Neural basis of non-repetition based rule learning in infants: a NIRS study

Martinez-Alvarez, A. 1, 2 , Gervain, J. 3, 4 , Pons, F. 1, 5 & de Diego-Balaguer, R. 1, 2, 6

1 University of Barcelona
2 Cognition and Brain Plasticity Unit
3 CNRS
4 Université Paris Descartes
5 Institute for Brain, Cognition and Behaviour
6 ICREA

A core ability in infant language development involves the extraction of linguistic rules. Previous studies investigated the brain basis of repetition-based rule learning (e.g. ABB: ?mubaba?, ?penana?, ABA: ?bamuba?, ?napena?) in infancy and found increased responses to the repetition sequences in temporal and left frontal regions (Gervain et al. 2008). However, the neural basis of infant non-repetition based rule learning (AXC-type) remains unexplored. Behaviorally, such rules are first learned at 15 months of age (Gomez & Maye, 2005). Such long-distance dependencies between non-identical items are commonly found in syntax and morphology (e.g. is walking, unbelievable). To examine this type of rules we presented 9-10-month-old infants (n=25) with sequences containing an AXC-type structure, where the A and C syllables predict one another with certainty (e.g., ?pedibu?, ?pegabu?) or a random control structure where there is no such dependency (e.g., ?dibupe?, ?bugape?). We increased the pitch of the first and third syllables, highlighting the items over which a rule could be extracted in the rule condition, but not in the no rule condition. Infants? brain activity was measured using functional near-infrared spectroscopy (fNIRS). Preliminary data show a tendency for the rule condition to produce a larger activation (oxyHb) in temporal and frontal areas, more prominent over right-lateralized regions. The present results suggest that pitch, orienting infants? attention, might highlight dependencies that cannot otherwise be learned yet at a given age. This study may contribute to our understanding of the brain basis of rule learning processes underlying language acquisition.