Sensory Features In Nonverbal Children with Autism

Background:  Roughly 30% of children with autism remain nonverbal into their school years (Lord et al., 2006).  This population of children is poorly understood and no published studies have directly investigated severity or patterns of sensory features in this subgroup.  In the general autism population, sensory phenotypes can be characterized by three patterns: hyper and hyporesponsiveness and sensory seeking (Baranek et al., 2006; Liss et al., 2006), with hyporesponsiveness being more characteristic of autism compared to other developmental disabilities.  Hyporesponsiveness and seeking are consistently associated with more impaired social-communication; evidence for hyporesponsiveness is mixed. This study aims to characterize the type and severity of sensory patterns in nonverbal ASD and their predictive utility in a cross-sectional as well as longitudinal sample.

Objectives:  To determine a) sensory features of nonverbal (fewer than five words; ADOS module 1 item A1.) children with autism, and b) indicators of change in nonverbal status over time.

Methods:  Extant data from two grants including parent and observational sensory measures were analyzed.  A total of 72 participants provided cross-sectional data (27 nonverbal CA 42.0 (14.5) mos.; 45 verbal, CA 57.6 (16.2) mos.) and a subset of 15 of these participants provided longitudinal data (7 nonverbal, CA 34.8 (4.5) mos.; 8 verbal, CA 35.4 (4.1) mos.).  Factor scores for three sensory patterns (hyper, hypo, seeking), were generated using structural equation modeling and were used to determine group (verbal / nonverbal) differences and change in verbal status over time.  Also, groups were analyzed for differences based on cognition, adaptive skills, and demographics.

Results:  Cross-sectional analyses indicated that nonverbal children have significantly more hyporesponsive (p=.019) and sensory seeking behaviors (p=.001).  Demographic data revealed significant group differences with nonverbal children having mothers with lower education (p=.029) and lower household incomes (p=.016). IQ (proxy) (nonverbal m=34.6, sd=12.7; verbal m=72.7, sd=26.3) was also significantly lower (p=.001).  Longitudinal analysis indicated children who remained nonverbal at time 2 were significantly more hyporesponsive (t=3.38; p=.005) and had significantly lower mental ages (p<.05) and poorer adaptive skills (p<.05) at time 1 compared to children who were verbal at time 2.  Group differences between hyperresponsiveness and seeking patterns were not significant. 

Conclusions:  Both cross-sectional and longitudinal data support higher levels of hyporesponsiveness, and cross-sectional data support sensory seeking behaviors being associated with nonverbal status in children with autism. This is consistent with findings regarding social-communication skills and hyporesponsiveness and sensory seeking patterns in the general autism population. Hyporesponsiveness could indicate the child is not participating in early learning opportunities and may indicate decreased attention to environmental stimuli overall.  Sensory seeking may yield overfocused attention to idiosyncratic interests to the exclusion of social and communicative opportunities.  An interesting finding was that nonverbal status is also associated with lower household income and lower maternal education.  These variables associated with lower SES may yield fewer intervention resources and a less rich language environment.  These findings have implications for assessment and intervention.