work

Research systems built to be used, checked, and challenged.

A few projects that capture the kind of problems I like: technically demanding, clinically grounded, and easier to trust when the assumptions are visible.

PROJECT 01 2026

OpenTrial

An evidence-grounded Bayesian trial-design engine that curates ClinicalTrials.gov, PubMed, openFDA, and DailyMed into source-cited design inputs. It produces evidence-derived priors, power and assurance analysis, prior-sensitivity checks, and simulation-based design outputs that a study team can trace back to the underlying evidence.

View repository ↗
Python
Streamlit
Bayesian inference
Clinical data APIs
Simulation
PROJECT 02 ONGOING

KinetiScan

A markerless 3D motion-analysis pipeline that lifts 2D pose into metric 3D coordinates from ordinary camera input. The research focuses on practical validation, numerical parity between research and deployment code, and a path toward accessible movement analysis outside a traditional motion-capture lab.

View GitHub ↗
Python
PyTorch
ONNX
Computer vision
Biomechanics
PROJECT 03 BIOMEDICAL AI

ECG Signal Quality Classification

A clinical signal-quality pipeline on the PTB-XL benchmark comparing a CNN-Transformer ensemble with interpretable statistical and neural baselines. The workflow includes uncertainty-aware modeling, threshold optimization, saliency analysis, and lead-level interpretation.

View repository ↗
PyTorch
Signal processing
Transformers
Interpretability
PTB-XL
04 / Trajectory

A clinical research foundation, expanded through data science.

2026 · present

Data Science Intern

Chinese American Biopharmaceutical Society

Developing OpenTrial and working across clinical evidence synthesis, Bayesian trial design, biomedical data, and regulatory-science questions in a biopharma data-science program.

2023 · present

Assistant Professor of Kinesiology

DePauw University

Lead the Motor Control Lab, design IRB-governed clinical and human-performance studies, analyze treatment and behavioral outcomes, and apply statistical and machine-learning methods to movement and health data.

2022 · 2023

Postdoctoral Research Fellow

MGH Institute of Health Professions · Brain Recovery Lab

Supported NIH-funded neurological research using longitudinal fMRI and TMS data, with responsibilities spanning data curation, multilevel analysis, interpretation, and research mentorship.

2017 · 2022

Doctoral Researcher

Louisiana State University

Studied visuomotor adaptation and Parkinson's disease, designed behavioral and treatment-response studies, and developed EEG signal-processing and statistical workflows.

2015 · 2016

Mitacs Accelerate Intern

Memorial University of Newfoundland

Evaluated a tele-ultrasound system for remote operations in a funded clinical-feasibility project that later produced a peer-reviewed publication.

05 / Research output

Selected publications, manuscripts, and presentations.

Work spanning clinical prediction, Parkinson's disease, motor learning, dystonia, and telemedicine.

Google Scholar ↗
2026
Evidence for preserved gait consistency after wild blueberry ingestion by individuals with Parkinson's disease
Hondzinski, J., et al., with R. N. Addison as co-author. Submitted to Journal of Parkinson's Disease. Manuscript PKN-26-0537.
Submitted editorial processing
2026
A Parsimonious Predictive Model for Hypertension in African American Adults in the Jackson Heart Study
Addison, R. N., et al. American Journal of Hypertension. Under review. Submitted July 10, 2026. Manuscript AJH-D-26-00235.
Under review clinical prediction · health equity
2023
Bilateral transfer of a visuomotor task in different workspace configurations
Addison, R. N., & Van Gemmert, A. W. Journal of Motor Behavior.
Peer reviewed 10.1080/00222895.2023.2293002 ↗
2018
Remote mentoring of point-of-care ultrasound skills to inexperienced operators using multiple telemedicine platforms: Is a cell phone good enough?
Smith, A., Addison, R., et al. Journal of Ultrasound in Medicine, 37(11), 2517-2525.
Peer reviewed 10.1002/jum.14609 ↗
2025
Wild blueberry supplementation and motor symptoms in Parkinson's disease
Gauss, T. M., Addison, R. N., et al. Society for Neuroscience, SFN 2025.
Conference
2024
Neural activity associated with visuomotor adaptation in various workspace locations
Addison, R. N., Steinberg, F., & Van Gemmert, A. W. NASPSPA 2024.
Conference
2024
Subtyping Parkinson's disease by symptom dominance reveals differences in postural sway
Gauss, T. M., Addison, R. N., et al. NASPSPA 2024.
Conference
2023
Functional changes in superior temporal gyrus in focal dystonia
Wang, Y., Hu, D., Addison, R. N., et al. 6th International Dystonia Symposium.
Conference
06 / Toolkit

Methods and tools I reach for when the problem calls for them.

Clinical trial design Bayesian inference Biostatistics Mixed-effects models Longitudinal analysis Treatment-effect estimation Experimental design Real-world evidence Python R SQL MATLAB PyMC PyTorch scikit-learn XGBoost CNNs Transformers Signal processing Computer vision Reproducible pipelines Clinical data APIs Git/GitHub Stakeholder reporting
07 / Contact

Good questions are usually worth a conversation.

I am interested in clinical research and biopharma collaborations where rigorous study design, statistical reasoning, and modern computational methods can strengthen evidence generation and decision-making. My background is especially relevant to clinical development, translational research, neurological and movement-disorder studies, and biomedical data science.

reuben.addison@gmail.com Google Scholar ↗
LinkedIn ↗ GitHub ↗ Google Scholar ↗