Neural correlates of form- and meaning-based predictability during speech perception

The processing of degraded speech can be facilitated by giving people access to its content beforehand, a phenomenon known as the "pop-out effect". This study investigates the neural correlates of this effect, focusing on the distinct contributions of predictability based on the form at the word level, or based on the meaning of the sentence level. Using functional magnetic resonance imaging (fMRI), we examined how these two types of predictability influence speech perception in normal-hearing listeners. Participants were exposed to noise-vocoded speech, with varying levels of intelligibility, and provided with either matching or non-matching visual text cues. Our findings reveal that form-based representations primarily engage the auditory cortex (superior temporal gyrus and sulcus), while meaning-based representations predominantly activate higher levels of language processing regions, such as the left inferior frontal gyrus. These results suggest that form- and meaning-based representations operate through additive top-down and integrative mechanisms, enhancing speech perception by leveraging top-down processes to organise auditory input. This study provides valuable insights into the neural mechanisms underlying speech perception and has implications for developing interventions to support individuals with auditory processing difficulties.