We present a computational model that learns a coupling between motor parameters and their sensory consequences in vocal production during a babbling phase. Based on the coupling, preferred motor parameters and prototypically perceived sounds develop concurrently. Exposure to an ambient language modifies perception to coincide with the sounds from the language. The model develops motor mirror neurons that are active when an external sound is perceived. An extension to visual mirror neurons for oral gestures is suggested.

The influence of a native language on learning new speech sounds in adulthood is addressed using a network model in which speech categories are attractors implemented through interactive activation and Hebbian learning. The network has a representation layer that receives topographic projections from an input layer and has reciprocal excitatory connections with deeper layers. When applied to an experiment in which Japanese adults were trained to distinguish the English /r/–/l / contrast (McCandliss, Fiez, Protopapas, Conway, & McClelland, 2002), the model can account for many aspects of the experimental results, such as the time course and outcome of the learning, how it varies as a function of feedback, the relative efficacy of adaptive and initially easy training stimuli versus nonadaptive and difficult stimuli, and the development of a discrimination peak at the acquired category boundary. The model is also able to capture some aspects of the individual differences in learning.

To examine the relationship between speech perception and production in second language acquisition, this study investigated whether training in the perception domain transfers to improvement in the production domain. Native speakers of Japanese were trained to identify English /r/-/l/ minimal pairs. Recordings were made of the subjects' productions of minimal pairs before and after identification training. American-English listeners then perceptually evaluated these productions. The subjects showed significant improvements from pretest to post-test in perception as well as in production. Furthermore, the subjects retained these abilities in follow-up tests given three months and six months after the conclusion of training. These results demonstrate that training in the perception domain produces long-term modifications in both perception and production, implying a close link between speech perception and production.

this article was supported by grants from the National Science Foundation (LIS-9720348), the National Institute of Mental Health (MH47566), and the J. S. McDonnell Foundation (9634) . We thank Helen Neville for asking the question that led to the Hebbian formulation of the basis of the stabilization of speech representations and Susan E. Holmes for helpful input in the early formation of this research. Correspondence concerning this article should be addressed to J. L. McClelland, Center for the Neural Basis of Cognition, 115 Mellon Institute, Pittsburgh, PA 15213 (e-mail: jlm@cnbc. cmu.edu)

This study investigated the effectiveness of Baldi for teaching non-native phonetic contrasts, by comparing instruction illustrating the internal articulatory processes of the oral cavity versus instruction providing just the normal view of the tutor's face. Eleven Japanese speakers of English as a second language were bimodally trained under both instruction methods to identify and produce American English /r/ and /l/ in a within-subject design. Speech identification and production improved under both training methods although training with a view of the internal articulators did not show an additional benefit. A generalization test showed that this learning transferred to the production of new words.

Student modeling is of great importance in intelligent tutoring and intelligent educational assessment applications. However, student modeling for computer-assisted language learning (CALL) applications differs from classic student modeling in several key ways, including the lack of observable intermediate steps (behavioral or cognitive) involved in successful performance. This

Abstract not found

Abstract not found

Available online at www.sciencedirect.com

Adults have difficulty discriminating nonnative phonetic contrasts, but under certain circumstances training can lead to improvement in this ability. Despite the ubiquitous use of performance feedback in training paradigms in this and many other domains, the mechanisms by which feedback affects learning are not well understood. In this event-related functional magnetic resonance imaging study, we examined how performance feedback is processed during perceptual learning. Thirteen Japanese speakers for whom the English phonemes [r] and [l] were nondistinct performed an identification task of the words ‘‘road’’ and ‘‘load’’ that has been shown to be effective in inducing learning only when performance feedback is present. Each subject performed alternating runs of training with and without feedback, followed by performance of a cardguessing task with monetary reward and punishment out-