Development of Real-Time Monitoring Systems with Comprehensive Intraoperative Electromyography (EMG) Data: A Preliminary Study on Brain Tumor Surgeries

Intraoperative neurophysiologic monitoring (IONM) is a vital technique to preserve the functional integrity of nerves during surgery. It plays an important role in preventing nerve damage by providing instant feedback to the surgeon, especially in brain tumor surgeries where motor and sensory nerves should not be damaged. Data on this surgery are very rare. In this study, for the first time in the literature, intraoperative electromyography (iEMG) data from 11 brain tumor surgery patients, labeled by professional neurophysiologists, were used for classification. This unique dataset, which is collected from 6 different muscle regions of the human face and contains 7 different signal events each, allows for signal identification and large-scale analysis. The features in this dataset were extracted using the 50% overlap method. These features were then analyzed with 5 different classifiers. In the analysis, the Bagged Trees model gave the most successful results with an accuracy rate of 97.13%. It is thought that this study may be effective in laying the foundations for a patient-independent, highly accurate and automatic warning system that surgeons can instantly monitor the patient's nerve damage status during brain tumor surgeries over a large data set.