AV speech elicits in the listener a motor plan for the production of the phoneme that the speaker might have been attempting to produce, and that feedback in the form of efference copy from the motor system ultimately influences the phonetic interpretation.I mentioned in a previous post that this claim is dubious based on the observation that infants, who can't generate speech, nonetheless exhibit McGurk-MacDonald effects. William Matchin in my lab recently followed up on this issue in two experiments, reported in JoCN.
Experiment 1 was a behavioral study in which participants viewed AV mismatched stimuli that give rise to fused McGurk perceptions. They view these stimuli either under typical conditions or during articulatory suppression (silent syllable articulation). The idea is that if the motor system is important in driving McGurk effects, then distracting the motor speech system with an articulatory task should affect the perception of AV mismatched stimuli, e.g., by reducing the strength of the McGurk effect.
Results: Articulatory suppression had no impact on the McGurk effect.
Experiment 2 was an fMRI study that explored whether motor speech areas, such as Broca's area exhibited activation profiles characteristic of cross-sensory integration, namely that the response to the AV stimulus is greater than the response to the auditory or visual stimuli alone.
Results: Motor speech areas do not exhibit a response pattern characteristic of AV integration. Instead, the STS exhibited such a profile (as previous work had shown).
Conclusion: The motor system is not the source of AV integration effects. All signs currently point to the STS. If you're interested in the evidence regarding the role of the STS, Michael Beauchamp's work is a good place to start.