Performing Identification and Discrimination Experiments for Vowels and Voiced Plosives by Using a Neurocomputational Model of Speech Production and Perception

A neurocomputational model of speech production and speech perception is introduced. After training, i.e. after mimicking early phases of speech acquisition, the model is capable of producing and perceiving vowels and CV-syllables (C = voiced plosives). Different instances of the model were trained for representing different “virtual subjects” which are then used as listeners in identification and discrimination experiments. First results indicate that a typical feature of speech perception – i.e. categorical perception – occurs in a straight forward way from our neurocomputational productionperception model.