Recovery after stroke
My PhD is on upper-limb motor control and recovery after stroke. Recovery differs enormously between people, and that variability is largely unexplained. Because of that, the mechanisms worth targeting cannot yet be identified. Characterisation is the route to knowing what to target.
Research
I study the interactions between the corticospinal and reticulospinal pathways. I design and validate paradigms that combine behavioural tasks with neurophysiological measurement, using TMS and wearable IMUs.
On the data side, I build quantitative models and statistical analyses of longitudinal motor and clinical datasets, to identify predictors of functional recovery. I also contribute to digital-health infrastructure and data pipelines that integrate sensor-derived motor assessments and patient-reported outcomes in clinical settings.
Instruments
- TMSTranscranial magnetic stimulation, the neurophysiological measurement I use to probe the corticospinal pathway.
- IMUWearable inertial sensors that record the movement of the upper limb during behavioural tasks.
- MODELStatistical models of longitudinal clinical and motor datasets, used to look for predictors of functional recovery.
MODUS · NCM 2026
“Context-Dependent Effects of Grip Force on the Strength–Dexterity Trade-Off in Dexterous Force Control”
A conference visit.
MSc thesis, ETH Zürich
ECG Construction from Fingertip and Remote Photoplethysmography
PPG to ECG
An inter-subject deep model (CNN, BiLSTM and dense layers) trained on a MIMIC-III subset, with filtering, alignment, segmentation, normalisation and derivation.
rPPG to rECG from video
Face and ROI detection on every frame, extraction and filtering of RGB time series, and rPPG features (CHROM, POS, LGI, R/G/B).
| PURE | r = 0.77 / 0.82 |
|---|---|
| LGI-PPGI | r = 0.89 / 0.68 |
Dr Mohamed Elgendi, Prof Carlo Menon, Prof Valentina Agostini
One paper accepted in Frontiers in Physiology (one of three submitted).
Maxwell Biosystems
- Temperature characterisation
- NTC sensor selection
- A Recording Unit test bench, built with an Arduino Mega and a custom board
- Development of a pressure test bench
- API integration
- Protocol design in Jira
- Documentation in Confluence
- ETL pipelines
- Data visualisation dashboards, analysis and reporting