Greater neural pattern similarity to the native language is associated with better novel word learning.

Introduction: Previous neuroimaging studies on bilingualism revealed that individuals tend to apply their native-language (L1) neural strategies to second language (L2) learning and processing. Nevertheless, it is still unclear how the utilization of the L1 neural strategies affects visual word learning in a new language. Methods: To address this question, the present study scanned native Chinese speakers while performing implicit reading tasks before 9-day form-meaning learning in Experiment 1 and before 12-day comprehensive word learning in Experiment 2. To quantify the application of the L1 neural strategies in novel word learning, representational similarity analysis (RSA) was used to compute the neural pattern similarity (PS) between the L1 and artificial language (i.e., cross-language PS) before training. Results: Univariate analysis revealed that reading both Chinese words (CWs) and artificial language words (ALWs) elicited activations in a typical reading network. More importantly, RSA revealed that greater pre-training cross-language PS in the left fusiform gyrus was associated with higher learning rate. Discussion: These findings directly reveal the facilitating role of the L1 neural strategies in novel word learning and further extend the assimilation hypothesis from the utilization of the L1 neural network in L2 learning to its learning outcomes.