A Novel Application of Deep Learning (Convolutional Neural Network) for Traumatic Spinal Cord Injury Classification Using Automatically Learned Features of EMG Signal

This study explores using a Convolutional Neural Network (CNN) to classify traumatic spinal cord injuries (TSCI) based on electromyography (EMG) signals in a non-human primate model. The CNN's performance is compared to a classical method (k-Nearest Neighbors, kNN). The goal is to develop a tool for evaluating the effectiveness of TSCI treatments. The study uses intramuscular EMG data from tail muscles of five monkeys before and after spinal cord lesion. The CNN uses filtered EMG signals, while kNN uses hand-crafted EMG features.

• 1
  The CNN shows promise as a classification technique for TSCI compared to traditional machine learning.
• 2
  kNN achieved F-measures of 89.7% and 92.7% for left and right side muscles, respectively, while CNN achieved 89.8% and 96.9%.
• 3
  The deep learning model (CNN) demonstrates high potential for use as a TSCI classification system.