ISCA Workshop on Plasticity in Speech Perception (PSP2005)

Senate House, London, UK
June 15-17, 2005

When and Why Feedback Matters in the Perceptual Learning of Visual Properties of Speech

Stephen Winters, David Pisoni

Indiana University, Bloomington, IN, USA

This study investigated the effects of feedback on the perception of words in point-light and fullyilluminated displays of speech. Participants attempted to identify individual words that they could see (but not hear) being spoken. After attempting to identify the word in each visual-only stimulus, the participants were informed of the identity of the word by receiving feedback in one of three forms: seeing the visual stimulus again, with sound (AV feedback); hearing only the audio track from the original visual stimulus (audio-only feedback); reading the word that was spoken (orthographic feedback). Another group of participants, in a control condition, did not receive any feedback after each trial (no feedback).

It was expected that correct identification scores would improve more over the duration of the experiment for the groups of participants that received feedback than for the group of participants that did not receive feedback. It was also expected that correct identification scores would improve more for the groups that received feedback which provided event-related information about the original stimulus (AV feedback, audio-only feedback) than for the group which received orthographic feedback, which is abstract in nature and not directly related to the event(s) that the participants are being asked to perceive.

It was found that effects of feedback on correct identification scores only emerged when the participants saw stimuli again, after they had already received feedback on them. Under these conditions, the groups of participants who received feedback correctly identified the visual-only stimuli significantly better than did the group of participants who received no feedback. Furthermore, the type of feedback also had an effect on the amount of perceptual improvement made by the participants who saw the point-light stimuli. With each successive repetition of a stimulus, the group of participants who received audio-visual feedback had increasingly higher correct identification scores than did the participants who received either audio-only or orthographic feedback. This pattern of results suggests that providing feedback to participants in the form of the original stimulus facilitates the perceptual learning of repeated stimuli more than does feedback which only provides derived or abstract information about the original stimulus.

Full Paper

Bibliographic reference.  Winters, Stephen / Pisoni, David (2005): "When and why feedback matters in the perceptual learning of visual properties of speech", In PSP2005, 148-151.