Statistical learning of an artificial language: combining ERP and behavioral measures to disentangle implicit and explicit processes

Sánchez-Blanco, L. 1 , François, C. 2, 3, 1 , Laine, M. 4 & Cunillera, T. 1

1 University of Barcelona
2 cognition and brain plasticity unit
3 Attention, perception and acquisition of language group
4 Abo Akademi University

Behavioral evidence for successful segmentation of speech streams by computation of statistical regularities is based on explicit tasks performed after the learning phase. On the other hand, the on-line learning process has been investigated with electrophysiological recordings. In the current study, we went one-step further and combined electrophysiological and implicit/explicit learning measures to explore speech segmentation of an artificial language. With this goal, we recorded EEG during an adaptation of the Rapid Visual Serial Repetition (RVSR) task in the auditory modality. During a first phase, participants performed a syllabic detection task (an implicit behavioral measure) while being exposed to short statistically structured sequences of tri-syllabic words. During a second phase, participants were tested with a word recognition test to explore their ability to recognize the words presented during the Rapid Auditory Serial Repetition task (an explicit behavioral measure). We will present preliminary behavioral and ERP results on the contribution of implicit and explicit processes to speech segmentation.