As one of neurological tests, needle electromygraphy exam (NEE) plays an important role to evaluate the conditions of nerves and muscles. Neurology interns and novice medical staff need repetitive training to improve their skills in performing the exam. However, no training systems are able to reproduce multiple pathological conditions to simulate real needle electromyogram exam. For the development of a robotic simulator, three components need to be realized: physical modeling of upper limb morphological features, position-dependent electromyogram generation, and needle localization; the latter is the focus of this study. Our idea is to couple two types of sensing mechanism in order to acquire the needle-tip position with high accuracy. One is to segment the needle from camera images and calculate its insertion point on the skin surface by a top-hat transform algorithm. The other is voltage-based depth measurement, in which a conductive tissue-like phantom was used to realize both needle-tip localization and physical sense of needle insertion. For that, a pair of electrodes was designed to generate a near-linear voltage distribution along the depth direction of the tissue-like phantom. The accuracy of the needle-tip position was investigated by the electric field potential and camera hybridization. The results showed that the needle tip could be detected with an accuracy of 1.05±0.57 mm.
Keywords: needle EMG exam; needle-tip localization; tissue-like phantom; top-hat transform; voltage distribution simulation.