Model-free and model-based learning mechanism for lexical acquisition

2015 Fiscal Year Final Research Report

• Project Area: Elucidation of neural computation for prediction and decision making: toward better human understanding and applications
• Project/Area Number: 23120003
• Research Category: Grant-in-Aid for Scientific Research on Innovative Areas (Research in a proposed research area)
• Allocation Type: Single-year Grants
• Review Section: Complex systems
• Research Institution: Keio University
• Principal Investigator: IMAI Mutsumi 慶應義塾大学, 環境情報学部, 教授 (60255601)
• Co-Investigator(Kenkyū-buntansha): MATSUI Tomoko 東京学芸大学, 国際教育センター, 教授 (20296792) ; OKADA Hiroyuki 玉川大学, 工学部, 教授 (10349326)
• Co-Investigator(Renkei-kenkyūsha): KITAJO Keiichi 理化学研究所, 脳リズム情報処理連携ユニット, ユニットリーダー (70302601) ; TOMONAGA Masaki 京都大学, 霊長類研究所, 教授 (70237139)
• Project Period (FY): 2011-04-01 – 2016-03-31
• Keywords: モデル化

Outline of Final Research Achievements

A fundamental question in language development is how infants start to assign meaning to words. We hypothesized that three cognitive skills play a critical role: sound-symbolism; symmetry (bidirectional) inference bias; and inference for other individuals’ intent. We experimentally established that: (1) Preverbal infants map auditory words onto visual input by recruiting a cross-modal perceptual processing system and a nascent semantic network. With fMRI study, we identified the right posterior STS as the key area for processing sound symbolic words. 2) In a cross-species comparison study with comparable procedure, human infants posses the symmetry inference bias but chimpanzees did not. 3) In inferring the reference of a novel word, 24-month-old infants are able to infer the speaker’s confidence, and believed that the object that the more confident speaker mapped the novel word would be the correct reference.

Free Research Field

実験系心理学