Parent tuning to children’s language processing speed

Angela Xiaoxue He, Rhiannon J Luyster, Sudha Arunachalam

Research output: Contribution to conferenceConference paperpeer-review

Abstract

Parents’ language input is tuned to children’s productive language abilities (e.g., Bornstein et al., 2007), but less is known about whether it is tuned to their language comprehension. This is important when considering children who show an unusually large gap between these skills. Children with autism spectrum disorder (ASD) show such a gap; their language comprehension lags farther behind age level than their productive language (e.g., Luyster, Lopez, & Lord, 2007). Thus, this population presents an opportunity to assess parents’ sensitivity to their child’s comprehension abilities.

Preschoolers with and without ASD (ages 2 to 6 years, N = 35) participated with one parent. On each trial, parents labeled one of six images while the child’s gaze was recorded using an eyetracker. The experimenter indicated to the parent which image was the target using its position in the array, so that the parent was not told what to call the target object, only which position it was in. There were two conditions. In the Difficult condition, the target had a competitor from the same basic-level category (e.g., target: open book, competitor: closed book) as well as four unrelated distractor objects (Figure 1). In the Easy condition, there were no competitors, only five unrelated distractors. In the Difficult condition, parents would need a more complex expression, such as a modified noun, to uniquely label the referent than in the Easy condition.

To assess how parent language might be tuned to support children’s comprehension, we analyzed features of the parent’s language, children’s latency to look to the target, and whether these patterns differed for ASD and TD children. The results indicated no differences between ASD and TD children in parent speech rate (TD: 0.33 syllables/second, ASD: 0.39; ß = -0.066, p = .27) or frequency of modifier use (TD: 63% overmodification, ASD: 54%; Fisher’s exact test p = 0.21).

However, there were differences in children’s latency to look to the target (Figure 2). Children with ASD were slower (mean latency = 1327 ms) than TD children (796 ms). A mixed-effects model with latency as the dependent measure and group (TD vs. ASD) and child age as predictors yielded a main effect of group (ß = 801.9, p < .05). When adding condition (Difficult vs. Easy) to ask if the group difference held over and above the difference between conditions, we found a main effect of condition (ß = -825.1, p = .048) and the group difference was no longer significant (but p = 0.052, ß = -865.7): children were slower in the Difficult condition (TD: 1193 ms, ASD: 1743 ms) than the Easy condition (TD: 452 ms, ASD: 819 ms).

In sum, children with ASD were slower to comprehend their parent’s language than TD children, but parents of both groups produced similar language. We suggest that parent language is not specifically tuned to processing speed. With TD children, this may not be detectable because language comprehension and production track each other predictably; ASD presents an opportunity to examine parents’ sensitivity to each separately.

Fingerprint

Dive into the research topics of 'Parent tuning to children’s language processing speed'. Together they form a unique fingerprint.

Cite this