Predicting speech: Cortical oscillations and perceptual dynamics

Speech is a complex signal. When we listen in our native language, however, we give this fact little thought; the task appears effortless and automatic. Simultaneously, through the acquisition of and experience with our native language, we subconsciously know how our language is structured at the level of sounds, words, sentences. One mechanism that could shape the fast and effortless nature of language comprehension is leveraging prior knowledge of our language’s structure to predict upcoming information at all levels of linguistic analysis, thereby facilitating the processing of future information.

This is a five-year research program aiming to identify the perceptual and neural dynamics, both temporal and spatial, of prediction at the earliest sensory stages during language comprehension: speech perception and spoken word recognition. Understanding how these predictive mechanisms operate during speech perception and spoken word recognition is uniquely important due the extent of physical variation in the speech signal and its importance for further stages of language comprehension. Ultimately, the research program will incorporate findings from linguistics with methods in cognitive neuroscience, utilizing both electroencephalography (EEG) and perceptual techniques. This interdisciplinary approach will focus on the when and where of prediction during speech perception: the nature of the speech sound representations involved in listeners’ generation of predictions and the cortical nature of the prediction mechanisms themselves.