Annelies Ceulemans

Enumeration of small and large numerosities in adolescents with mathematical learning disorders

The accuracy and speed in an enumeration task were investigated in adolescents with typical and atypically poor development of arithmetic skills. The number naming performances on small and large non-symbolic numerosities of 18 adolescents with mathematical learning disorders (MLD) and 28 typically achieving age-matched (TA) adolescents were compared. A mixed logistic regression model showed that adolescents with MLD were not significantly less accurate on numbers within the subitizing range than control peers. Moreover, no significant differences in reaction times were found between both groups. Nevertheless, we found that within the control group adolescents with higher ability tended to respond faster when taking into account the whole range (1-9) of numerosities. This correlation was much weaker in the MLD group. When looking more closely at the data, however, it became clear that the correlation between accuracy and speed within the control group differed in direction dependent on the range (subitizing or counting) of the numerosities. As such, our findings did not support a limited capacity of subitizing in MLD. However, the data stressed a different correlation between speed and accuracy for both groups of adolescents and a different behavioral pattern depending on the numerosity range as well. Implications for the understanding and approach of MLD are considered.

Number Sense in Siblings of Children with Mathematical Learning Disabilities: A Longitudinal Study

Number sense, counting and logical thinking were assessed in 14 siblings of children with Mathematical Learning Disabilities (MLD) and in 41 age matched children without family members with MLD. The children were tested in kindergarten and followed up in grade 1. A 0-100 number line estimation paradigm with three formats (Arabic digits, dots and number) was used as a measure of number sense. Results reveal that siblings of children with MLD are less proficient in number line placements compared to non-siblings, with both groups having a logarithmic representation in kindergarten and grade 1. Siblings also differ from non siblings on procedural and conceptual counting knowledge and logical thinking in kindergarten. In addition, our findings suggest that nnumber line estimation in kindergarten is especially predictive for untimed procedural calculation performances in grade 1, whereas procedural counting knowledge is related to timed fact retrieval skills in grade 1. Our findings also reveal that MLD had a familial aggregation. Clinical siblings especially differ from non-clinical siblings on the estimation with Arabic numbers (in kindergarten and grade 1) and number words (in grade 1), pointing to the fact that especially symbolic number line estimation tasks on a 0-100 scale can be used as screeners for MLD. Implications for the understanding and diagnosis of MLD are discussed.