Growth in Narrative Retelling Abilities of Higher-Functioning Children with ASD: Associations with ASD Symptomatology, Verbal Ability, and Reading Comprehension

Background: Narrative retelling, or story memory (SM), relies on structural and pragmatic language, social cognition, and global/gist processing skills, and has been shown to pose difficulties for children with ASD (Diehl et al.,2006). These language and cognitive skills influence reading development, especially the growth of text comprehension ability (van den Broek et al., 2003). The development of reading comprehension (RC) is a significant challenge for many children with ASD (Nation et al., 2006; Ricketts et al., 2013). These challenges are so pronounced that RC impairment may be an important part of the social communication phenotype of ASD in higher-functioning children (McIntyre et al., 2016; Ricketts et al., 2013). This study was designed to examine the relations between SM, RC, language, and the social communication phenotype of school-aged children with higher-functioning ASD (HFASD).

Objectives:  1) To examine the development of SM over a 30-month period in school-aged children with HFASD using latent growth curve modeling, 2) To test the hypothesis that autism symptom severity and verbal abilities are significantly related to SM development among these children, and 3) To examine the hypothesis that SM development predicts the development of HFASD reading comprehension.

Methods:  Participants included seventy-eight 8- to 18-year-old children with HFASD (Full-scale IQ ≥ 75). Data were collected at 3 time points across 30 months. ASD symptoms were confirmed with the ADOS-2, which provided symptom severity data. Verbal ability (VIQ) was measured by WASI-2. Narrative retelling was measured with the WRAML2 Story Memory task (SM). Reading comprehension (RC) was measured with the Gray Oral Reading Test-5 (GORT-5).

Results:  The final growth model fit the data well, χ² = 8.95 (6 df), p=0.18, CFI=0.98, RMSEA=0.08. 1) Participants varied significantly on SM intercept at T1 (SM1; M=24.99, p=0.009), but individual growth curves did not differ significantly on slope (4.2; p=0.38). 2) ADOS and VIQ both significantly related to the intercept (-1.11, p=.002;0.45, p<.001, respectively) but not slope (-0.07, p=0.77;-0.01, p=0.91, respectively) of SM growth curves. For every 1-unit increase in ADOS scores, a decrease of 1.11 units was seen in SM1; for every 1-unit increase in VIQ, an increase of 0.45 units was seen in SM1. Age was included in the model, but was not significant and was dropped as a result. 3) RC was significantly associated with the intercept (0.90, p<.001) but not the slope (-0.69, p=0.61) of SM growth curves. For every 1-unit increase in SM1, a 0.90 unit increase was seen in RC3 scores.

Conclusions:  HFASD improved in their narrative retelling skills with similar 30-month growth rates regardless of initial scores. Lower ASD symptomatology and higher VIQ, but not age, were significantly associated with better SM1 performance, but did not impact growth rates. SM1 was a significant predictor of RC3, but there was no relation between the rate of SM growth and RC. These data provided unique details about the social communication development of school-aged children with HFASD, and raise the possibility that narrative retelling may be an important intervention target to improve reading comprehension in these children.