Body sides analysis in hemiparesis patients (Convergence point)
In this collaborative project, including the Reha Clinic in Rheinfelden and the ETH Zürich, we investigated recovery trends in hemiparetic patients afters troke using wearable IMUs. We evaluated algorithms for unsupervised stride detection, walking parameter estimation and trend quantification.
Innovation
We investigated how potential recovery can be described intuitively and developed the convergence point metric. The convergence point could be used by the clinician as an indicator of patients progress during the rehabilitation and indicate potential differences in body sides during walking.
Figure 1. The illustration shows our convergence point estimation for unsupervised walking analysis in free-living, including (1) IMU sensor data processing, (2) unsupervised walking segment extraction, (3) stride segmentation and movement parameter estimation, and (4) the recovery trend analysis based on convergence estimation.
We showed how ubiquitous wearable computing could be used for continuous,
objective movement quantification to devise clinical relevant health recommendations. Published results indicate, that the clinical evaluated approach could be extended to analysis in free-living, i.e. at the patients home.
Bilateral trend analysis using convergence points for the movement parameter stride duration (mean and standard deviation). Recording days where walking was extracted relative to study begin are indicated by markers, dashed lines indicate recovery trends.
With our research we showed that health marker can be derived in unsupervised settings, thus enabling continuous monitoring. Such continuous monitoring of patients provides insights in patient behaviour and valuable information for clinicians, and caregivers to adapt therapy recommendations.
Work with us
For this project, we aim to extend our approach towards home-monitoring of patients. Hence, we offer multiple projects ranging from the development of battery-efficient sensor platforms for long-term recordings till the implementation of unsupervised algorithms to detect activities and behaviour patterns.
Publications
[publication-list ids=”VADX2QCE”]
Contact
Adrian Derungs
- Job title: Researcher
- Phone number: +49 9131 85-23605
- Email: adrian.derungs@fau.de