Are Sensory Features a Subdomain of Restricted and Repetitive Behaviors? Evaluating Empirical Support for the DSM-5...

Following calls to formally recognize sensory differences in the autism diagnostic criteria, the DSM-5 introduced sensory features as a subdomain of restricted and repetitive behaviors (RRBs). However, this categorization was developed based on expert consensus, rather than being empirically derived. Subsequent psychometric investigations have largely examined the structure of sensory features and RRBs separately, precluding evaluation of whether these constructs are best represented by shared or distinct latent dimensions.

The current study used cross-measure factor analysis with items from across the Dimensional Assessment of Repetitive Behaviors (DARB) and the Sensory Experiences Questionnaire Version 3 (SEQ-3) to expand coverage of sensory features and to compare the DSM-5 RRB structure with alternative empirically derived structures identified in the literature. The best-fitting first-order model was then operationalized to compare different general factor solutions.

Across first-order models, the 4-factor DSM-aligned model showed the poorest fit, whereas a 10-factor empirically derived model showed the best fit. Across bifactor models, the unidimensional general RRB/sensory model showed the poorest fit. Two alternative 2-bifactor structures showed superior fit. Among those bifactor models, an exploratory model-whereby Factor 1 comprised insistence on sameness, restricted interests, unusual interests, repetitive language, obsessive-compulsive behavior, self-injurious behaviors, and sensory hypersensitivity, and Factor 2 comprised repetitive motor behavior, sensory seeking, and sensory hyposensitivity-showed the highest internal consistency and explained the most variance in related clinical variables.

Findings suggest the need to further revise the RRB DSM-5 criteria, indicating that a unidimensional general RRB/sensory domain may not be empirically valid.