About this srcd poster session

---

Panel information
Panel 5. Developmental Disabilities

Abstract

---

Every object label applies to a variety of entities in the world. For instance, the word spoon can refer to metal spoons, plastic-colored spoons, serving spoons or toy spoons. By the time non-autistic (N-AS) infants have learned 100-200 count nouns (≈24 months), they recognize that words such as spoon indicate a characteristic shape (Baldwin, 1992; Imai et al., 1994; Landau et al., 1988; Smith et al., 2002). Importantly, this “shape bias” is only helpful for extending count nouns: Mass nouns (e.g., mud, water) do not have characteristic shapes and neither do early learned adjectives (e.g., red, big), abstract nouns (e.g., idea, plan), verbs (love, go). It is appropriate then, that N-AS children do not generally prioritize shape when extending mass nouns (Li et al., 2009), animates (Booth et al., 2005), nor when judging general similarity or “sameness” (Landau et al., 1988; Smith et al., 2002).
Autistic children tend to be delayed in language learning generally and are delayed in learning this shape bias as well, although prior results are mixed on the extent of the delay (Field et al., 2016; Hartley et al., 2014; Hartley et al., 2019; Potrzeba et al., 2015; Tek et al., 2008; Tovar et al., 2020). Here autistic children aged 3-5 (N = 109, from SPARK) were tested on two tasks before an intervention designed to help them learn to prioritize shape when extending novel object labels (Smith et al., 2002). Parents also filled out the MB-CDI, a measure of productive vocabulary (Mvocab = 457, SD = 190) and the SCQ, a screening tool for autism referral (MSCQ = 26, SD = 5.5).
Autistic children were asked to extend novel Object Labels to objects matching in shape, color, or texture. Several days later, participants were asked to extend novel Adjectives to shape matches or color/texture matches. After exclusions for failure to understand or complete the tasks (n = 20), data from 89 autistic children (Mage = 56.7, [36-70]) were analyzed. On the Object Label task, 93% children appropriately prioritized shape at above chance levels. On the novel Adjective task, 68% appropriately prioritized color and texture rather than shape (Fig. 1).
All children with lower than 60% accuracy on the novel object task were invited to take part in an eight-week online intervention, inspired by Smith et al. (2002). They watched brief videos each week that illustrated novel object labels being extended by shape rather than color or texture. Only nine children completed the intervention, but despite the small sample, accuracy on extending new labels for new novel objects improved significantly (Fig. 2: Meanincrease = .31; 95% CI[.14, .48]; t(8) = 4.12, p = 0.003), and without any inappropriate overgeneralization of shape to new novel adjectives (Meanincrease = .03). No evidence of increased vocabulary was evident after the intervention in the small sample: 95% CI[-59, 27]. To summarize, the majority of autistic children may learn to prioritize shape appropriately on their own, and others can be helped to do so.

Author information

Author	Role
Arielle Belluck, Princeton University	Presenting author
Julia Nguyen, Princeton University	Non-presenting author
Momna Ahmed, Princeton University	Non-presenting author
Adele Goldberg, Princeton University	Non-presenting author

⇦ Back to session