Faculty opponent: Matthew Davis, MRC

In everyday listening, speech perception occurs under conditions where acoustic information may be incomplete or ambiguous, including background noise and variability in signal quality. Listeners must navigate an ambiguous, masked or rapidly unfolding speech signal in order to comprehend spoken language. While prediction has been widely discussed and acknowledged in speech perception, less is known about how listeners predict upcoming information when acoustic cues are limited or ambiguous particularly at the early points of the speech signal. This thesis investigated how listeners anticipate upcoming sounds and update predictions during speech perception, focusing on the fine-grained acoustic cues in the rapidly unfolding speech signal. It also explored how these processes are influenced by individual extended high-frequency hearing thresholds, background noise and spectrotemporal dynamics in the signal. By combining behavioral tasks, electroencephalography and audiological assessments, the thesis traced speech processing from acoustic detail to auditory perception and neural activity in the cortex. Taken together, across studies, the findings suggested that listeners used fine-grained acoustic cues to anticipate upcoming speech sounds and that the differences in these predictions were affected by what listeners were able to access from the signal. By integrating behavioral, audiological, and neural evidence across paradigms, this thesis offers insights into how speech perception unfolds dynamically over time through the interaction of early acoustic information and listeners’ expectations and individual differences in hearing.