汉语神经分析系统研究现状与展望

Abstract

Simulating and describing the functions of brain regions involved in speech acquisition and production based on neurophysiology and neuroanatomy is an important research topic in artificial speech synthesis systems. To facilitate this research, a team led by Professor Frank Guenther at the Boston University Speech Lab developed the Neuralynx System, an instrument that allows users to accurately express their ideas using a speech synthesis system based on 29 basic English phonemes. However, without modifications, it cannot decipher the thought processes of a Chinese speaker. The manners of articulation and the processing methods of brain mechanisms differ greatly between Chinese and English speakers. For example, the Chinese language has more than 70 basic phonemes. Therefore, the design established by Professor Guenther and his team will need to be supplemented and modified in order to construct a Chinese Neuralynx System (CNS) suitable for Chinese thought processes. To achieve this goal, we analyze the development trends, important features, and difficult points related to CNS. These are the main issues described and discussed in this article. The main content of this article is as follows: (1) the Neuralynx System's research status; (2) the international and domestic research status related to CNS, and its existing problems; and (3) CNS development trends. Through the information in this article, we expect that researchers will gain inspiration and learn from other researchers involved in the areas of speech acquisition and production, and Chinese brain mechanisms.