Cross-modal talker familiarity: The search for amodal lexical episodes

Familiarity with lip-read (visual-only) speech from a talker facilitates perception of heard (audio-only) speech-in-noise from that same talker [L. Rosenblum et al., Psychol. Sci. 18, 392 (2007)]. This could suggest that stored speech episodes are composed of lexical items that retain amodal talker-specific characteristics. However, the Rosenblum et al. study used sentential material which differed between the lip-reading and speech-in-noise tasks. That study did not examine whether cross-modal talker facilitation was based on lexical items used in both tasks. The current experiment addressed whether amodal talker-specific characteristics are stored as lexical episodes by modifying Rosenblum et al.’s design. Subjects were first asked to lip-read words from a talker and were then asked to listen to words embedded in noise spoken by the same or a different talker. Preliminary evidence showed that speech-in-noise words were identified better when spoken by the same talker from the lip-reading task. In addition, preliminary results showed a lexical frequency by word interaction, where low-frequency old words (presented at training) were better identified than low-frequency new words (not presented at training) during the speech-in-noise task. These preliminary results suggest that cross-modal talker facilitation can be based on the retention of amodal lexical episodes.