A neurocomputational account of mental curve tracing (CROSBI ID 619240)
Prilog sa skupa u zborniku | sažetak izlaganja sa skupa | međunarodna recenzija
Podaci o odgovornosti
Domijan, Dražen
engleski
A neurocomputational account of mental curve tracing
A neural model is proposed whose temporal dynamics simulates the properties of mental curve tracing. Behavioral studies revealed that the tracing is a time consuming process that serially spreads attentional label along the target curve. However, the speed of tracing is not fixed and it could be flexibly adjusted depending on the density of image elements. Single-unit recordings in the monkey primary visual cortex showed that tracing is associated with elevated firing rate for neurons whose receptive field falls on the traced curve. In order to explain behavioral and neurophysiological findings, the proposed model implements a novel form of neural filling-in that enables activity spreading along the target curve. Filling-in occurs at multiple spatial scales in order to account for different speeds of tracing. The model implements object-level competition and selection among distinct image elements. Computer simulations showed that the model exhibits appropriate scaling of tracing speed with distance between curves thus simulating the hot spots of attention. The speed of tracing slows down when the curves are close to each other and speeds up when they are far apart. Furthermore, the model is able to store the traced pattern in short-term memory.
pažnja; praćenje konture; neuronski model
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
Podaci o prilogu
148-x.
2012.
objavljeno
Podaci o matičnoj publikaciji
Perception
Pion
Podaci o skupu
35th European Conference on Visual Perception
poster
02.09.2012-06.09.2012
Alghero, Italija