Speaking in front of a mirror improves visual speech intelligibility in healthy adults

Previous research has shown that talkers who could see themselves in a mirror were judged to be less intelligible than those who did not have access to real-time visual self-feedback (Casserly & Marino, 2024). However, this finding was specific to acoustic intelligibility; the present study examined the same speech for changes in VISUAL intelligibility by asking naive participants to complete a lipreading discrimination task with visual recordings of the previously analyzed tokens. In those tokens, speakers (N = 13) produced 139 isolated English words with and without mirror feedback. Perceivers (N = 39) were then shown orthographic prompts followed by silent video of a speaker articulating and were asked to indicate whether each word/articulation pair matched. Prompts were either accurate representations of the speech (e.g., “postcard”) or a close phonetic “lure” (e.g., “coast guard”) with minimal segmental differences from the target (mean edit distance 1.3 segments), randomized across trials. Overall accuracy in this lipreading discrimination task was 67.78% (SD = 6.5%), with significantly better performance in response to speech produced in front of a mirror (69.8% vs. 65.7% no-mirror). Although negative effects of a mirror have been observed for acoustic intelligibility, therefore, there appear to be corresponding positive effects on intelligibility in the visual domain.