The Number Race

The Number Race software is designed for remediation of dyscalculia in children aged 4-8. It may also be useful for prevention of dyscalculia, or to teach number sense in kindergarten children without specific learning disabilities. The software was developed in our laboratory by Anna Wilson and Stanislas Dehaene, and is based on our current knowledge of the brain circuits underlying numerical cognition. Details of the design of the software are published in Behavioral and Brain Functions:

Wilson, A. J., Dehaene, S., Pinel, P., Revkin, S. K., Cohen, L., & Cohen, D. (2006). Principles underlying the design of “the number race”, an adaptive computer game for remediation of dyscalculia. Behavioral and Brain Functions, 2(19).

Click here to download The Number Race software (currently available in Dutch, English, French and German & Spanish).

News

1/4/2007: Version 2.3 now released.

31/3/2007: New Spanish version now available, thanks to Josep M Serra-Grabulosa!

Who can use this software?

Parents, teachers and researchers are equally welcome to download and use the software with children. (Researchers see notes below). It is designed for children to use with a minimum of supervision.

What does the software contain?

The software trains number sense with single digit numbers, numerical comparison, enumeration, one-to-one correspondance, counting, reading arabic digits, and simple (1-digit) addition and subtraction.

How does the software work ?

The software is an adaptive game, which trains children on an entertaining numerical comparison task. It adjusts the difficulty of the task to their level of performance, maintaining their average success at 75%. Children have to choose the larger of two quantities of treasure (or coconuts) before the competitor character (controlled by the computer) steals the largest. In higher levels, children may have to add or subtract in order to make the comparison. They then use the set of gold pieces or coconuts that they selected in order to advance in a race against the competitor character.  At higher levels, they have to try and avoid hazards, which appear at random on some board squares and cause characters to go backwards when landed on. When children reach the end of the board before their computer-driven competitor does, they collect a reward. After enough rewards are collected, they can unlock a new character to play with. The adaptive algorithm of the game adjusts the numerical distance between the quantities to be compared, the length of the response deadline, and the complexity of the choice (in terms of ratio of symbolic to non symbolic representations of number, and presence of addition and or subtraction).

What is the impact on children?

Scientific evaluation of the efficacy of the software is currently at the second of  two stages.

The first stage was testing the software in a 4-month open-trial study, in which we examined whether a group of 9 children aged 7-9 with mathematical learning difficulties showed improvement in basic numerical cognition tasks after 5 weeks of using the software for 2 hours a week. The children improved in small number perception, comparison, and simple arithmetic (subtraction). Results from this study have been published in Behavioral and Brain Functions:

Wilson, A. J., Revkin, S. K., Cohen, D., Cohen, L., & Dehaene, S. (2006). An open trial assessment of “the number race”, an adaptive computer game for remediation of dyscalculia. Behavioral and Brain Functions, 2(20).

The second stage is ongoing, and entails studies including a control condition (e.g. using cross-over or waiting list designs). The aims of this stage are to establish a) that the improvements seen in the first study were not due to normal development, teaching in the classroom, or test-retest effects, and b) in which populations the software is most effective.

As of 27/12/2006 three second stage studies have been conducted by ourselves and our collaborators:

• Clermont-Ferrand (2006), in collaboration with Michel Fayol and Ophélie Dubois. This study was a cross-over design, using a control (reading) software. The sample was 53 kindergarten children of low SES. Results showed clear cross-over effects in symbolic and cross-format  (eg. digits to dots) numerical comparison tasks. The manuscript for this study is submitted for publication.
• Aachen (2006), in collaboration with Klaus Willmes and Helga Krinzinger. This study was a waiting-list design. The sample was 18 seven to nine year old dyscalculic children (below 10th percentile in maths), and a group of 26 control children. Initial results show cross-over effects in symbolic and cross-format numerical comparison tasks. We are currently conducting further analyses and preparing a manuscript for publication.
• Paris (2006) in collaboration with Monique Plaza. This study was a cross-over design including a control reading software. The sample was 24 7-year-old children with mild difficulties in both reading and mathematics, and 23 control children. So far we have not found much evidence for an effect of the software in the analyses of this data. In the presence of the positive results above, we are not interpreting this as evidence against the software, but rather we think it is probably due to the sample, which due to recruitment constraints was only below the 35th percentile in mathematics, and included children with low IQ, ADHD, and bilingualism. Thus children in this sample could not be described as dyscalculic, and likely had difficulties in math due to other reasons.

Acknowledgements

Funding for development of this software was provided by a post-doctoral fellowship to Anna Wilson from the Foundation Fyssen and a McDonnell Foundation centennial grant to Stanislas Dehaene. Funding for further development and testing of the software was provided by INSERM.

In addition, we gratefully acknowledge a consultancy contract to Anna Wilson from the Brain and Learning Project of the OECD.

Using the software for research

Researchers interested or planning to use the software in studies are encouraged to for suggestions about research protocol, and information about how to analyse data from the software. This will also help avoid unnecessary duplication of research studies.

Support

Please note that the software is provided without user support. Consult the software documentation for installation and use instructions.

Known bugs

The software contains some minor known bugs, which can be viewed in the SourceForge bug database. You may with reports of further bugs, and we will endeavour to fix major bugs where possible. Currently there are no major known bugs.

Further development

The software is written in Java, and is open-source, distributed under the GNU general public license. If you would like to further develop the software, you need to first download the English version source code.

Please note that you are in turn obliged to distribute any developments you make under the conditions of the GNU general public license.

We recommend to inform us that you are developing the software, because she can help you with developer documentation.

The software makes use of two other open-source modules: BeanShell (GNU LPGL), and JMAT (GNU LPGL), as well as selected modules from David Brackeen.

Software License

Copyright (C) Anna Wilson and Stanislas Dehaene, 2004

This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for more details.

Back to the unicog main page.