Language is hierarchically organized into syllables, words, phrases, and sentences. For spoken language processing, building hierarchical linguistic structures is a fundamental yet challenging task. In general, the boundaries between words and phrases are not clearly defined by acoustic features and finding such boundaries critically relies on the listener's linguistic knowledge. In a series of magnetoencephalography (MEG) experiments, it is shown that cortical activity is concurrently entrained to the rhythms of syllables, phrases, and sentences, which is not confounded by the tracking of acoustic or statistical properties of speech. Recent studies also show that the grouping of syllables into larger linguistic units, e.g., words, critically relies on top-down attention. In summary, under the modulation of top-down attention, cortical computations driven by linguistic knowledge can generate slow rhythms matching the time scales of larger linguistic structures, which provides a plausible mechanism for online building of large linguistic structures.
Biography
Ding Nai is an Assistant Professor at the College of Biomedical Engineering and Instrument Sciences at Zhejiang University.
Neuroscience Seminar Series by the NYU-ECNU Institute of Brain and Cognitive Science at NYU Shanghai