Computational Linguistics Colloquium
Thursday, 17 June 2010, 16:15
Neural correlates of learning a novel vocabulary
Language learning in adulthood becomes more and more of a necessity in in times of globalisation and increasing mobility. Concentrating on vocabulary, we investigated how novel words can be acquired with associative learning, and the cognitive and neural changes accompanying such learning. Associative learning mimics naturalistic learning situations. It constitutes a powerful training method in which to-be-learned novel words are coupled with existing concepts more frequently than word/concept combinations that should not be learned. We demonstrated that a large amount of new vocabulary - nouns and versbs - can be acquired with this method, and that once learned, the novel words display characteristics very similar to words from one’s native language. Associative learning is good for learning novel vocabulary, but is not suited to induce changes in the phonemic inventory of one’s native language - as we found out in an additional study.
We validated word-learning success by means of behavioural responses and neurophysiological correlates from MEG. The component of interest is the N400m, the magnetoencephalographic counterpart of the N400, which is considered to be an index for semantic (in)congruencies and/or the ease with which cues can access conceptual knowledge. As expected, novel words induced large M400 components before learning, and this amplitude was reduced after learning. Importantly, the N400 was reduced only when novel words and concepts matched. Depending on the type of linguistic material that had to be learned, different neural networks were involved in the integration of novel information into already existing lexical and semantic networks.
If you would like to meet with the speaker, please contact