• Top page
  • Timetable
  • Per session
  • Per presentation
  • How to
  • Meeting Planner

演題詳細

Poster

感覚運動系の学習・可塑性
Sensorimotor Learning/Plasticity

開催日 2014/9/13
時間 14:00 - 15:00
会場 Poster / Exhibition(Event Hall B)

劣化雑音音声の理解プロセスの聞き取り能力と脳活動の相関関係による検討
Correlation between brain activity and performance of listening test: recognition process of the noise-vocoded speech sounds

  • P3-106
  • 村井 翔太 / Shota Murai:1 小林 耕太 / Kohta I Kobayasi:2,3 力丸 裕 / Hiroshi Rikimaroux:1,2,3 
  • 1:同志社大院生命医科学 / Grad Sch of Life and Med Sci, Doshisha Univ, Japan 2:同志社大院生命医科学 / Dept of Biomed Information, Fac of Life and Med Sci, Doshisha Univ, Japan 3:同志社大学ニューロセンシング・バイオナビゲーション研究センター / Neurosensing and Bionavigation Res. Ctr, Doshisha Univ, Japan 

Noise-vocoded speech sound (NVSS) has been used to reveal speech perception when the sound is severely deteriorated. It contains no information about the fundamental frequency or precise formant peaks but only slow temporal changes in the amplitude envelope. Recent studies revealed that normal-hearing subjects were able to recognize NVSS after training. Though there were individual differences in the comprehension of distorted speech, the neural mechanisms involved in the individual difference were still largely unknown. To investigate brain regions contributing the differences, this study compared the individual NVSS comprehension scores in listening test with brain activities measured by functional MRI with a sparse sampling paradigm. The original speech was a sentence consisted of 13-16 morae Japanese sentence (2-3 s). NVSS was created by dividing the original speech into 3 bands (60-600, 600-1500, 1500-4000 Hz). The amplitude envelope from each band was extracted and the each envelope was multiplied with a band noise in the same bandwidth. In test sessions, comprehension of NVSS was evaluated before the MRI session. Adult native Japanese speakers listened to NVSS and reported how they perceived NVSS without feedback in a soundproof chamber. In MRI session, participants listened to another NVSS. Then, the same NVSS was repeated while the original sentence was provided on screen as text. As a result, average percentage of correct morae was 15%, and standard deviation of correct percentage was 17%. The individual differences were quite variable within subjects. A multiple regression analysis with the individual scores of pre-scanning test was performed on individual contrast images while listening to NVSS of each subject. The individual variation in the NVSS comprehension score were significantly correlated with BOLD signal of left inferior frontal gyrus (IFG). These results suggested that brain activity in the left IFG may play a critical role in subjects' ability to understand NVSS.

Copyright © Neuroscience2014. All Right Reserved.