ORIGINAL

Abstract

The relationship between language and cognition has long been a point of investigation. This is particularly relevant in the case of infants who are bombarded with sensory information that they must organize into mental representations, which have been shown to elicit unique patterns of neural activity. Evidence suggests that around the age of 9-months infants may use language to help them in this overwhelming task. Recent advances in fNIRS research have shown that the BOLD signal recorded via this medium can be decoded to yield semantic representations in adults as well as in infants. This study investigates how early language knowledge influences object representations in 7- to 10-month-old infants. Language knowledge is quantified using a modified MCDI test to collect and analyze data specific to each infant. Neural object representations are compared based on: a) variation in vocabulary size b) variation in word knowledge, and c) variation in vocabulary structure. Two analysis methods are employed to study infant object representations - within-subject decoding and semantic decoding. We observe no significant relationships between object decoding and language development. Exploratory analyses suggest that representations of infants with different vocabulary structures lend themselves better to one decoding method than the other. Future directions based on this study are proposed.