Needle detection by electro-localization for a needle EMG exam robotic simulator

Abstract

Needle EMG (Electromyogram) Exam (NEE) is an important neurological exam, and neurology interns and novice medical need repetitive training to gain the necessary skill to perform the exam. However, until now it has been difficult to reproduce multiple pathological conditions for their training, since in most cases, trainees serve as human subjects for each other. A robotic simulator that could reproduce behavior with various pathological disorders can be of great help for NEE skill training. Needle tip localization is a key component of the robotic simulator, since position-dependent-EMG is the signal source for skilled neurologists to determine the pathological situation. The needle tip localization has been investigated for many medical tests and applications, such as prostate brachytherapy, etc. However, only few studies have been reported on the process of needle operation in muscle based on EMG signals dependent on needle tip position. Our idea is to apply a tissue-like conductive phantom so as to realize both physical sense of insertion, and needle localization for the NEE robotic simulator. A pair of electrodes was designed to generate a near-linear voltage distribution along the depth direction of the tissue-like phantom, by which the inserted needle could be localized. The influence of the shape of phantom and electrodes on detection accuracy were investigated by a set of measurement experiment and a computer simulation. The results showed that, the estimated depth values agreed with that of the computer simulation model, and the curved phantom had a much steeper distribution in the deeper region (better accuracy for needle tip detection).