Looking at early word segmentation and mapping through pupillometry

Looking at early word segmentation and mapping through pupillometry

Maria Teixidó and Laura Bosch

Universitat de Barcelona

Previous research has shown that 6-month-olds can use prosody to simultaneously extract one word from an artificial language and map it onto a referent (Shukla, White and Aslin, 2011). To further explore this ability using natural speech, 6- and 9-month-old infants were tested with an audiovisual segmentation and mapping task, in which objects moved aligned with prosodically marked words. Visual fixation patterns and pupil dilation measures were recorded. Visual fixation measures yielded significant between-group differences (p=0.02), with only 9-month-olds succeeding at this dual task. An ANOVA using mean pupil size values to words (baseline and learning phase) as a within-group factor, and age (6 and 9 months) as a between-group factor showed a significant interaction (p = 0.01), with only 9-month-olds increasing pupil size during learning. Two additional experiments, with similar material, testing segmentation and mapping separately, confirmed that these abilities are present by 6 months of age. Increases in pupil size were only found in the mapping task, suggesting that pupil dilation might reflect object-label association processing, rather than segmentation. Taken together, results indicate that the dual ability of simultaneously segmenting and mapping two words extracted from a natural language is cognitively too challenging by 6 months of age.