Infant-directed song potentiates infants' selective attention to adults' mouths over the first year of life

Abstract

The mechanisms by which infant-directed (ID) speech and song support language development in infancy are poorly understood, with most prior investigations focused on the auditory components of these signals. However, the visual components of ID communication are also of fundamental importance for language learning: over the first year of life, infants' visual attention to caregivers' faces during ID speech switches from a focus on the eyes to a focus on the mouth, which provides synchronous visual cues that support speech and language development. Caregivers' facial displays during ID song are highly effective for sustaining infants' attention. Here we investigate if ID song specifically enhances infants' attention to caregivers' mouths. 299 typically developing infants watched clips of female actors engaging them with ID song and speech longitudinally at six time points from 3 to 12 months of age while eye-tracking data was collected. Infants' mouth-looking significantly increased over the first year of life with a significantly greater increase during ID song versus speech. This difference was early-emerging (evident in the first 6 months of age) and sustained over the first year. Follow-up analyses indicated specific properties inherent to ID song (e.g., slower tempo, reduced rhythmic variability) in part contribute to infants' increased mouth-looking, with effects increasing with age. The exaggerated and expressive facial features that naturally accompany ID song may make it a particularly effective context for modulating infants' visual attention and supporting speech and language development in both typically developing infants and those with or at risk for communication challenges. A video abstract of this article can be viewed at https://youtu.be/SZ8xQW8h93A. RESEARCH HIGHLIGHTS: Infants' visual attention to adults' mouths during infant-directed speech has been found to support speech and language development. Infant-directed (ID) song promotes mouth-looking by infants to a greater extent than does ID speech across the first year of life. Features characteristic of ID song such as slower tempo, increased rhythmicity, increased audiovisual synchrony, and increased positive affect, all increase infants' attention to the mouth. The effects of song on infants' attention to the mouth are more prominent during the second half of the first year of life.

Keywords: eye-tracking; infant-directed song; infant-directed speech; infants; language development.

Figure 1.

Example of (A) still images and (B) corresponding regions of interest (ROIs: eyes: orange; mouth: cyan; body: fuchsia, and object: yellow) from a video clip in the current study. Our analyses focused on the proportion of face looking time spent fixating on the mouth ROI (PFLT-m: cyan ROI/(cyan ROI+ orange ROI)).

Figure 2.

Clip feature distributions for infant-directed song (orange - left) and infant-directed speech (blue - right): (A) Mean pitch; (B) Pitch variability (measured as standard deviation); (C) Tempo; (D) Rhythmic Variability; (E) Saliency of Audiovisual Synchrony (AVS) in the Mouth ROI; (F) Positive Affect. The grey squares show each feature’s average per clip type.

Figure 3.

Mouth-looking to infant-directed song and speech as a function of age. Mouth-looking is quantified as the percentage of face-looking time spent on the mouth region (PFLT-m). Note this measure is the complementary percentage of eye-looking and can be reversed and read with respect to attention to eyes (i.e., 60% mouth-looking can also be interpreted as 40% eye-looking). The individual points show average mouth-looking time per child at each age point for song (orange circles) and speech (blue triangles), and the lines show the model predictions for each clip type (song: orange solid; speech: blue dashed). The shaded regions represent 95% confidence intervals around model predictions.

Figure 4.

Model predictions of mouth-looking (PFLT-m) as a function of clip features across age groups (indexed by color and line type): Mouth-looking as a function of (A) tempo; (B) rhythmic variability; (C) saliency of audiovisual synchrony (AVS) in the mouth ROI; (D) positive affect. Slower tempo, reduced rhythmic variability, and increased mouth AVS saliency all increased mouth-looking more in older infants while greater positive affect increased mouth-looking more for younger infants. The shaded regions represent 95% confidence intervals around the model predictions.