Abstract

An enduring question about sensorimotor learning is how specific the acquired input-output relationship is. In this presentation, we review a series of studies in which the stimuli and conditions during speech motor learning were manipulated to study when and if generalization occurs. In our studies, the first and second formants of vowels were shifted using a real-time signal processing system; when subjects spoke one vowel, they heard themselves saying another vowel. In response to this auditory feedback perturbation, talkers spontaneously compensated by producing formants in the opposite direction in frequency to the perturbation. These compensations persisted after feedback was returned to normal, indicating that a form of sensorimotor learning had taken place. When different vowels were tested following the perturbation of one vowel, they were found not to show evidence of the feedback perturbation nor did the new vowels have any influence on the perturbed vowels’ return to normal baseline levels. Data from studies in which listening versus producing were compared and studies in which the similarity of the feedback voice quality was manipulated will also be presented. In general, the studies suggest that learning is quite local and, thus, that learning does not generalize beyond restricted conditions.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call