Speakers' abstracts

Professor Michael Thomas - Birbeck, University of London

Title: Computational modelling approaches to understanding the causes of language delay and opportunities for intervention

In this talk, I will present computational modelling work that explores possible causes of developmental language delay. Computational models present a double-edged sword, encouraging theoretical clarification and generating testable predictions, but at the expense of simplification. The model I discuss targets the acquisition of inflectional morphology and considers the distinction between early language delay that persists and requires intervention, and delay that resolves to the normal range by mid-childhood. The model incorporates a polygenic approach to the causes of delay, and considers language development in the context of variable levels of language input (for example, as associated with differences in socioeconomic status). The model: (1) shows how in a population of learners, apparent subgroups can emerge from continuously varying underlying causes; (2) tests the extent to which developmental outcomes can be predicted from early developmental markers; and (3) considers whether intervention outcomes can be predicted from model properties.

Erica Ellis, Ph.D. CCC-SLP - Department of Communication Disorders - Cal State University, Los Angeles

Title: Patterns of Word Learning in Late Talking Toddlers

The course of typical vocabulary development varies greatly between children. This variability makes it challenging to tease apart typical versus delayed language development during the toddler years. Longitudinal studies of language outcomes in Late Talkers (LTs) have only had moderate success in accurately predicting which LTs are likely to be identified with developmental language disorder (DLD). Our work considers how to best identify which LTs will continue to have language deficits at school age, in order to maximize the opportunity for appropriate early intervention and support during the toddler and preschool years. We examine variables during early lexical development to help predict language and academic outcomes in LT toddlers. Using eye tracking methodology, we explore several lexical and cognitive processing skills in young children that may underlie these ongoing differences in early word acquisition. In this talk, various patterns of word learning performance among LT and typically developing toddlers will be presented. Clinical implications will be discussed.

Professor Mary Alt - The University of Arizona

Title: The Vocabulary Acquisition and Usage for Late Talkers (VAULT) Treatment Approach

The Vocabulary Acquisition and Usage for Late Talkers (VAULT) treatment approach is a new combination of some classic ideas that is showing promise as an option for intervention. This talk will explain the principles on which VAULT is built (i.e., statistical learning, focused stimulation), the ways in which the approach can be implemented, and review the evidence for the treatment. We will highlight characteristics of late talkers that appear to predict who will respond to VAULT and discuss future directions for the treatment.

Michelle MacRoy-Higgins PhD CCC-SLP, TSHH - Hunter College, City University of New York (CUNY)

Title: Attention Allocation During Word Learning and Free Play in Children Who Are Late Talkers

The cause of slow expressive language development in children identified as late talkers is unknown. It is possible that attention skills may play a role in early vocabulary development in this population. This presentation will report research findings related to attention allocation during word learning tasks and free play in a group of late talkers at 2;0 years and 3;0 years, as compared with age-matched peers. Clinical implications will be discussed.

Dr. Lynn K. Perry - University of Miami

Title: What structural differences in late talkers’ early vocabulary knowledge can tell us about long-term outcomes

Late talkers are heterogeneous in their developmental trajectories. Some can be considered late bloomers and eventually “catch-up”, but others have persisting delay or are later diagnosed with developmental language disorder (DLD). Early in development it remains unclear which children will belong to which group. Here, I use vocabulary structure as a tool for exploring long-term outcomes amongst late talkers. Most children with typical development (TD) have vocabularies dominated by names for categories organized by similarity in shape (e.g., cup), and show a bias to attend to shape when they generalize the names of novel nouns—a bias associated with an accelerated rate of vocabulary development. However, as a group, late talkers tend to know fewer names for categories organized by shape and are less likely to show a “shape bias” than TD children. Thus, in a retrospective analysis of 850 children, I examine specific differences in vocabulary structure at 1.5 years, comparing late talkers had persisting delay and those who were late bloomers. At 1.5 years, late talkers with persisting delay knew a smaller proportion of shape-based nouns than both TD children and late bloomers who “caught up” to more typically sized vocabularies. Additionally, children who went on to receive a DLD diagnosis between 4 and 7 years knew a smaller proportion of shape-based nouns at 1.5 years than TD children or children who received other diagnoses (e.g., dyslexia). Together, these findings bring new insight into sources of heterogeneity amongst late talkers and offer a new metric for assessing risk.

Dr. Cecilia Zuniga-Montanez - University of Leeds

Title: Can Late Talkers Learn a Shape Bias for Noun Generalisation?

Late talkers show a vocabulary delay in the absence of any conditions or disorders that may account for this delay. Contrary to typically developing children, late talkers do not seem to generalise object labels by shape. Instead, they do so based on texture or do not show a preference for any particular property. The shape bias has been considered as an important strategy for successful noun learning and generalisation in early childhood. Thus, a lack of a shape bias could be one factor contributing to late talkers’ vocabulary delay. In this talk, I will discuss an intervention study that assessed if late talkers can be taught a shape bias and the effect that this training has on vocabulary growth. Fourteen late talkers between 24 and 47 months were randomly allocated to a training group or a control group. Over seven weekly sessions, participants in the training group were taught that objects similar in shape had the same name. Participants in the control group were taught seven sets of labels of existing objects (e.g. giraffe). In the subsequent two weeks, we assessed participants´ noun generalisation strategies. Results showed that participants in the shape training group generalised labels by shape (instead of colour or texture) for known object labels, but did not show a preference when generalising novel object labels. Participants in the control group responded randomly for both types of labels. The two groups did not differ in terms of expressive or receptive noun vocabulary growth over the course of the study. Thus, an explicit shape bias training did not help late talkers to pick up on this helpful word learning strategy and did not boost vocabulary growth.

Dr. Eva Jimenéz - University of Warwick

Title: Semantic Maturation in Word Acquisition in Typical and Late Talking Toddlers

In this talk, I will explore the semantic maturation process involved in word acquisition during the comprehension phase (words understood but not produced) in typical- and late-talking toddlers. To do this, I investigate the influence of contextual diversity -- a word feature associated with the semantic enrichment of the mental representation of words --. I will present an observational study (Study 1) and three computational models motivated by that empirical data (Study 2). We conclude from Study 1 that there is an interaction between the type of talker and the type of word (verbs vs nouns), i.e., late talkers produce fewer nouns with high contextual diversity and more verbs with high contextual diversity. However, the parameter values for our best computational model in Study 2 suggest that late talkers make less use of the contextual diversity strategy to learn verbs and nouns, suggesting a weaker semantic representation of the words they know. I will discuss how the sole examination of network properties of children’s networks might be insufficient to determine the degree of children’s word-learning abilities.

Dr. Eliana Colunga - University of Colorado Boulder

Title: Using computational models to predict and alter word learning trajectories.

One of the big accomplishments of children in their first years of life is learning the language or languages of their environment, but this task does not come easily to all children. In this talk I will present work using computational models to understand, predict, and intervene in early word learning. The idea is that language is a highly structured system of similarities and relationships, and that children become more and more skilled learners by exploiting the regularities in this structure. If this is the case, we may be able to nudge developmental trajectories by shifting the structure present in a child’s language environment. I will discuss computational models that can be used to characterize and predict the early vocabulary trajectories of children with different language abilities, present evidence suggesting that it is indeed possible to make model-based target word selections that will have a greater impact on subsequent vocabulary growth, and discuss different ways of delivering this environmental “nudge”.

Dr. Nina Capone Singleton - Seton Hall University

Title: Using Theory and Data to Inform Gesture as a Word Learning Scaffold with Late Talkers.

In this talk, I present the theory and a line of research that supports the use of gesture as a word teaching scaffold. Gesture has cognitive and social underpinnings. These social and cognitive underpinnings intersect well with what we know of word learning. Over time, cognitive theory and the extant gesture development literature guided us to compare different types of gestures for their effect on word learning. In addition, we measured word learning in toddlers, preschoolers, and late talkers across several tasks to show the depth of learning. We place the results in a clinical context.