Development of a neuro-searching system

本研究的目的是开发一个系统，以表明神经的位置，相对于断层成像图像，在外科手术中。在该系统中，我们通过专门设置的多偶极电极刺激周围神经并检测相应的诱发电位。电极相对于层析成像图像的位置从铰接导航系统读取。然后将估计的神经位置叠加在CRT上的断层图像上。为了成功地运行，提高信噪比和减少平均过程所需的迭代测量次数是非常重要的。在本研究中，我们首先开发了一种特征映射型神经网络对特征空间中的诱发电位数据进行平均，以提高平均效率。其次，我们应用反向传播网络来估计神经位置，相对于探测电极，从一组诱发电位波形。通过计算机模拟和/或志愿者的基础实验，我们证实了这些网络的有效性。

The objective of the research is to develop a system to indicate the position of the nerves, with respect to tomographic images, during surgical operations. In this system, we stimulate the peripheral nerves and detect the corresponding evoked potentials via multi-dipole electrode, specially arranged for this measurement. The position of the electrode, with respect to the tomographic images, is read from the articulated navigation system. Then the estimated nerve position is superimposed on the tomographic image on CRT.For the successful operation, it is very important to improve the signal to noise ratio and reduce the number of iterative measurements required for the averaging process. In this research we, at first, developed a feature-map type neural network to average the evoked potential data in the feature space, to improve the averaging efficiency. Secondly, we applied a backpropagation network to estimate the nerve position, with respect to the probing electrodes, from a set of evoked potential waveforms. With computer simulation and/or fundamental experiments with volunteers, we confirmed the effectiveness of these networks.