Predictive language comprehension and neural synchrony

Do brain rhythms and neural synchrony relate to how predictable the words in stories are? We recorded magnetic fields (magnetoencephalography) when participants listened to short stories. With techniques from natural language processing, we built computer models of how surprising the words in the stories are. We then looked at thow brain waves at different frequencies (4Hz and up to 60 Hz) change when words hit you as surprising. It turns out the mid-fast brain waves (at around 20 Hz) are tracking how unexpected the words are. The slow waves (around 4 Hz) seem to track how unpredictable the sentence context is, rather than words alone.

To learn more, read the paper.

Avatar
Kristijan Armeni
postdoctoral fellow

Publications