Orofacial somatosensory inputs modulates perceptual shifts induced by the speech motor learning
Résumé
Orofacial somatosensory inputs during listening to speech sounds change the speech perception (Ito et al. 2009). This implies auditory and somatosensory inputs would be consolidated during speech learning and, as a result, this linkage effects on subsequent speech perception. This study examined how auditory and somatosensory inputs during speech learning would contribute to speech perception. We focused on perceptual shifts of speech sounds induced by the auditory-feedback learning during speech production (Shiller et al., 2009; Lametti et al., 2014) and examined if the shifts can be modulated by additional somatosensory stimulation associated with facial skin deformation. Specifically, we compared performance of perceptual tests prior and post to speech training. In the speech training, subjects repeated a task phrase ‘a shed’ and received altered auditory feedback in on-line manner. We manipulated the spectral centroid frequency (Shiller et al., 2009) so that the produced /sh/ sounds can be heard as more /s/-like sounds. In addition to the altered auditory feedback, a robotic device pulled subjects’ cheek forward or backward to apply the somatosensory stimulation. Native speakers of American English were assigned into one of four conditions, which are the altered-auditory feedback with the forward skin stretch (ASF), the altered-auditory feedback with the backward skin stretch (ASB), the altered-auditory feedback (A) and no perturbation (control) conditions. As for the perceptual tests, identification testing using /s/-/sh/ acoustic continuum was carried out. The adaptation effect was quantified as a change in the spectral centroid frequency of the produced /sh/ sounds. The perceptual shift was quantified as a change in the perceptual threshold for /s/-/sh/ between prior and post to the speech training. The adaptation effect in the three perturbed conditions (ASF, ASB and A) were larger than one in the control condition, and the effect in the three perturbed conditions were not different. This suggests that subjects were adapted to the altered auditory feedback, and the somatosensory stimulation does not have a large effect on the speech motor adaptation. Nevertheless, most notably, the perceptual shift in ASB was different from the perceptual shifts in ASF and A, suggesting that somatosensory inputs during the speech training modulated the perceptual shifts depending on the direction of stimulation. Taken together, the somatosensation during speech production may be a key to tune the auditory representation for speech perception rather than motor outputs.