From raw data to clinical insight.
My research follows a single chain — from signal, image, and video data, through reproducible analysis, to results a clinician could act on. I work across clinical, animal, and computational studies, and my edge is that I also build the instruments that capture the data in the first place.
Signal, Image & Video Analysis
Turning raw data into meaning is the core of my work. I analyze tactile / contact-sensor time series, markerless kinematics from high-frame-rate video (DeepLabCut), and medical images — across clinical, animal, and computational studies, with explicit quality control so unreliable data never reaches the results.
Representative work: DeepLabCut kinematics from 60 fps video across an 84-trial dataset with documented retention (95.2% sensor, 89.3% kinematic); Digital Image Correlation analysis of ventricular deformation in my Master's.
See the projectsReproducible Pipelines & Statistics
The point of careful analysis is a defensible result. I write Python pipelines and use linear mixed-effects modeling and effect-size estimation, with the individual subject as the unit of analysis — built so another researcher can rerun the pipeline and reach the same conclusion.
Representative work: the LME trajectory models and Mann–Whitney effect-size analysis behind my PhD thesis, packaged as a reproducible pipeline.
See the projectsInstrumentation & Sensing
The edge most data analysts don't have: I also build the acquisition systems behind the data. I design behavioral apparatus and hardware around Arduino and microcontroller systems, with timestamped tactile / contact sensing — which means I understand and control the signal from the sensor up, not just after it lands in a file.
Representative work: the custom apparatus and touch-sensor acquisition for my PhD paradigm; an Arduino + stepper CPM wrist rehabilitation device with a Processing 3 serious game.
See the projectsToward Clinical Digital Biomarkers
These computational methods don't care about species — time-series analysis, markerless kinematics, imaging, and reproducible pipelines are exactly the tools of modern clinical and digital-health research. My interest going forward is carrying this analysis stack toward clinical digital biomarkers, and releasing the tools openly through the Neuroprocessing initiative I'm developing.
In development: Neuroprocessing — an open-source neuroscience tools and education initiative (neuroprocessing.org), currently being built.
Get in touch