Temporal Predictions in Perception and Language Comprehension

During perception, information is not uniformly distributed over time; some time-points are more informative than others. How does the brain get ready to perceive upcoming bursts of information? When people listen to quasi-periodic stimuli such as speech or music, electrophysiological activity in cerebral cortex synchronizes with pulses in the stimulus. Cortical synchronization helps to align information in the stimulus with periods of maximal neural excitability. However, it remains unknown how the brain uses temporal structure in the stimulus to drive cortical synchronization. In Study 1, I test whether cortical synchronization depends on characteristics of sensory cortex, or on informational characteristics of the stimulus. I develop a metric to quantify visual information over time, and use electroencephalography (EEG) to test for cortical synchronization to temporal structure in sign language. Results show that cortical synchronization depends on the structure of information in the stimulus, not on modality-specific perceptual processes. Cortical synchronization is often thought to rely on oscillatory resonance to rhythms in the stimulus, analogous to harmonic resonance in other physical and biological systems. However, the brain synchronizes to both isochronous and non-isochronous sequences, raising the possibility that temporal expectations may rely on a non-oscillatory mechanism. In Study 2, with two EEG experiments and a computational model, I show that oscillatory resonance can generate predictive neural activity to predictable non-isochronous rhythms. Resonance can lead to behavior that appears "smart", despite being solely stimulus-driven. Together, these studies suggest that oscillatory resonance provides a flexible, domain-general mechanism that can support temporal predictions across sensory modalities, enabling the cerebral cortex to maximize sensitivity to informational peaks in time-varying signals