Assessment of Cognitive and Language Abilities in Toddlers with and without Autism Spectrum Disorders: Comparison of the Mullen Scales of Early Learning and the Bayley Scales of Infant and Toddler Development, 3rd Edition

Background: Valid and reliable assessment of cognitive and language development in young children is a key component of diagnostic evaluations for autism spectrum disorders (ASD), developmental delay, and language disorders. The Mullen Scales of Early Learning and Bayley Scales of Infant and Toddler Development, 3^rdedition are the most commonly used standardized measures to assess early cognitive development because they are normed from infancy through the preschool years. While the Bayley Scales are frequently used in clinical settings, the Mullen Scales are more widely used in research studies, particularly with children with or at risk for ASD.  A study presented at IMFAR 2008 (Esler et al, 2008) found strong correlations between the measures for nonverbal mental age but standard scores were not correlated. These findings need to be replicated in a more heterogeneous sample in order to better understand the implications of translating research findings to the community. In addition, Esler and colleagues (2008) only looked at the nonverbal domains; as such it is important to also understand the comparability of the language domains.

Objectives: Examine convergent validity of the nonverbal cognitive, receptive language, and expressive language subscales on the Mullen and Bayley Scales in toddlers with and without ASD.

Methods: The Mullen and Bayley scales were administered to participants from 12-42 months of age who were seen for a research or clinical assessment due to a diagnostic question of ASD, as well as typically developing (TD) controls. Data has been collected on 11 individuals to date, with expectations of 50 participants by April 2014. ASD diagnosis was determined through a full evaluation including cognitive and diagnostic measures (i.e., Autism Diagnostic Observation Schedule (ADOS)), as well as behavioral observations and parent interviews.

Results: Toddlers with ASD had lower standard scores on both the Mullen and Bayley scales than TD controls. Preliminary results across the sample indicate strong correlations for standard scores and age equivalence scores for the Mullen and Bayley nonverbal cognitive, receptive language, and expressive language subscales (r's>0.8, p's<=0.001). However, for some TD and ASD participants, scores on any particular subscale differed by more than one standard deviation, though performance was not consistently greater for one test versus the other.

Conclusions: Preliminary findings suggest, unlike the Esler et al (2009) findings, good concordance between the Mullen and Bayley nonverbal cognitive, receptive language, and expressive language scales when examining a sample that includes toddlers both with and without ASD. Indeed, correlations between the Mullen and Bayley scales were similar to the test-retest correlations for each of the individual measures. Differences in participants' level of functioning between the current and Esler studies may explain this variability. Analyses based on specific diagnostic group are ongoing. Research studies often examine cognitive or language performance on the Mullen as a predictor or covariate in paradigms of behavior or brain function in ASD versus TD individuals. Results suggest that these findings in research settings may translate to toddlers seen through community practices that rely on the Bayley scales.