Formation of an Auditory Map for Invariant Perception of Vowel Sounds: Listening to a Variety of Speakers to Make Unified Vowel Representation
We propose a neural network model that function as a cognitive map for the perception of vowel sounds. By simulating the model, we investigated neuronal bases for the encoding and perception of vowel sounds. The model consists of two networks, by which vowel sounds are processed in a hierarchical manner. The first network, which is tonotopically organized, detects spectral peaks catted formant frequencies. The second network receives input from the first network in a convergent manner and detects the combinatory information of the first (Fl) and second (FS) formant frequencies. We trained the model with five Japanese vowels spoken by different people and modify synoptic connection strengths of the second network according to the Hebbian learning rule. The present model can recognize not only learned vowel sounds but also some unknown vowel sounds. We suggest that such perception of the unknown vowel sound is possible if the sound activates the neurons that are members of the cell assembly whose activation encodes the information about the category of the vowel to which the sound belongs. We also suggest that the synaptic reorganization of the neural network may be a key mechanism for neural representation of the information about vowel sounds in the brain.
M. Miyamoto O. Hoshino M.Zheng K. Kuroiwa
Department of Electrical and Electronic Engineering of Oita university 700, Dannoharu, Oita, 870-119 Department of Welfare Engineering of Oita university.700, Dannoharu, Oita, 870-1192, Japan Department of Information Network Science, Graduate School of Information Systems,The university of
国际会议
8th International Conference on Neural Information Processing(ICONIP 2001)(第八届国际神经信息处理大会)
上海
英文
1093-1098
2001-11-14（万方平台首次上网日期，不代表论文的发表时间）