Asymmetrical interference between number and item size perception provide evidence for a domain specific impairment in dyscalculia

Abstract

EC) 19 20 42 discriminability of these features is matched, as here in control subjects. By quantifying, for 43 the first time, dyscalculic subjects' degree of interference on another orthogonal dimension of 44 the same stimuli, we are able to exclude a domain-general inhibition deficit as explanation for 45 their poor / biased numerical judgement. We suggest that enhanced reliance on non-46 numerical cues during numerosity discrimination can represent a strategy to cope with a less 47 precise number sense. 48 49 50 Developmental dyscalculia, Inhibitory control 52 53 effect, see for example Barth, 2008;Hurewitz et al., 2006;Nys and Content, 2012; Rousselle 89 and Noël, 2008). This theory places the origin of interference at the level of the response 90 selection. Alternatively, it has been proposed that interference may originate at the level of 91 sensory extraction: models based on the stimulus energy at different spatial scales can yield 92 non-veridical estimates of the number of items in a display resembling the biases of human 93 observers (Dakin et al., 2011), and within hierarchical generative networks, interference from 94 non-numerical quantities has been related to the efficiency of a normalization process 95 embedded into the extraction of numerosity representations (Cappelletti et al., 2014b; 96 Stoianov and Zorzi, 2017). Nevertheless, some authors have interpreted interference to 97 indicate that numerosity is indirectly inferred from a combination of non-numerical 98 quantitative features (though without specifying which combination of features in detail), 99 sometimes going as far as to completely deny the existence of a dedicated perceptual 100 mechanisms for numerosity (for a review see: Leibovich et al., 2016a). 101It is noteworthy that among the studies that found strong interference of non-102 numerical dimensions on numerosity comparison, many required participants to judge rather 103 difficult numerical ratios, even between 0.9 and 1.1 (DeWind et al., 2015; Nys and Content, 104 2012;Tokita and Ishiguchi, 2010). Importantly, the strongest interference is usually observed 105 for the most difficult numerical ratios with a tendency to decrease for the easier comparisons 106 (Hurewitz et al., 2006;Nys and Content, 2012). It is well-known that comparative judgments 107 without counting are not perfect but approximate, depending on the ratio of the compared 108 numbers with a precision that is commonly operationalized by the Weber fraction. It is hence 109 conceivable that when subjects are required to make decisions close to or beyond the 110 precision of their numerosity processing system, they would attempt to rely on associated 111 quantities to solve the task, especially since in everyday life these often provide correlated 112 information. However, such heuristic use of non-numerical information need not be the only 113 possibility: even in symbolic number-size interference tasks, which are not limited by 114 sensory/perceptual precision to the same extent as non-symbolic numerosit...

Recruitment of the occipital cortex by arithmetic processing follows computational bias in the congenitally blind

“…Interestingly, some studies demonstrated that mathematical abilities do not correlate with specific computations: size and density sensitivity do not correlate with math, in contrast to visual numerosity (Anobile et al, 2016(Anobile et al, , 2018. Similarly, average size (Castaldi et al, 2018), line length (Cappelletti et al, 2014;De Visscher et al, 2017) and area (Iuculano et al, 2008) are not impaired by dyscalculia, a learning deficit affecting number processing. The evidence that numerical concepts can be acquired in CB raise the question as to whether these perceptual factors that link to the development of arithmetic abilities in the sighted are independent from visual experience or whether, in contrast, CB people rely on separate perceptual processes in order to develop adequate arithmetic abilities through separate cognitive mechanisms (Dormal et al, 2016b; Crollen et al, 2017).…”

Random Forest multiclasse: a diagnostic study of mathematical learning difficulties

“…Among the psychometric tests used in the diagnosis of mathematics difficulties are the TDE (School Performance Test), Raven's Colored Progressive Matrices (RCPM), Simple Calculation Task, Arabic Numeral Writing, Woodcock-Johnson IV and TEDI Math Grands as well as neuropsychological tests to evaluate other cognitive functions such as verbal and non-verbal IQ (intelligence quotient) assessment with WAIS subtests (Wechsler Adult Intelligence Scale), evaluation of the verbal component of working memory with the 'digit span' (measures short-term auditory memory and attention) and evaluation of visual spatial working memory, with Corsi Blocks. (PAIVA, 2021;SILVA et al, 2020;CASTALDI et al, 2018).…”

“…In some contexts, where the goal is classification into multiple categories or classes, the simplified approach is built with the generation of binary models (single output model) for each class analyzed, distinctly. Alternatively, the decision tree method allows you to build a unique multiclass model, based on multi output model, whose main benefit is that the dependency between variables can be considered in the model (KOCEV et al, 2007;CHAUDHARY;KOLHE;KAMAL, 2016;LOUPPE, 2014).…”