How do we learn new meanings for words already known? Evidence from EEG and MEG studies

In addition to learning new words, people often learn new meanings for words they already know. For example, one might learn that the word ‘skate’ refers to a type of fish long after knowing its more common roller- or ice-skating meaning. Different from learning a new word, this type of learning requires updating the lexical knowledge by associating new information with an existing word and involves the co-activation of new and prior word knowledge. This dissertation investigates the mechanisms underlying the learning of new meanings for known words. In particular, it focuses on the influence of overnight consolidation on the learning of new meanings for known words and the role of the left posterior middle temporal gyrus (pMTG) in binding new meanings to known words. Study 1 showed that the processing of both new and original meanings became faster after overnight sleep. This indicated reduced interference between new and original meanings over time, especially after overnight consolidation occurred. However, the event-related potential (ERP) data showed that accessing the new meanings was still mainly supported by episodic retrieval even 24 hours after learning. To investigate how new meanings are associated with known words, Study 2a first demonstrated that presenting word meanings as verbal definitions were sufficient to drive a semantic category effect. Based on this result, Study 2b further showed that the left pMTG, in addition to sensorimotor cortices relevant to the representation of new meanings, was involved in binding new meanings to known words. Combined with the previous findings on learning novel words, the results suggest that the co-activation of new and prior knowledge is essential to the integration of new word knowledge into the mental lexicon. The left pMTG not only supports the formation of novel form-meaning associations, but also the associations between new meanings and previously known words.