The Identification and Clustering Analysis of Auditory Neuronsfor Salicylate-Induced Rat Model

• Main Authors: Li-hui Chen, 陳俐卉
• Other Authors: Kuo-Sheng Cheng
• Format: Others
• Language: en_US
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/55832954871713464698

碩士 === 國立成功大學 === 醫學工程研究所碩博士班 === 93 === 　　Salicylate-induced rat model is one of the animal models for tinnitus study. Since the mechanisms for the tinnitus are varied a lot, there is no break-through progress in related studies. Through the study of the salicylate-induced rat model, the conduction pathway for the auditory neurons in the brain may be activated. These activated neurons may then be fos-labeled for analysis. From the identification and distribution analysis, some mechanism of tinnitus may be revealed from the Salicylate-induced rat model study. In this study, in order to overcome the disadvantages of the manual identification of auditory neurons, a radial basis function neural network for automatic identification is developed due to its features of easy training and learning. From the experimental results, the recognition rate is demonstrated to be as high as 98%. Not only the recognition rate is improved, but also it is very objective in analysis. In addition, a support vector clustering is applied to neurons distribution analysis. Traditionally, the distribution is only characterized in number or density of neurons. Using the support vector machine, the clustering feature may be obtained as another possible parameter for analysis. Base on the clustering analysis, it is found that the cluster number and distribution area for the Salicylated-induced fos-labeled neurons are very different from those of controlled group. However, it needs to be further investigated in physiology.