Learning with and without feedback in children with developmental language disorder

Arbel, Y., Fitzpatrick, I., & He, X. (2021). Learning with and without feedback in children with developmental language disorder. Journal of Speech, Language, and Hearing Research, 64(5), 1696–1711. https://doi.org/10.1044/2021_JSLHR-20-00499

When a child says, ‘tup’, and we reply, ‘no, that’s a cup. Say cup’, we’re providing feedback that we expect will help the child learn the correct form. This feedback is considered to provide an avenue for ‘explicit learning’, that is, the conscious effort to learn the right form. This type of learning requires the child to monitor and evaluate feedback, so it places demands on working memory and executive functioning. On the other hand, when a child says, ‘tup’, and we say, ‘yes, that’s a cup. Let’s have a drink’, we’re creating opportunities for ‘implicit learning’, that is, the unconscious learning of patterns. This type of learning without feedback does not require self-monitoring, and so does not place demands on working memory and executive functioning.

The ability to learn from feedback is also tied to certain brain regions such as the frontal cortex and basal ganglia. Prior work suggest that feedback processing may be impaired in children with developmental language disorders (DLD) due to poor working memory skills and brain abnormalities. It would follow that children with DLD may learn better if they can bypass feedback processing.

In this study, 14 typically developing (TD) children and 13 children with DLD learned new words. During the learning session, children learned the names of novel objects and EEG data was recorded. On each trial, children saw two images and heard a name. In the feedback trials, participants had to decide which object matched with the name followed by feedback. Green checkmarks indicated correct responses while red Xs indicated incorrect responses. In the no feedback trials, correct responses were highlighted with a green box around the object and participants did not have to respond. During immediate and delayed (1-week later) testing, children had to decide which of the two objects matched a name and no feedback was provided. As expected, TD children performed better than DLD children overall. Interestingly, for both groups of children, learning was better without feedback than with feedback when tested immediately and after a delay. Further, when learning from feedback, both groups of children were more likely to benefit from positive feedback (repeat a correct response) than change after negative feedback (switch to the correct response). The EEG data showed that TD children were more sensitive to negative feedback than children with DLD.

It is interesting to consider how the effect of feedback, or lack thereof in this case, fits into clinical practice. This might be one way to reduce working memory demands, that is, to prioritize the learning of new information by minimizing the need to process feedback simultaneously. The findings also suggest the importance of positive feedback. More broadly, the findings regarding feedback were surprising, and further replication and study of feedback in DLD is important.

Blogger: Theresa Pham is a student in the combined SLP MClSc/PhD program, supervised by Dr. Lisa Archibald

Reducing low-value practices: a functional-contextual consideration to aid in de-implementation efforts

Farmer, R.L., Zaheer, I., Duhon, G.J., & Ghazal, S. (2020). Reducing low-value practices: a functional-contextual consideration to aid in de-implementation efforts. Canadian Journal of School Psychology. https://doi.org/10.1177/0829573520974915

Low-value practices are practices that are either unproven or discredited by evidence. Low-value practices sometimes persist in practice, and may interfere with implementation efforts for competing strategies. The de-implementation of low-values practices may be a necessary step in an implementation science approach to the adoption of evidence-based practices. Steps in de-implementation include identifying a low-value practice to be reduced, evaluating variables that may affect the de-implementation process, determining and using strategies that should reduce or eliminate the practice, and evaluation of the effects of de-implementation. Strategies for de-implementation could include policy change at the system level or ‘unlearning’ at the individual.

The authors of this paper consider how applied behaviour analysis could inform de-implementation due to the focus on functional relationships between context and future behaviour (i.e., functional-contextual lens). Briefly, the authors explain that the likelihood of a behaviour can be altered by changing the reinforcement, salience of contextual cues, or effort to engage in the behaviour. Through this lens, the authors suggest that low-value practices viewed as behaviours have little consequences and are rather rule governed behaviours. Behaviour reduction strategies are described including extinction through removal of reinforcement, differential reinforcement where extinction is paired with targeting of alternate behaviour, increasing response effort by requiring approvals or extra paperwork to use a behaviour, and punishment involving the addition or removal of stimulus when engaged in behaviour. The authors argue that these behavioural reduction strategies have the potential to assist in de-implementation of low-value practices but stress the importance of pairing this goal with efforts to implement evidence-based practices. 

The focus on the persistent use of low-value practices as a barrier to implementation of evidence-based practice is interesting. When implementing a new evidence-based practice, clinicians would do well to consider how it fits in with existing practice and any need for change.

Blogger: Lisa Archibald

Relearn faster and retain longer: Along with practice, sleep makes perfect

Mazza, S., Gerbier, E., Gustin, M., Kasikci, Z., Koenig, O., Toppino, T., & Magnin, M. (2016). Relearn faster and retain longer: Along with practice, sleep makes perfect. Psychological Science, 27(10), 1321–1330. https://doi.org/10.1177/0956797616659930

Two of the most effective study techniques are spaced learning and sleep. Spaced learning (or distributed practice) is when learning is spaced out over multiple sessions rather than presented in one long session. Sleep, on the other hand, has many learning benefits. After sleep, newly learned information becomes more stable thereby enhancing learning. The goal of the current study was to determine if combining both strategies could lead to better learning.

The study had 3 parts, the studying session, relearning session, and delayed testing (1 week and 6 months later). In the studying session, participants read 16 Swahili-French word pairs (nyanya-tomate) and then studied each word (given nyanya-____, what is the French translation?). For words they recalled incorrectly, the correct translation was shown, and participants practiced until they got all 16 correct. The relearning session happened 12 hours later and proceeded in a similar way. Delayed testing occurred 1 week and 6 months later. Participants were divided into 3 groups: (i) sleep group: studied at 9 pm, slept, and then relearned at 9 am the next day; (ii) no sleep group: studied at 9 am and relearned at 9 pm on the same day; and, (iii) control group: studied at 9 pm, slept, and then completed a recall-only session at 9 am the next day. Findings revealed that sleeping after learning (sleep and control groups) led to better retention the next day than not sleeping (no sleep group). Further, the sleep group required fewer trials to successfully recall all 16 pairs correctly than the no sleep group (i.e., relearning happened faster). Strikingly, the benefits of sleep and relearning were amplified 1-week later and were maintained 6-months later: only the sleep group remembered significantly more word pairs than both the no sleep and control group.

The results suggest that sleeping after learning is a good strategy but additional learning after sleep is especially beneficial to long-term memory. It would follow from these findings that it would be beneficial to children’s learning if a clinician could introduce new learning in therapy one day and then incorporated a re-learning phase (perhaps with home practice materials) after sleep the next day.

Blogger: Theresa Pham is a student in the combined SLP MClSc/PhD program, supervised by Dr. Lisa Archibald

Semantic effects in sentence recall: The contribution of immediate vs delayed recall in language assessment

Polišenská, K., Chiat, S., Comer, A., & McKenzie, K. (2014). Semantic effects in sentence recall: The contribution of immediate vs delayed recall in language assessment. Journal of Communication Disorders, 52, 65–77. https://doi.org/10.1016/j.jcomdis.2014.08.002

Sentence recall, the immediate repetition of spoken sentences, is used clinically to assess language skills. This seemingly easy task draws on different types of language and memory skills. For instance, phonological (speech sound) skills support the remembering of the word form itself. Semantic skills (knowledge of meaning) aid in remembering the meaning of the sentence. Typically, sentence repetition is tested by having the person repeat the sentence immediately after hearing it (immediate recall). The ability to recall a sentence after a delay (delayed recall) is usually not tested. The goal of this study was to understand the skills involved in immediate and delayed sentence recall.

Across two studies, participants completed immediate and delayed sentence recall. In study 1 involving adult participants, a distracting task was completed prior to delayed recall: either counting backwards from 10 (less demanding) or naming a series of images (more demanding). In study 2 with children, the task was either waiting quietly for 10 s (less demanding) or counting from 1 to 10 (more demanding). Further, sentences were either semantically plausible (The red bus was late so we drove by car) or implausible (The red grass was brave so we spoke to jam) with the idea that more familiar and meaningful sentences would benefit recall. The results were similar across both studies. Immediate sentence recall was better than delayed recall. Plausible sentences were recalled more accurately than implausible sentences. Finally, the more demanding the distracting task, the more participants relied on the semantics of the sentence to support recall (i.e., accuracy declined more steeply for implausible than plausible sentences).

Overall, results suggest that sentence recall draws on both phonological and semantic knowledge, but their contributions may differ. Phonology has a relatively greater role in immediate recall, whereas semantics has a relatively greater role in delayed recall. As well, familiar information supports memory overall. Clinicians who have a good understand about the language and memory skills involved in sentence recall will be in a better position to interpret their assessment findings. Further, there may be a need to incorporate delayed testing into practice.

Blogger: Theresa Pham is a student in the combined SLP MClSc/PhD program, supervised by Dr. Lisa Archibald

Exploring mentorship as a strategy to build capacity for knowledge translation research and practice: A scoping systematic review

Gagliardi, A. R., Webster, F., Perrier, L., Bell, M., & Straus, S. (2014). Exploring mentorship as a strategy to build capacity for knowledge translation research and practice: A scoping systematic review. Implementation Science, 9(1), 1-10.

Knowledge translation (KT) refers to the movement of research knowledge into practice. It improves healthcare outcomes by promoting the use of new findings in clinical settings, management settings, and decision-making regarding healthcare policies. Within KT there is knowledge transfer which is the movement of new knowledge from research to practice, and knowledge exchange which is the multidirectional movement of knowledge between researchers and knowledge users (i.e., clinicians, decision makers, policy makers). Engaging in KT is a complex process and for these approaches to be successful the appropriate knowledge of KT, infrastructure for KT, and incentive to engage in KT need to be in place. 

It has been identified that those involved in KT would benefit from receiving support to build their KT capacity. In the past, building KT capacity involved KT training for those involved in a KT project. The current authors were interested in the use of mentorship to increase KT capacity. Mentorship provides the chance for an interactive experience and a partnership between the mentor and mentee to promote learning and development of KT. 

In this study the authors completed a scoping review to examine the effectiveness of mentorships as a way to support the development of job-related knowledge. They were interested in understanding the components that support a successful mentorship (e.g., design of mentorship, goals of mentorship). The scoping review identified 13 articles that reported on the use of mentorships to increase job-related knowledge. Results revealed that in 12/13 studies those in the mentorship self-reported that they achieved their goals related to the mentorship. In most studies the mentorships were formally established, mentees were specifically matched with mentors, and most mentorships were hierarchical. Some barriers faced in mentorships included issues if specific goals were not laid out by mentorship program, and if the mentee felt that their mentor was untrustworthy. 

 

These results provide insight into the importance of preparing, educating and supporting those who are engaging in a new project. Mentorships are an interactive way to support the development of new knowledge and skill. This scoping review was the first step in understanding what components are necessary for a successful mentorship. As more researchers and knowledge users begin to engage in KT approaches, understanding how best to prepare and support those involved in the process will be important for project success.   

                          

Blogger: Meghan Vollebregt is a student in the combined SLP MClSc/PhD program working under the supervision of Dr. Lisa Archibald.

Language repetition and short-term memory: an integrative framework

Majerus, S. (2013). Language repetition and short-term memory: an integrative framework. Frontiers in Human Neuroscience, 7, 357–357. https://doi.org/10.3389/fnhum.2013.00357

Tasks such as repeating words and sentences likely rely on both language skills and verbal short-term memory (the ability to temporarily hold verbal information in mind). To capture the interaction between language knowledge and verbal short-term memory, this paper proposes a three-component framework consisting of:

1.   Item information: This is the knowledge associated with ‘knowing a word’ including phonological (speech sounds that form the word), semantic (meaning-based information), and syntactic information (grammar). All these different types of knowledge are said to be immediately activated upon encountering a (familiar) word. 

2.    Serial order information: Order information refers to keeping information in the exact order as it was presented. Think of learning a new word like kipser. Because you likely know nothing about the word, prior knowledge would not support learning. Instead, it is important to keep the order of the speech sounds in mind in order to retain it. Serial order is also important when trying to remember a phone number, for example.

3.   Attention: Attention is the ability to focus on what matters and keep that information active in mind until the end of a task. As the task becomes more difficult, more attention is needed to keep items in mind. Attention is also important in coordinating and maintaining both item and order information during language tasks.

Measuring brain activity during language tasks has been a helpful way to understand how these three domains interact with each other to support verbal short-term memory. Language pathways in the front (ventral) and back (dorsal) of the brain maintain phonological and semantic information, respectively. In addition to these language pathways, more difficult tasks like new word learning and recalling complex sentences also activate brain areas tied to serial order and attention processes. There is a network in the right front-parietal region of the brain that is important for maintaining serial order information. The left fronto-parietal network, on the other hand, is important for actively maintaining items in the focus of attention. 

This framework shows the importance of recognizing that some language tasks used in clinical assessments may be as much a measure of language skills as verbal short-term memory. Interpreting children’s performance should take into account task demands related to linguistic knowledge (or lack thereof), efficiency of serial order processing, and attention. 

Blogger: Theresa Pham is a student in the combined SLP MClSc/PhD program, supervised by Dr. Lisa Archibald.

Using a Design-Based Research Approach to Develop and Study a Web-Based Tool to Support Collaborative Learning

Lyons, K. M., Lobczowski, N. G., Greene, J. A., Whitley, J., & McLaughlin, J. E. (2020). Using a design-based research approach to develop and study a web-based tool to support collaborative learning. Computers & Education, 104064.

Engaging in collaborative learning is beneficial for students as it creates supportive learning environments. Difficulties in collaborative learning can arise when conflicts in the group are not dealt with in productive ways. Theories of social regulation of learning suggest that successful collaborative learning includes students monitoring and regulating their own and their groups’ cognition, motivation and emotions. To support students in successful collaborative learning, researchers have developed and implemented web-based tools focusing on 3 principles for group regulation (promoting metacognitive awareness of the learning process, supporting externalization of the learning process and promoting the acquisition and activation of regulation process). 

 

The current article identified the need for a comprehensive web-based tool that included all 3 principles and additionally addressed the need to scaffold students’ metacognitive knowledge of how to use these principles. A design-based research approach was used to develop a comprehensive tool to foster social regulation of learning. Design-based research is an approach to research that takes an iterative and systematic approach to investigate the design and implementation of a tool in a specific learning environment. 

 

In this study, authors developed and implemented the tool, Collabucate, over two design-based cycles. Each cycle included the following phases: focus on the problem, understand the problem, determine existing solutions, define the current goal, conceive the solution, build the solution, test the solution, and understand results. Cycle 1 involved 29 participants who completed the tool twice a week for six weeks. Results included data from their weekly submissions and qualitative data from a survey and focus group. Students identified advantages (e.g., increased metacognitive awareness) and disadvantages (e.g., the need for a group discussion surrounding strategy implementation) of the tool that were considered in cycle 2. In the second iteration, disadvantages of the tool were addressed by creating a new section in the web-based tool and making the tool visually appealing. Cycle 2 involved 83 participants and results revealed that students indicated previously seen disadvantages were now advantages for the tool. 

 

These results support the use of a web-based tool that aids in social regulation of learning within collaborative learning environments. Additionally, this study demonstrates the iterative nature of design-based research and describes the different phases involved in developing a tool suited to a specific learning environment. This project reveals both the complexity and success that can come from design-research.

                          

Blogger: Meghan Vollebregt is a student in the combined SLP MClSc/PhD program working under the supervision of Dr. Lisa Archibald.

Lexical Diversity Versus Lexical Error in the Language Transcripts of Children With Developmental Language Disorder: Different Conclusions About Lexical Ability

Charest, M., & Skoczylas, M. J. (2019). Lexical Diversity Versus Lexical Error in the Language Transcripts of Children With Developmental Language Disorder: Different Conclusions About Lexical Ability. American journal of speech-language pathology, 28(3), 1275-1282.

 

It is well documented that the profiles of children with development language disorder (DLD) prominently feature grammatical challenges. However, evidence shows that children with DLD also struggle with word learning and use. Studies have shown that compared to children with typical language development (TLD), children with DLD frequently forget newly learned words, make more confrontation naming errors, have less robust understanding of word meanings, and need more exposures to new words in order to learn them. These results suggest that children with DLD have limited lexical knowledge when compared to children with typical language development (TLD). 

 

Previous studies have used lexical diversity as a measure of lexical ability in children with DLD and TLD, and have shown mixed results. The authors of the current work posit that these mixed findings do not necessarily indicate that lexical ability is unreliable in distinguishing children with DLD from those with TLD. Alternatively, it may be the case that clinical measures that more accurately capture the lexical-semantic challenges noted in children with DLD are needed. The authors hypothesize that rate and type of lexical errors may be more indicative of differences in lexical ability between children with DLD and those with TLD. 

 

In this study, 7 children with DLD and 7 children with TLD ranging from 5-7 years of age  completed the Edmonton Narrative Norms Instrument (ENNI). The resulting transcripts were coded and analyzed both in terms of lexical diversity and lexical errors. Similar to previous work, various measures of lexical diversity were not found to discriminate between DLD and TLD in the current study. Lexical errors were identified by three separate coders. In total, 198 lexical-semantic errors were identified, with a significantly higher number of errors observed in the DLD group. The authors noted, however, that coding lexical errors was a highly subjective process. Only 38% of all errors in the DLD group were initially flagged by all three coders. Therefore, although focusing on lexical errors has potential for characterizing the lexical-semantic abilities of children with DLD, a detailed framework for describing these errors is needed for it to prove clinically useful. 

Blogger: Taylor Bardell is a combined MClSc/PhD student in Speech-Language Pathology, supervised by Dr. Lisa Archibald. 

  

Evaluating Part V of the German version of the Token Test as a screening of specific language impairment in preschoolers.

Schmoeger, M., Deckert, M., Eisenwort, B., Loader, B., Hofmair, A., Auff, E., & Willinger, U. (2019). Evaluating Part V of the German version of the Token Test as a screening of specific language impairment in preschoolers. Applied Psycholinguistics, 41(1), 237–258. https://doi.org/10.1017/s0142716419000493

Children with Developmental Language Disorders (DLD) have a persistent problem learning language despite normal development. DLD is also associated with weak verbal working memory, which is the ability to store and process some aspect of language in mind, and other impairments related to cognition and attention. A valid and efficient screening tool for DLD is therefore critical.

One tool that may be well suited to screen for DLD is Part 5 of the Token Test (De Renzi & Vignolo, 1962). The Token Test is a simple and easy tool used to assess language. In the Token Test, the child listens to the command and then points to the sequence of shapes, with commands increasing in length and complexity each time. Commands in Parts 1 to 4 involve pointing to shapes by colour and size and increase in length (e.g., Part 1: Point to the circle; Part 4: Point to the small, red circle and the large, blue square). Part 5, on the other hand, draws on different skills because it requires understanding long and linguistically complex commands (“Instead of touching the white square, touch the yellow circle”). The authors were interested in whether the Token Test may be a reliable tool for screening language and cognitive difficulties in DLD.

In this study, 4-6 year old children with DLD and typically developing children completed the Token Test and a test of intelligence. Not surprising, children with DLD made more mistakes than typically developing children on all parts of the Token Test, including the easiest and hardest parts. When looking at the whole sample, Part 5 was not shown to be an effective screening tool, but results were more promising for 4- and 5-year-olds. Part 5 was able to correctly classify children between ages 4-5. Further, Part 5 may also be a good indicator for general cognitive abilities. Children with DLD who were correctly classified by Part 5 showed worse performance on the verbal and non-verbal scales of the intelligence test than incorrectly classified children with DLD, whereas only the verbal scale correlated with Part 5 for typically developing children. 

The authors suggest that there are several advantages to the use of Part 5 of the Token Test as a clinical tool. But more research is needed to determine whether Part 5 and the Token Test more generally can be seen as a valid and efficient tool for screening DLD and intellectual skills.

De Renzi, A., & Vignolo, L. A. (1962). Token test: A sensitive test to detect receptive disturbances in aphasics. Brain, 85,665–678.

Blogger: Theresa Pham is a student in the combined SLP MClSc/PhD program, supervised by Dr. Lisa Archibald.

Lessons learned about the effective operationalization of champions as an implementation strategy: Results from a qualitative process evaluation of a pragmatic trial.

Bunce, A. E., Gruß, I., Davis, J. V., Cowburn, S., Cohen, D., Oakley, J., & Gold, R. (2020). Lessons learned about the effective operationalization of champions as an implementation strategy: Results from a qualitative process evaluation of a pragmatic trial. Implementation Science, 15(1), 1-12.

 

The field of implementation science focuses on identifying barriers to practice change and introducing implementation strategies to address such barriers. The current authors previously completed a project where they utilized 5 implementation strategies to integrate a new guideline into clinical practice. Clinics involved received varying level of implementation intensity, and they hypothesized that the greater number of intensive implementation supports received the greater the improvement in guideline integration. Results from this study revealed that greater supports did not lead to greater integration. This result led the authors to look further at specific implementation strategies and the success of specific strategies in individual community clinics. This also demonstrated the importance of operationalizing implementation strategies to better aid in the selection of strategies. 

 

In this study, the authors re-analyzed quantitative and qualitative data from the previous project. Results revealed that for the clinic sites demonstrating a significant pre-post difference in guideline integration, the two factors associated with this change were the influence of the “implementers” or “champions” and their impact on the organizational context. Implementers were defined as a champion for promoting guideline related activity and acting as a liaison between the health centre and the researchers.  In an analysis of the factors associated with change, the 4 key implementer characteristics were engagement in promoting intervention, influence to foster trust, credibility, and capacity. 

 

These results add to the literature supporting the use of implementers/champions and organizational support in efforts to change practice. Further these results open the discussion surrounding how to support these individuals and better operationalize what factors are associated with a successful implementer. Broadly, this research highlights the complexity of identifying what implementation strategies will support change in a clinical context.

                                            

Blogger: Meghan Vollebregt is a student in the combined SLP                                                                                                       MClSc/PhD program working under the supervision of Dr. Lisa Archibald.