A new framework of neurocomputational model for speech production

Speech production is complex for the brain to control, since it involves many neural processes such as speech planning, motor control, auditory and somatosensory feedback. Those functions are thought to work both in cascaded and parallel, and the control signals are transformed from one brain area to others with “one-to-many” relations. To describe this situation, in this study, we developed a new framework for a neuro-computational model for speech production based on our previous studies. The proposed model is used to deal with dynamic properties of speech articulation for consonant-vowel (CV-) syllables. In our simulation, the neuronal groups (i.e., motor, auditory and somatosensory) were acquired by learning and stored in the self-organizing maps (SOMs), and those relations between the SOMs were investigated. The results show that the time-varying properties were represented properly. In the control signal flow, the model demonstrated “one-to-many” projections between the SOMs, where one neuron in an SOM on average was projected onto 1.64 neurons in another SOM.