Numerical cognition

# Friday, September 30th, 2011 [17:20 - 19:20]

## PS_1.044 - Brain activity during multiplication in more- and less-skilled adults: an event-related potential study

Sobanska, M. , Szumska, I. , Warakomski, D. & Jaskowski, P.

University of Finance and Management

It is widely assumed that high-skilled adults use retrieval strategy to solve single-digit multiplication problems more frequently that others, who often base on more time-consuming and more error-prone procedural strategy. The latter is usually applied in problems of large products which may be a source of so-called problem-size effect. The aim of the present study was to compare the brain activity during multiplication in more- and less-skilled adults. Previously Jost et all. (2004) showed that topography of a slow negative wave, associated with the implicite production of multiplication result, varied with the problem type (small v. large), finding which was explained in terms of using different solution strategies (retrieval v. non-retrieval). However the level of participants’ arithmetic skills was not controlled. To investigate the cause of the topographical differences described before we used the same implicit production task with event-related potentials (ERPs) recoded from 64 scalp positions in 70 subjects. Basing on the behavioral data, e.g. speed and accuracy of simple arithmetical problems solving, we established two groups: 11 more-skilled participants, and 15 less-skilled participants. Further analysis will be carried out and discussed in the framework of models of mental arithmetic.

## PS_1.045 - A study on the operand-order effect in single-digit multiplications in Italian and English mother tongues

Didino, D. ^{1, } ^{2 } , Vespignani, F. ^{1 } & Lombardi, L. ^{1 }

1 Dipartimento di Scienze della Cognizione e della Formazione, University of Trento, Trento, Italy

2 Institute of Cognitive Neuroscience, University College of London, London, United Kingdom

In order to better understand the internal organization of the memory network that encodes arithmetic facts we studied the effect of operand-order on the speed of solution of one-digit multiplications. 24 Italian participants were asked to report the result of the operations. The analysis of the RTs showed an interaction between the size of the problem (both operands larger than 5, e.g. 7x8; one larger and the other smaller than 5, 3x7; and both smaller than 5, 3x4) and the order of the operands (first larger, 8x4; and second larger, 4x8). When both operands were larger than 5, the problems with the second operand larger were solved faster, e.g. RT(7x8)<RT(8x7). When one operand was larger and the other was smaller than 5, the problems with the first operand larger were solved faster, RT(7x3)<RT(3x7). Similar results were obtained in a second experiment in English (24 participants). We exclude the interaction is due to the operand-order during the acquisition of the multiplication table, given that the two languages differ in the order of operands during the table learning. The results can be rather explained by a model of retrieval with an asymmetric right skewed activation of multiples around ties.

## PS_1.046 - Numerical abilities in deaf children with Cochlear Implant. Evidences from magnitude comparison tasks

Iza, M. , Rodriguez, J. M. , Calleja, M. , Garcia, J. & Damas, J.

University of Malaga

Usually deaf children show lower scores in numerical tasks than normal hearing peers. Explanation of mathematical disabilities in hearing children are based on a quantity representations deficit (Geary 1994) or on an access deficit to such representations (Rousselle&Noël 2008). The aim of this study is to verify whether deaf people show a deficit in representation or in access to numerical representations by using both symbolic (Arabic digits) and non-symbolic (dot constellations and hands) magnitude comparison tasks. 10 profoundly deaf children using cochlear implants (mean age 9) and 10 normal-hearing children, matched in IQ, visual STM, oral language skills and age, participated in the study. Numerical distance (1vs3-4) and magnitude (1-5vs5-9) were manipulated. RT analysis show a significant interaction TaskxGroup [F(2,36)=3.68; p=.035]: No differences were found between both groups in non-symbolic comparison tasks, however deaf participants were slower than hearing participants in the symbolic task Magnitude and distance effects were found across groups and tasks. Our results suggest that magnitude representations are similar in both groups. However, deaf children seem to have difficulties in accessing magnitude representations through symbolic codes. Following Budgen&Ansari (2011), a slower activation of semantic numerical information might explain deaf children lower scores in numerical tasks.

## PS_1.047 - What is important for numerical-spatial association?

Herrera, A. ^{1 } , Macizo, P. ^{2 } , Flores, A. ^{1 } & Juárez, V. ^{2 }

1 University of Murcia

2 University of Granada

A debated question in cognitive psychology regards to the origin of the numerical-spatial associations that are usually found when individuals process numbers. Recent works have shown that visuospatial or verbal short-term representations of numerical information could be responsible of these spatial effects. Moreover, ordinal information, instead the longstanding assumed magnitude information, seems to have a prominent role when participants perform a parity task. In the present work, we conducted a series of experiments in which different types of non-verbal information (ordinal, visual or spatial) had to be maintained while participants performed a comparison task or a parity judgment task. The results showed that type of information to be maintained produced a differential effect on the numerical-spatial association. However, it is also dependent on the numerical task.

## PS_1.048 - Discrete and continuous quantity judgment in adults: Number counts more than area!

Nys, J. & Content, A.

Laboratoire Cognition Langage Développement (LCLD), Université Libre de Bruxelles (ULB), Brussels, Belgium

A Stroop-like paradigm was used with 57 adult participants who were asked to perform (1) a number judgment and (2) an area judgment on dot collections orthogonally varying along a discrete dimension (number of dots) and a continuous dimension (cumulative dot area). In the number comparison task, as expected, incongruent trials for which the largest number of dots corresponded to the smallest cumulative area led to larger error rate and reaction times than congruent trials for which number and area covaried positively. This finding suggests that area is automatically processed and integrated during a discrete quantity judgment task. Interestingly, a similar interference effect was observed in the area comparison task, providing evidence that adults are unable to ignore numerical features of the stimuli even when task-irrelevant. Moreover, participants tended to select the numerically largest collection when cumulative area, the task-relevant dimension, was equal for both sets. By contrast, in the number comparison task, they showed no preference for the set with the largest area when the number of dots was identical for both collections. Contrasting with earlier statements, these results support the view that, in adults, number is automatically extracted and acts as a more salient cue than area.

## PS_1.049 - The stability of the SNARC effect - can we reach it?

Cipora, K. ^{1 } & Wood, G. ^{2 }

1 Institute of Psychology. Jagiellonian University. Cracow, Poland.

2 Institute of Psychology. Karl-Franzens-University of Graz. Graz, Austria.

The SNARC (Spatial Numerical Association of Response Codes) effect is regarded as an index of the association of numbers and space. Although the SNARC effect has been consistently replicated, almost no studies reported between-groups differences. This led some to question the sensitivity of the SNARC effect to individual differences. Here we examined whether the lack of differences is due to poor estimation. The impact of sample size, number of repetitions per condition, and intraindividual variability on the probability of finding a non-zero SNARC effect was investigated. Simulations revealed that the most important factor determining the probability of detecting non-zero SNARC effect was the number of repetitions per condition. Moreover, very small sample sizes jeopardize the detection of SNARC, particularly when the number of repetitions is small. Therefore, failures to find significant between-groups differences is due foremost to lack of power to detect them, rather than to the absence of those differences. Fortunately, an adequate estimation of the SNARC effect is possible even with modest sample sizes when using 20 or more repetitions per condition. We conclude that the SNARC effect has the potential to reveal much more individual differences than what was reported before.