Since the mid-1990s scientists have floated the idea that representations of numeric quantities, whether expressed as digits or as written words, are codified by the parietal cortex, a higher-processing region in the brain located just above the forehead. The notion is supported by calculation deficits observed in patients with damage to that brain region and in low birth weight children who exhibit reduced gray matter in adulthood.

Still, it is unknown whether the parietal cortex—specifically its left hemisphere—processes numeric value independent of notation. New research indicates that the brain region may prefer symbolic notation to other numeric representations—a finding that could open the door to helping kids plagued by dyscalculia, a learning disability characterized by severely impaired mathematical ability.

"Knowing how we represent numbers is very important for education, to have rehabilitation, to know exactly what is the core of deficit and know how we can train [dyscalculia sufferers]," says researcher Roi Cohen Kadosh, a cognitive neuroscientist at University College London.

Two studies, published this week in the journal Neuron, employ a slight tweak on functional magnetic resonance imaging (fMRI) known as "fMRI adaptation." As activity was monitored, researchers noted changes under the assumption that the brain reduces activity as it becomes accustomed to a stimulus and then reactivates when a novel stimulus is presented.

In a trial led by Cohen Kadosh, 12 subjects were presented with a sequence of two numbers, represented as either the digits (2 or 6) or the words (two or six). The sequence proceeded with the digit 2 followed either by the word two or by the numeral 2. Each numeric representation was flashed for 350 milliseconds; then, after a time lapse of 1,300 milliseconds, the second symbol flashed. The values also appeared in different patterns in which the number 2 could be followed by the word or the digit 6.

Cohen Kadosh's team found that, regardless of notation, there was recovery of activity in the left parietal lobe when the numeric quantity was changed, as opposed to when it remained the same. In the right lobe, however, the adaptation effect was stronger in trials involving digits than it was for written words, Cohen Kadosh says, hinting at a notation-dependence. This indicates, he explains, "there are different neuronal populations for digits and different populations for written numbers."

Manuela Piazza, a neuroscientist at France's National Institute of Health and Medical Research, used a slightly different approach. For 30 seconds she exposed 14 volunteers to a specific range of numbers (like, 17 to 19) in the same notation&emdash;either numerals or a cluster of dots. After the initial half minute, over the next 90 seconds, some deviant quantities could be inserted while staying within the same range; for example, 17 dots in a sequence of numerals between 17 and 19. Next, there was a jump to a new range of numbers, like 47 to 49, although the primary mode of expression, dots or numerals, remained consistent. After the initial 30 seconds needed for the subjects to adapt to the new range, deviants (of the previous range or a different notation) could appear once again.

Piazza and her colleagues noticed that if a deviant quantity was farther from the adapted quantity, the rebound activation in the brain was larger regardless of notation. In the left parietal region, however, the fMRI measured slightly more activity when deviant dots appeared in a set of numerals than when numerals appeared in a set of dots. The researchers interpret this to mean that "at least in the adult brain, numerical symbols and nonnumerical numerosities converge onto shared neural representations," meaning that there are "shared neuronal populations" that can map symbols onto nonsymbolic quantities.

Daniel Ansari, a cognitive scientist at the University of Western Ontario in Canada, says that while the new findings appear to disagree, both studies suggest "that the left parietal lobe is more finely tuned to the symbolic representation of numerical magnitude."

2 Comments

1. 1. Hassan 05:58 AM 11/27/07

   ]Application of Artificial Neural Networks Modelling is a very relevant tool for systimatic investigation of the presented issue . That is supported by my recent research work .As an example for my claim is the following paper H.M.Hassan , et.al.: "Towards Evaluation of Phonics Method for Teaching of Reading Using Artificial Neural Networks (A Cognitive Modeling Approach)",to be published at IEEE Symposium on Signal Processing and Information Technology Seventh Symposium will be held in Cairo-Egypt , during 15-18 December 2007.

   Reply | Report Abuse | Link to this

2. 2. tripathi 10:39 AM 12/18/07

   67 11 11 71 19 23 67 23 67 43 71 41 11 2 43 71 71 47 3 11 19 11 61 11 83 19 2 71 23 67 23 71 67 61 23 17 19 71 53 37 2 5 11

   Reply | Report Abuse | Link to this