Brain activation during novel word encoding predicts lexical integration

Acquisition of a novel word involves the integration of a newly formed representation into the mental lexicon, a process which is thought to benefit from offline consolidation. Brain activity during post-learning sleep has been shown to relate to behavioural measures of lexicalisation (Tamminen et al., 2010; 2013), suggesting that the outcome of acquisition is indeed at least partly determined after encoding. It is however unknown to what degree the neural response during the learning phase itself influences successful lexicalisation. A consistent body of evidence indicates that activation in medial temporal, parietal and frontal areas during encoding predicts subsequent memory strength (Kim, 2011), suggesting that encoding-related factors may also affect offline integration processes. In the present study we combined and extended these two lines of research and asked whether encoding-related neural activity is related to subsequent lexical integration as well as explicit memory. Specifically, we hypothesised that immediate orthographic and semantic integration during the first few encounters with novel words predicts their later ability to interact with existing words. Participants studied 40 novel printed words, each paired with a picture of a common object illustrating its meaning, while their neural responses were measured using functional magnetic resonance imaging. A primed visual lexical decision task was administered approximately 24 hours after encoding. In this task, participants made lexical decisions to existing and pseudo-word targets, which were each preceded by a briefly presented novel word that was either semantically related or unrelated to the target. Faster response times to related versus unrelated pairs suggest that links have been formed between the novel-word representations and their semantic associates. Priming effects can therefore be considered a strong indication that novel words have been lexically integrated. Following the priming task, cued and free recall tasks probed explicit memory for the learned novel words. A significant priming effect was observed, suggesting that those novel words that had been encoded successfully were sufficiently lexicalised to influence recognition of their existing semantic associates. In line with previous findings, words that were correctly recalled in the test session elicited enhanced activation in the left inferior frontal gyrus (IFG) during encoding. Similarly, words that subsequently produced priming effects showed enhanced IFG activation compared to words that had no facilitating effect. Crucially, a set of additional clusters predicted subsequent priming but not memory persistence. These were found in left temporalparietal regions involved in semantic processing, as well as in a posterior portion of the left fusiform gyrus known as the visual word form area (VWFA). These data suggest that increased orthographic and semantic processing during encoding facilitates lexicalisation. We argue that enhanced VWFA activation during encoding reflects the formation and integration of a stable orthographic representation. This enables rapid lexical access to the novel word, which in turn facilitates retrieval of related words and hence boosts their recognition. In conclusion, successful lexicalisation is determined in part by the engagement of encoding mechanisms that stimulate memory integration, above and beyond those supporting memory formation