Naslov
A model of numerical representation in the brain

Autori
Domijan, Dražen

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
7th Alps-Adria Conference in Psychology, Abstracts / Manenica, Ilija - Zadar : Odjel za psihologiju Sveučilišta u Zadru, 2005

Skup
7th Alps-Adria Conference in Psychology

Mjesto i datum
Zadar, Hrvatska, 02.06.2005. - 04.06.2005

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Nije recenziran

Ključne riječi
numbers; neural model; computer simulation

Sažetak
Investigations with adults, infants, patients and animals point to the existence of the language-independent numerosity processing system. This system is not endowed with abstract algebraic rules but it can solve simple numerosity discrimination tasks and represent evolutionary precursor for more elaborate numerical skills observed in humans. Numerical processing provide a challenge for neural network modeling because general purpose learning algorithms are not well suited for dealing with abstract concepts such as number. Here, a new neural network model is proposed with the ability to extract abstract numerical representation from visual input. It simulates properties of a numerosity detection system which is hypothesized to underlie simple language-independent numerical abilities. The network has three layers where the first layer computes the sum of the nearest neighbour inputs. The first layer is also augmented with multiplicative gating and gradient tonic activation which prevents interference. The second layer implements local lateral inhibition which enables a single node to represent a single object. The third layer exhibits number-tuning similar to recently described responses of neurons in the prefrontal cortex. Computer simulations showed that network response does not depend on visual attributes like the object’ s size, position or shape. The model exhibits numerical distance effect and numerical size effect consistent with psychophysical findings. This is demonstrated by tuning curves of nodes in the third layer which are overlapping with maximal response for certain numerosity and lower response for adjacent numerosities. Width of the tuning curve increased with increasing numerosity. The model is based on several biophysical mechanisms such as multiplicative interaction on dendrites, independent processing on different dendritic branches and disinhibition by glutamate spill-over on kainate receptors on inhibitory axons.

Izvorni jezik
Engleski

Znanstvena područja
Psihologija

POVEZANOST RADA