Document resource
With an incidence rate of 3.19 per 100,000 persons in the United States and a median age of 64, Glioblastoma is the most aggressive astrocytic brain malignancy, characterized by rapid growth, invasive behaviour, and limited therapeutic options. Despite neurosurgical resection being central to management, preserving critical neurological function remains challenging, necessitating precise intraoperative strategies. Intraoperative neurophysiological monitoring (IONM) offers real-time feedback on the integrity of eloquent regions, enabling safer, more accurate resections.In this work, a deep learning strategy based on grey wolf optimization (GWO) and Convolutional Neural Network Long Short-Term Memory (CNN-LSTM) is suggested to identify the set of most relevant channels that provides a high trade-off between model accuracy and model simplicity for the detection and differentiation of glioblastoma from normal tissue using data collected from seven participants.The performance of the proposed evolutionary strategy is compared to other state-of-art techniques such as Genetic Algorithms (GA) and Particle Swarm Optimization (PSO), while the classification performance of CNN-LSTM is evaluated using traditional gradient descent optimization. Early algorithmic validation demonstrated an average testing accuracy of 93%.This suggest that Artificial Intelligence approaches can be successfully implemented for optimal channel selection to allow representation learning for the detection and classification of IONM signals. Future work would seek to incorporate these algorithms into the intraoperative workflow for realtime diagnostic and prognostic insights.jdavids@ic.ac.uk