At Charité – Universitätsmedizin Berlin’s Kühn Lab, I built an automated pipeline to detect beta-power drift across chronic LFP segments recorded from Parkinson’s patients with implanted DBS devices. I designed per-day median/MAD normalisation and clock-hour windowing after global z-scoring suppressed the drift signal of interest.
I traced batch failures to a daylight-savings timestamp error affecting six aDBS segments, corrected it, and brought the batch to 109/109 successfully processed. I also standardised six diagnostic plot types per segment for lab-wide quality control.
Through my ongoing NEURO REACH project, I’m developing a closed-loop rehabilitation system that detects voluntary eye movements to trigger functional electrical stimulation. My other ongoing project is a biomaterials design concept for an enzyme-responsive cortical scaffold to guide regeneration after stroke.
