Heterogeneity at the Minimally Verbal End of the Spectrum

Background: Approximately 30% of children with autism spectrum disorders (ASD) are minimally verbal past age 5 (Anderson et al., 2007), and these children are rarely reported in research. Preliminary research suggests that despite limited expressive language, there may be considerable variability in their other abilities, such as non-verbal cognition and receptive language.  Better characterization of these abilities can lead to increased understanding of the minimally verbal ASD population, more targeted interventions and better longitudinal outcomes. 

 

Objectives: 1) To describe the range of abilities across domains in a sample of minimally verbal children with ASD, 2) determine how those abilities relate to expressive language, and 3) find common skill profiles. 

Methods: Participants included 61 minimally verbal children with ASD from a multi-site intervention study (Kasari et al., 2014), ages 4.5 – 9 years.  Data was analyzed from the baseline timepoint. Assessments included the Autism Diagnostic Observation Schedule (ADOS), Peabody Picture Vocabulary Test – 4^thEdition (PPVT), Natural Language Sample (NLS; Kasari et al., 2014), Leiter-R (non-verbal IQ), Early Social Communication Scales (ESCS) and Repetitive Behaviors Scale – Revised (RBSR).  Descriptive and correlational statistics were used to describe ability ranges across domains and their relationships to expressive language.  Skill profiles were determined using K-means analysis. 

Results:   All participants were assessed with the ADOS module 1 (no or very little expressive language).  On the ADOS, 44 participants used five or fewer words, 12 used more than 5 words and 5 used short, inflexible phrases.  Severity scores ranged from 4 – 10, indicating a variety of ASD symptoms. NVIQ ranged from 36 to 123, with a mean of 68.  NVIQ was correlated with receptive language (r=.50, p<.001), but not with expressive language or any other measure.  Receptive language age-equivalent ranged from 1.9 to 4.7 years (M=2.6) and was related to expressive vocabulary size (r=.39, p=.002) but no other variables.  Joint attention (from the ESCS) was correlated with ADOS severity score (r=.36, p=.005).  K-means cluster analysis was used to identify 4 participant clusters, based on a variety of measures.  Cluster 1 (N=10) had low abilities across domains, the highest level of repetitive behaviors and highest ADOS scores.  Clusters 2 (N=5) had a mix of abilities across domains.  Cluster 3 (N=32) was the largest group and had the lowest rates of repetitive behaviors with scores near the mean on other measures.  Cluster 4 (N=11) had a moderate rate of communicative utterances, moderate repetitive behaviors and high abilities across other domains. 

 

Conclusions: In this sample of minimally verbal children with ASD, there was substantial heterogeneity across abilities.  Of particular interest, children showed great variability in NVIQ, which was not correlated with most other domains.  The skill profiles that emerged from the cluster analysis highlight the dissociations among expressive language, NVIQ, joint attention and repetitive behaviors.  These characteristics should be explored further, to see if similar skill profiles can be found in additional samples of minimally verbal children with ASD.  Particularly, research is needed to determine if these characteristics can be related to intervention response and longitudinal outcomes.