Marker Evaluation of characteristic quantitative Variables in Electroneurography
In this project – a group of three students – are analyzing how precise quantitative variables could be manually or automatically extracted in Electroneurography, which is an essential measurement method for the assessment of peripheral nerve disorders. For electrophysiological diagnostics, muscle signals are recorded with surface electrodes. The signals are triggered by electrical stimulation of efferent nerve fibers at different sites along the nerve. The recorded waveforms are analyzed by measuring quantities such as nerve conduction velocity, signal amplitude, distal motor latency. These are compared with age-matched reference values. These objective data are a relevant contribution to the diagnosis of nerve pathologies in addition to the clinical examination.
Compared to algorithms from digital signal processing, these measurements are either imprecise setting characteristic signal variables by hand in the recorded signals, or they result in non-uniform results if electroneurography equipment of different manufacturers is used for the measurement. Additionally, no standardized algorithms are currently available for the calculation of quantitative variables, which makes a comparison of clinical studies quite challenging and prohibits the usage of big data and artificial intelligence.
The project task is to develop a web-based survey, which enables medical persons to participate from their clinical computer and to set time- and amplitude markers by hand. This survey allows to collect statistics characteristic markers. Afterwards, the results of the different markers are compared in terms of accuracy to a standard electroneurography system and advanced digital signal processing algorithms. The motivation of this project is to enable the use of modern signal analysis methods in the neurologic application by showing the impact of optimal calculated quantitative variables in electroneurography. This could pave the way for machine learning for electrophysiological diagnostics in the near future.
- Tobias Philipp
- Lukas Fonk
- Christian Kanarski
- Eric Elzenheimer
- Gerhard Schmidt
- Kevin Prehn