Tuesday, June 21, 2022
The Role of Effort in Novel Word and Grammar Learning
Leah Brainin
University of Western Ontario
View the digital poster here. (Please use Safari for best results. If you are using Google Chrome, please click here to view the poster.)
Compared to children, adults typically excel at tasks involving higher-order cognitive processes, especially those involving effort and attention. However, language learning, particularly grammar, is a notable exception. We propose that adults’ more mature cognitive processes may come with costs by interfering with certain aspects of language learning. To examine this hypothesis, we used a statistical language learning task comprised of novel words presented in grammatical sequences. Effortful engagement during language exposure was manipulated through the use of two group conditions: passive listening vs. effortful learning. Standard forced-choice recognition tests may come with shortfalls pertaining to test sensitivity, influence from confounding factors, and barriers when it comes to developmental research. We therefore included reaction time measures through the use of target detection tasks to measure word learning and grammar generalization. We discuss how effort interferes with some aspects of adult language learning while facilitating others. Additionally, we discuss how reaction time can be used to assess various aspects of language learning and highlight its versatility for use with a wider variety of populations. This work is critical in developing child-friendly measures that better capture learning outcomes by reducing influence from cognitive processes that tend to contaminate standard statistical learning approaches. In turn, these measures can be used to examine whether cognitive maturation of executive function processes significantly contribute to age-related language learning differences.
Question Generation for Reading Comprehension Assessment by Modeling How and What to Ask
Bilal Ghanem
University of Alberta
View a digital copy of the poster here. (Please use Safari for best results. If you are using Google Chrome, please click here to view the poster.)
Reading is integral to everyday life, and yet learning to read is a struggle for many young learners. During lessons, teachers can use comprehension questions to increase engagement, test reading skills, and improve retention. Historically such questions were written by skilled teachers, but recently language models have been used to generate comprehension questions. However, many existing Question Generation (QG) systems focus on generating extractive questions from the text, and have no way to control the type of the generated question. In this paper, we study QG for reading comprehension where inferential questions are critical and extractive techniques cannot be used. We propose a two-step model (HTA-WTA) that takes advantage of previous datasets, and can generate questions for a specific targeted comprehension skill. We propose a new reading comprehension dataset that contains questions annotated with story-based reading comprehension skills (SBRCS), allowing for a more complete reader assessment. Across several experiments, our results show that HTA-WTA outperforms multiple strong baselines on this new dataset. We show that the HTA-WTA model tests for strong SCRS by asking deep inferential questions.
Neural entertainment of natural language in a large-scale sample of school-aged children
Christine Moreau
University of Western Ontario
View the digital poster here. (Please use Safari for best results. If you are using Google Chrome, please click here to view the poster.)
Recent work has demonstrated that statistical learning can be indexed online using electroencephalography (EEG)-based neural entrainment, which is the alignment of neural oscillations with sensory stimuli. In this study we examined how entrainment to natural speech is associated with language and reading skills. We focused on frequency bands associated with processing syllabic and phonemic information (theta, alpha, and beta) to determine whether weaker neural entrainment at these frequencies could be linked to deficits in language and reading abilities. We used EEG data from the Child Mind Institute’s Healthy Brain Network from 713 children (5-18 years, M = 10.17 years) passively viewing and listening to a 2.72-minute educational video. Cerebro-acoustic phase coherence was calculated for each child to quantify how well their neural oscillations align to language information. Multiple regressions were conducted to determine whether entrainment predicted scores on standardized language and reading assessments. Results demonstrate that entrainment at the syllabic and phonemic rates predicted phonemic decoding, receptive and expressive vocabulary, and verbal comprehension abilities. Together, these findings highlight the important role neural entrainment plays in language and reading and could further our understanding of how entrainment is impacted in children with language and reading disorders.
Statistical Learning in relation to ASD and ADHD traits: Further evidence for a spectrum of impairment
Kaitlyn Parks
University of Western Ontario
View the digital poster here. (Please use Safari for best results. If you are using Google Chrome, please click here to view the poster.)
Links between statistical learning (SL) and language have primarily been explored in children with language or reading disorders. Little is known about how SL relates to other prevalent disorders such as autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorder (ADHD), where language and communication are also impacted. SL was measured by exposing participants (N=95) to an artificial language stream consisting of six nonsense words. Participants then completed a two-alternative forced-choice test to assess whether they could identify the words. Traits associated with ASD and ADHD were measured using the Autism Quotient (AQ), Broad Autism Phenotype Questionnaire (BAPQ), and Adult ADHD Self-Report Scale. Poor performance on the SL task was significantly related to increased ASD, but not overall ADHD traits. Further, neither inattention nor hyperactivity characteristics of ADHD significantly related to SL. Our results indicate that those with higher ASD traits demonstrate increased SL difficulties. Deficits in SL may therefore contribute to the language and social challenges commonly reported in ASD, while other mechanisms may underlie these challenges in ADHD. Our findings highlight the importance of examining a range of traits associated with ASD, ADHD, and other disorders to gain a better understanding of how specific symptomatology is linked to SL.
Use of Diverse Children’s Literature by Speech-Language Pathologists
Carlos Perez Valle
McGill University
View the digital poster here. (Please use Safari for best results. If you are using Google Chrome, please click here to view the poster.)
Overview: Books are popular tools used by speech-language pathologists (SLPs) with pediatric clients. ‘Diverse children’s books’ authentically highlight diverse experiences, identities, and realities for young readers and may improve services for diverse clinical populations. Despite the demand for such books, the lack of diversity among children’s literature is a historical barrier to using them in practice. As such, this on-going study provides a summary of demographic moderators, facilitators, and barriers which influence the use of diverse children’s literature in pediatric services across Canada.
Method: Canadian pediatric SLPs treating clients in pre-school through age 13 were surveyed about their knowledge and use of diverse children’s books along with the ethno-racial and linguistic diversity of their caseload. Currently, we have 35 complete responses from SLPs across 7 Canadian provinces and territories. Participants listed up to 10 books routinely used in their practice. Responses were coded for the ethno-racial diversity of the book’s protagonist. Participants also provided perceived facilitators or barriers to using diverse children’s books.
Results: Participants reported having somewhat or very diverse ethno-racial and/or linguistic caseloads. 22 of 35 SLPs noted not having enough tools, resources, or professional learning for servicing diverse clients. These perceptions are reflected among books frequently used.
Word Semantic Representations for Infants
Rohan Saha
University of Alberta
View a digital copy of the poster here. (Please use Safari for best results. If you are using Google Chrome, please click here to view the poster.)
Infants start developing rudimentary language skills and start understanding simple words before their first birthday. While existing works use Event Related Potentials to validate the presence of word semantic representations in infants, the mental processes manifesting these semantic representations is unclear. Accordingly, we use a decoding approach where we use machine learning techniques on Electroencephalography (EEG) data to predict the semantic representations of words found in the brain activity of infants (9-month and 12-month old). We observed signs of word comprehension immediately after word onset, marked by significantly above chance prediction accuracy of word semantics. We also observed strong neural representations of word phonetics for both age groups, some possibly correlated to word decoding accuracy and others not. Lastly, we discovered similarity of neural representations of word semantics in both infant age groups. Our results give us insights into the evolution of neural representations of word meaning in infants.
Literacy learning in a digital world: Considering the roles of sex and executive function in learning differences
Dahlia Thompson, Tanya Matthews, Susan Rvachew
McGill University
View a digital copy of the poster here. (Please use Safari for best results. If you are using Google Chrome, please click here to view the poster.)
In the present study, we examined relations among sex, two components of executive function (inhibitory control and working memory), and young children’s learning from digital stories in a remote learning context; ubiquitous even among preschool children. Specifically, we hypothesized that executive functions would predict children’s ability to retell the story, understand the story, and recognize story words. As a secondary hypothesis, we predicted that boys might have weaker executive functions, thereby leading to worse learning from shared reading online. Twenty-five English-speaking children (10 boys, 15 girls) between the ages of four and five participated in this online study. Children completed intake measures of early literacy and executive function skills, shared a digital story, then received post-read assessments of oral language and decoding. Independent Samples t Test and Analysis of Variance did not reveal any statistically significant differences between boys and girls in their intake executive function skills, and boys did not have significantly worse learning outcomes than girls. The results of Hierarchical Linear Regression indicate that executive function skills significantly predicted young children’s learning outcomes, above and beyond what was accounted for by only intake pre-literacy skills. Taken together, the findings suggest that learning differences at this stage of development are more likely associated with differences in executive functions than child sex. These findings prompt us to consider what executive function strategies we can embed within literacy learning contexts such as online learning to support boys’ and girls’ learning.
Keywords: online learning; shared reading; digital books; early literacy; sex; executive function skills; oral story retelling; listening comprehension; word recognition
Conflict of interest statement
The authors report no conflicts of interest. This study was carried out in accordance with the McGill University Institutional Review Board guidelines. Parental informed consent and child assent were obtained from study participants prior to participation.
Disclosure
The Social Sciences and Humanities Research Council (SSHRC) of Canada provided financial support for this study, through a research grant awarded to the third author, Susan Rvachew. Funding Code: G256437, SSHRC 435-2021–0591
Representational Similarity Analysis of the Neural Representations of Orthographic-Phonologic and Orthographic-Semantic Mappings Across Development
Deanne Tak On Wah
University of Western Ontario
View the digital poster here. (Please use Safari for best results. If you are using Google Chrome, please click here to view the poster.)
Previous studies show that children recruit different brain regions and show less activation of key brain regions compared to adults when doing rhyming and lexical decision tasks. These differences are thought to reflect the development of efficient orthographic-phonologic processing. Here, I propose an fMRI study to determine brain regions recruited while processing orthography, phonology, and semantics, as well as the underlying mappings in the direction of orthography to phonology and orthography to semantics. Children and adults will silently read words in an fMRI scanner, while monitoring proper names using a button press. Analyses will be conducted using a multivoxel pattern analysis-representational similarity analysis approach to examine how different aspects of visual word representation are reflected in patterns of brain activity. It is hypothesized that in adults, orthographic information will be represented in the left ventral occipital temporal cortex and angular gyrus, while phonological information will be represented in both the left and right inferior frontal gyrus. Semantic information will be represented in the medial temporal gyrus and angular gyrus. However, children will show reduced strength in the similarity of the patterns of activation for the orthographic and phonologic models compared with adults. They may also show greater strength in the similarity of patterns of activation for the semantic model, suggesting that early readers rely more on semantic information and less on orthographic and phonologic information than skilled readers. This study will uncover the patterns of activation involved in specific reading processes, particularly how the strength of these representations may influence reading proficiency across development.
Current Projects: The Multilingualism and Literacy Lab
Diana Burchell*, Krystina Raymond*, Zein Abuosbeh*, Songtao Wang*, Michelle Chin*, Dianne Macdonald* & Xi Chen
University of Toronto, OISE
View the digital poster here. (Please use Safari for best results. If you are using Google Chrome, please click here to view the poster.)
The Multilingualism and Literacy Lab is a developmental psycholinguistic lab, led by Dr. Xi Chen, which focuses on the development of languages in multilingual children, both within and across languages. This poster will overview the purpose, methodology, results and implications of all the MLL current projects. These projects include:
- The International Bilingual Education (IBE) project: Evaluating the efficacy and accuracy of phonological and orthographic specificity as longitudinal predictors of reading ability in emergent English-French bilingual children.
- The IBE stakeholder study: Interviewing parents, teachers, and administrators on the accessibility of French Immersion for diverse students.
- The Digital Reading project: Evaluating the effectiveness of passive versus active digital scaffolds embedded in online text in supporting French word learning and reading comprehension among a sample of French immersion (FI) students.
- The Online Learning project: Investigating the impact of online schooling during the pandemic on language and reading development in English and French among elementary French immersion (FI) students.
- The Componential Model of Reading study: Examining the contribution of ecological factors to English word reading among Syrian refugee children.
- The Writing Modality study: Comparing the effects of transcription mode on the quality of written French narratives produced by French L2 learners.
- The Reading Impairment project: Investigating the impact of reading in the home language (English), for school-aged children with reading impairment who attend French school, on a battery of English and French language and literacy measures.
