Learning of new words: behavioural and neural correlates of encoding and integration
Notwithstanding the great advances in the study of neurocognitive processes involved in learning of new words in the past three decades, many questions remain regarding the nature of these processes, how they can be measured empirically and which factors influence processing during word learning. This thesis focuses on two fundamental processes in word learning, integration into semantic memory and encoding, in learning of novel names for novel concepts (simulating novel word learning in one’s first language) and for familiar concepts (simulating novel word learning in a foreign language) in neurotypical adults. Thereby, both behavioural and electrophysiological methods were employed, and tasks included both word recognition (primed lexical decision) and word production (recall, translation, picture-word interference) paradigms.
The first experimental study (Chapter 2) explored neural correlates of successful encoding using novel names for novel concepts. This study was inspired by work in the recognition memory literature, where successful encoding is typically studied with the subsequent-memory paradigm and using words that are well-known to participants. The goal of this study was to test whether the findings with well-known words can be extended to novel words, and, to achieve this, three measures of neural activity were considered: event-related potentials, time-frequency representations of power, and phase-amplitude coupling. No differences were observed during encoding of words that were later recalled as opposed to not recalled for any of the three measures. These results are discussed in the context of theories of encoding that are based on the previous findings with well-known words.
The experimental study reported in Chapter 3 focused on behavioural and electrophysiological markers of integration into semantic memory in learning of novel names for novel concepts. Integration was assessed by means of a continuous primed lexical decision task, which has been argued to be less susceptible to controlled processes of lexical-semantic retrieval than other tasks commonly used in the literature (e.g., paired primed lexical decision or primed semantic-relatedness judgment tasks). Irrespective of the length of the consolidation opportunity (0-day vs. 1-day), Bayesian inference provided evidence against the effects of interest both in the behavioural and in the electrophysiological data. Yet, there was evidence that the length of the consolidation period impacted processing of the novel names in episodic memory. These findings are interpreted in light of the Complementary Learning Systems model of word learning and general accounts of memory.
In contrast to the previous two chapters, the focus of Chapter 4 was on learning of novel names for familiar concepts (i.e., simulating word learning in a foreign language). This chapter investigated the influence of learning context (teaching of novel words blocked by semantic category vs. in mixed semantic categories) on encoding, immediate recall, and integration of these novel names. This experimental manipulation was motivated by the fact that introducing novel vocabulary blocked by semantic categories is a common practice in second language teaching, despite the limited evidence for its effectiveness. The categorically related learning context resulted in lower accuracy during training, longer response times at immediate recall and greater semantic interference in the picture-word interference task. These effects are interpreted as being due to differences in episodic rather than semantic memory.
Overall, this thesis contributes to understanding of the encoding and integration processes and their manifestations in behaviour and electrophysiology in learning of novel names both for familiar and for novel concepts. It highlights which aspects of the most prominent theories of learning and memory remain under-specified and/or under-researched, and proposes directions for future research.