Several species of animal can count, but only humans can do so using abstract number symbols like 鈥�2鈥� or 鈥�4鈥�. Now we know we do so using a different set of brain cells from those we use to recognise numbers by simply counting groups of objects. This in turn may underpin the unique evolution of mathematical ability in humans.
Evidence for this came from experiments involving nine patients with epilepsy whose brains had temporarily been implanted with electrodes to identify the source of their seizures. While the electrodes were implanted, researchers recorded the electrical activity from 585 individual neurons in each participant.
The neurons were located in the medial temporal lobe, a region vital for processing sound and vision, memory and cognition. As participants each performed 320 simple arithmetical additions and subtractions of numbers from one to five, electrical signals revealed which neurons were active.
Advertisement
Crucially, half the sums were presented using customary numerical symbols. The other half were presented using quantities of blobs within a circle to represent numbers. So in one case, the participant had to recognise abstract symbols as numbers to do the sums while in the other case they had to count the blobs before they began the calculation.
Neural fingerprint
The researchers discovered that small sets of neurons lit up when a participant recognised an abstract number symbol 鈥� almost like a neural fingerprint for that symbol. They also found there was a neural fingerprint when a participant had to count blobs. But the two fingerprints associated with any given number were distinct 鈥� in other words, the neurons that lit up when a participant saw the number symbol 鈥�4鈥� were almost completely different from the neurons that lit up when they had to count four blobs in a circle.
鈥淥n average, and across all participants, 16 per cent of the randomly selected neurons responded to visible quantities, and 3 per cent to abstract numbers,鈥� says Andreas Nieder of the University of T眉bingen in Germany, and co-leader of the research team. Only 1 per cent 鈥� or six individual neurons 鈥� were involved in both. 鈥淭his indicates that we have two segregated populations of neurons in our brains 鈥� one responsible for visible quantities and another for number [symbols],鈥� he says.
Elizabeth Brannon at the University of Pennsylvania says it鈥檚 surprising that the medial temporal lobe seems to play such an important role in number identification, because this brain region is usually implicated simply in learning and memory. 鈥淭hese findings suggest much more of the brain might be involved in representing number than we previously thought,鈥� she says.
Monkeys counting
Previous research by Nieder鈥檚 team in 2007 revealed that when monkeys count, each number is registered by specific neurons. He says that the same probably applies in other animals known to count. What鈥檚 different in humans is that we have this ability but we also use a different set of neurons to count using abstract symbols. This abstract number ability probably evolved from the more basic, simpler form of counting, he says.
Nieder says the findings suggest that abstract number symbols gain their numerical meaning during our childhood because we learn to link them to the visible quantities that we and some other animals can count. 鈥淲e now know how single numbers 鈥� the units of arithmetic 鈥� are represented in our brains, something not known prior to our study,鈥� he says.
鈥淚 think this is very important work,鈥� says Christian Agrillo of the University of Padua in Italy, who has shown that fish can count. 鈥淚t shows that [our] symbolic numerical system is related to the most ancient, non-symbolic one, but evolved in the human brain independently,鈥� he says. 鈥淭his might explain why, although different animals recently showed surprising numerical abilities, the use of verbal labels and complex mathematical abilities are uniquely human.鈥�
Neuron
