Slide Slam P15 Sandbox Series
The Emergence of the Neural Representation of Novel Words through Narrative Reading
Isil Poyraz Bilgin1,2, Samuel Guay3,4, Robert Oostenveld5, Valentina Borghesani3,6, Slawomir Jaroslaw Nasuto1, Douglas James Saddy7; 1Biomedical Engineering, Biological Sciences, University of Reading, Reading, UK, 2Allied Health Professions, The University of the West of England, Bristol, UK, 3Department of Psychology, University of Montreal, Montreal, Canada, 4Centre de recherche CIUSSS NÎM, Montreal, QC, Canada, 5Donders Institute for Brain, Cognition and Behaviour; Radboud University, Nijmegen, The Netherlands, 6Centre de recherche de l'Institut universitaire de gériatrie de Montréal, Montreal, QC, Canada, 7School of Psychology and Clinical Language Sciences, University of Reading, Reading, UK
Learning words is a process that requires gathering cues from the context the word appears in and integrating these cues to build a meaningful semantic representation (Caramazza et al.,1990). Throughout learning selective attention processes are employed in capturing contextual statistical regularities and building meaningful associations for further categorisation of the novel word (Devlin et al. 1998, Nofosky et al, 1986). A wealth of studies show that the successful semantic processing and categorisation of a word is distributed over specific linguistic and sensory-motor systems, to the extent that the category of the word can be reliably predicted from patterns of activations within those brain regions (Mitchell et al., 2008; Huth et al., 2012; Shinkareva et al., 2012). Given the distributed nature of semantic processing, the emergence of the neural representation of a novel word is expected to follow a distributed fashion over the functional areas associated with the processing of the corresponding semantic information. Leveraging known differentiating neural signatures associated with the processing of familiar vs unfamiliar (David & Gaskell, 2009), old vs new (Vuilleumier et al., 2002) and frequent vs infrequent (Fiebach et al., 2002) words, we set out to investigate the emergence of the neural representations of novel words in an experimental setting in which the amount and the sequence of semantic information available were experimentally controlled. We acquired EEG and fMRI data simultaneously from 28 healthy right-handed subjects on a silent narrative reading task. Subjects were trained on a set of sentences introducing three different semantic features (habitat, visual, functional) of a group of made-up and real objects from two semantic categories (tools, animals). Each semantic feature was introduced in a separate block and repeated six times to reinforce learning. Each learning phase was followed by a semantic congruency test to quantify the learning success. Our preliminary analyses of behavioural performance confirm that subjects achieved successful word learning. We will now focus on testing the temporal and topographical hypotheses leveraging the complementary strength of simultaneous recording of the two neuroimaging modality. Our initial aim is to observe the temporal trajectory of the comprehension of the novel word knowledge by investigating one of the well-established neurophysiological markers of the meaning integration, N400, a negative deflecting EEG component following the 400 ms after the stimulus onset. As an indicator of the successful encoding of a novel meaning, we expect to observe a training-related attenuation of the N400 amplitude that is elicited due to the incongruent matchings between the novel word and an unrelated semantic context. Secondly, we expect the emergence of multivoxel activity patterns in a priori defined brain regions based on previous work (i.e. parahippocampal gyrus, precuneus, occipito-temporal gyrus, premotor cortex, left inferior frontal gyrus, fusiform gyrus) (Bauer & Just, 2015; Chao & Martin, 2000; Binder et al., 2009) that follows the sequential introduction of the semantic attributes associated with the novel word, such that it predicts the semantic property and the level of novelty of the information that is encoded.