Proceedings on Automation in Medical Engineering
Vol. 2 No. 1 (2023): Proc AUTOMED

Rehabilitation technology, ID 729

Machine-learning based evaluation of mechanic muscle responses elicited by tSCS

Main Article Content

Eira Lotta Spieker (Charité), Constantin Wiesener (SensorStim Neurotechnology GmbH), Ardit Dvorani (Technische Universität Berlin), Christina Salchow-Hömmen (Charité), Nikolaus Wenger (Charité), Thomas Schauer (Technische Universität Berlin)


Transcutaneous spinal cord stimulation can reduce spasticity and enhance voluntary movement. However, electrode position and therapy intensity must be determined in a tuning process before the treatment. For that, electromyographic signals of the leg muscles are recorded during isolated double-pulses and categorized as “no response”, “reflex response” and “muscular response”. This procedure involves time-consuming skin preparation and electrode placement. In this contribution, mechanical muscle responses (accelerations) are recorded additionally to the electromyogram in nine healthy subjects to train a machine learning algorithm, which classifies the acceleration signals with an accuracy of 86 %, when considering EMG classification as ground truth.

Article Details

Similar Articles

1 2 3 > >> 

You may also start an advanced similarity search for this article.