Linguistic pitch is hierarchically encoded in the right ventral stream

A challenge for speech perception research is to explain how listeners transform speech sounds into meaning. Using MEG and complementary analytical methods, we expose the neural dynamics that generate meaningful linguistic units (e.g., sentence type: question, statement) that are abstracted from lower-level perceptual cues (fundamental frequency (f0) modulations). In analogy to the well-established left-lateralized hierarchical organization of a ventral stream in the domains of vision and auditory processing, we show that the emergence of abstract f0 representations is supported by a right-lateralized progression that starts in early auditory areas and advances along the posterior-anterior axis of the superior temporal gyrus (STG). Furthermore, our data indicate that the accumulation and integration of sensory information is supported by fast (gamma) and slow (alpha) interhemispheric oscillatory dynamics in a more distributed, bilateral network that informs the decision processes that drive participants' behavior. In short, hierarchical processing of f0 in the right ventral stream gives rise to abstract neural representations of linguistic pitch critical for speech perception and behavior.