The conference is over. See the video recording of the conference that includes the presentations from Cara Verwim, Giada Guerra, Daria Khanolainen and Irene Mammarella below:
Questions and answers from the live session:
Q:
How much of the results is dyslexia-specific and how much in applicable for reading development in general?
Cara Verwimp:
In our sample it were all children with reading problems
Q:
You talked about centrality before. Which variable was considered central in your model?
Cara Verwimp 02:23 PM:
Our reading measures were the most central
Let me elaborate on this. There are different measures to calculate the centrality, of which node strength, betweenness and closeness are the most well known. However, in the context of psychology, it is not clear what betweenness and closeness mean, it is related to shortest paths as you would go from one metro station to another and take the shortest way, but in psychological context it is not always the ‘shortest path’ that is taken. If you want to know more about this you can look into the work of Laura Bringmann or Sacha Epskamp :)
Q:
Are non-significant edges fixed to zero? (Then this may affect the whole network but depend merely on the N.)
Cara Verwimp 02:22 PM:
Yes these are fixed to zero indeed.
Q:
I liked this presentation very much, thank you! I may have missed it, but how would you evalute model parsimony in such a model? Which variables should be excluded because "redundant" for example?
Cara Verwimp 02:21 PM:
There’s not a real ‘gold standard’ to decide on this, but you can use theories in the field to decide on this. For example, we decided to include our reading measures as separate variables as studies have shown that contribution of PA and RAN shift at different timepoints for these, i.e highly frequent words tend to become automatized faster so especially related to RAN whereas pseudowords always rely on decoding so especially related to PA
Response:
I understand, thank you. Of course there are theoretical reasons to include these variables. I was just wondering whether the inclusion of highly correlated variables as for example reading and spelling accuracy, reading fluency and RAN etc... may have a negative impact on the model.
Cara Verwimp 02:50 PM:
Yes this could ofcourse have a negative impact on the model. I suggest for checking these relations before running such a network model and if you have correlations higher than .80 or so can consider to exclude one of them. There are also some techniques for controlling for false positives to get the most parsimonious model :)
Q:
I liked your presentation Cara! Thanks!
Cara Verwimp 02:37 PM:
Thank you Marianne! Decoding strategy refers to the initial mapping of speech sounds to visual symbol, which is especially important in early phases of accurate reading :) these are also trained in reading intervention. Don’t know whether this answers your question but reach out to me if it doesn’t!
Response:
Sure, thanks! I was confuse by the term ‘strategy’. I thought it was a special form of decoding. Sorry for the confusion! Nice presentation!
Cara Verwimp 02:53 PM:
Sorry for the confusion. It refers to strategies that are used to learn the asscociations in intervention. In the Netherlands we start with teaching the very easy one-letter correspondences and then go to more complex such as diphthongs etc :)
Q:
Thanks for your presentation, Giada! Regarding the correlation you mentioned between auditory attention and reading skills, was it significant within each single group as well (dyslexics only and controls only)?
Also, could it be that the divergence of scores in the LSS tasks is due to greater fatigability of dyslexic kids?
Giada Guerra 02:47 PM:
Hi, thank you for your question! When checking the correlation separately for each group, the significant correlation remained within the dyslexia group but not in the typical reders one. One possibility is the lack of power (we had 51 with dyslexia and 38 without), but it could also possibly means that children with dyslexia rely more on the attention to read due to their lack of automatised skills.
About your second question, I am not sure if this could be a possible interpretation given the short task time.
SRLD:
The SRLD is a scientific society, including senior and young researchers, facilitating exchange between researchers who work on specific learning disorders in the fields of reading (decoding and comprehension), writing (spelling, text production), and maths (number, calculation, and mathematical reasoning).
The mission of SRLD is to promote the study of cognitive, linguistic, social, emotional, cross-cultural, neural, genetic, and environmental influences on learning disorders. This Society will also have the role of promoting exchange of scientific information, encouraging international collaboration, organizing meetings, and supporting young researchers in the field. The Society will favour exchanges with other societies covering the same area but with partly different goals.
Programme
1. Keynote presentation by Minna Torppa: "Predicting PISA reading: the roles of cognitive skills, motivation, and gender"
2. Symposium "Multivariate structure in learning disabilities: Stepping into the complex world behind behavioral symptoms":
i. Cara Verwimp (University of Amsterdam): "A network approach to dyslexia: Mapping the reading network"
ii. Giada Guerra (University of London): "Individual differences in auditory attention and letter-speech sound learning in children with dyslexia"
iii. Daria Khanolainen (University of Jyväskylä): "The impact of family risk and the home learning environment on reading and mathematical development across grades 1 and 9"
3. Keynote presentation by Irene Mammarella: "No evidence for a core deficit in developmental dyscalculia or mathematical learning disabilities"
The keynotes
Minna Torppa is an Associate Professor at the University in Jyväskylä, Finland. She has published a number of studies on learning and learning disabilities, and have datasets that are rather unique. Her focus has been on risk and supportive factors of reading difficulties from birth to adulthood including genetic, environmental, cognitive, and motivational factors. She also has studies on learning difficulties and psycho-social well-being.
Irene Mammarella is an Associate Professor at the University in Padova. She is highly recognized researcher on learning disorders, and her main research areas are: Specific Learning Disorders (Developmental Dyscalculia, Dyslexia), high functioning autism, visuospatial working memory in developmental disorders and emotional aspects (mathematics anxiety, anxiety) in relation to academic achievement.
Registration
- The Conference is free of charge. For registration please use this registration form.
- Link to Zoom: We will be sending all registered participants the link to Zoom by E-mail.
-
Deadline for registration: December 2, 2020.
Zoom
The conference will be streamed through the Zoom video conference software. You will receive an email link to the Zoom conference that will enable viewing it through your web-browser, or through the Zoom app (can be downloaded from here (https://zoom.us/download): not required).
