Sensory Subtypes in Children with ASD: Latent Profile Analysis Using a National Survey of Sensory Features

Background:  Sensory features are highly prevalent in children with ASD and have been suggested to have negative consequences in the daily life of children and families.  Lacking in the literature is the identification of homogeneous sensory phenotypes that can be used to analyze the relationship of such subtypes with functional outcomes as well as inform precise diagnostic instruments and targeted treatment strategies.

Objectives: This study describes the methodology of characterizing sensory features in children with ASD using latent profile analysis (LPA) to create sensory subtypes and presents the association of the subtypes to child characteristics such as autism severity and mental age.

Methods:  Data were collected as part of a national online survey from 1307 participants with ASD, ages 2-12 years.  Sample consisted of 1068 boys (CA 93 (34) mos.) and 239 girls (CA 96 (35) mos.).  ASD symptom severity was assessed using the Social Responsiveness Scale.  The Sensory Experience Questionnaire 3.0 (SEQ) was used in a confirmatory factor analysis (CFA), which yielded four factors of sensory response patterns (i.e., hyporesponsiveness, hyperresponsiveness, seeking, and enhanced perception).  Factor scores from the CFA were exported for analysis of latent profiles. LPA was used to detect the presence of distinct groups in the data set. Model fit was assessed with BIC and AIC as well as the Lo-Mendell-Rubin test.  Mixed models, allowing for nesting of observation within family, were used to explore the association of autism severity and mental age to the subtypes.

Results:  Four distinct profiles (sensory subtypes) emerged as supported by model fit indices with increasingly small changes in both AIC and BIC values after the inclusion of the fourth profile. Further, the Lo-Mendell-Rubin test of change in the likelihood ratio was significant only up to the fourth profile. The first subtype (n=402, 31%) describes children who scored low on all sensory patterns, while those in the second subtype (n=288, 22%) showed exactly the opposite profile with the high scores in all four sensory patterns. The remaining two subtypes showed a split in their factor scores. The third subtype (n=404, 31%) scored close to the mean on all patterns, with some tendency to score low on seeking and hyporesponsiveness, but slightly above on hyperresponsiveness and enhanced perception. The fourth subtype (n=213, 16%) had the opposite pattern of the third subtype with scores tending to be more extreme on seeking and hyporesponsiveness.  Autism severity was significantly related the subtypes while controlling for mental age, with subtypes two and four expressing the highest levels of autism severity, followed by the third subtype, and finally, the first subtype that expressed the least amount of autism severity.

Conclusions:  The LPA further explores the co-existence of different sensory patterns in children with ASD.  The identification of homogenous subtypes will allow for characterization of children within the subtype to functional child outcomes leading to improved assessment and treatment as well as potential identification of biological markers. Further analyses will determine the extent to which there are significant associations between the identified sensory subtypes with various child characteristics.