2701 A model of ventral pathway of temporal lobe reproducing neural responses of superior temporal sulcus

2701 A model of ventral pathway of temporal lobe reproducing neural responses of superior temporal sulcus

S266 27. System theory 2701 Syst. Japan. A MODEL OF VENTRAL PATHWAY OF TZMPORAL LOSE REPRODUCING NEURAL RESPONSES of Intelliqence OF SUPERIOR TEMPO...

70KB Sizes 1 Downloads 51 Views

S266

27. System theory

2701 Syst. Japan.

A MODEL OF VENTRAL PATHWAY OF TZMPORAL LOSE REPRODUCING NEURAL RESPONSES of Intelliqence OF SUPERIOR TEMPORAL SULCUS. KIYOEIKO NAKAMURA, Dept. Enq., Tokyo Znstitute of Tech., Yokohama 227, Sci., Grad. Sch. of Sci.

temporal sulcus(STS) of the monkey discriminate Neurons of the superior onset(Oram and Perrett 1992). This study visual stimuli within 5 ms of response presents a biologically realistic model of ventral pathway of temporal lobe which Results obtained from the model analysis are: reproduces the fast discrimination. First, computer simulation shows difference between cell responses of STS of the model reaches significant level within 5 ms of response onset. Second, the fast mechanism of competition on firing discrimination is produced by the neural synaptic plasticity controlled by attetion/reward system latency. Third, reinforces cortical circuits to work the competition mechanism well. Fourth, response latency of the STS decreases with increase of response magnitude both in predicts. Fifth, the model the experiment and the model, as the ajove mechanism shows the mechanism using neuronal pop,ulations is robust against noise. From these, the hypothesis that the latency competition functions as neural mechanism to perform cortical processing of the rrlllisecond time scale.

2702

Research Japan.

SPATIAL TEXTURE.

FREQUENCY PROCESSING IN KO SAKAI, The Institute

(RIKEN), FRP, Laboratory Email:kol?yugiri.riken.go.jp

for

THE PERCEPTION

OF SHAPE

FROM

of Physical and Chemical Neural Modeling, Wako, Saitama,

I seek to investigate what computational mechanisms are utilized by the visual system to perceive 3D shape and depth from changes in texture. It has been shown that the visual system characterizes the local spatial frequency spectrum by Average Peak Frequency (K. Sakai and L. H. Finkel, J. Opt. Sot. Amer., 1995). I investigate what visual information is utilized and what is discarded by generating a series of stimuli whose frequency spectrum is artificially modified. Based on the psychophysical experiments, I propose a network model for how the cortex discriminates shape and depth from texture. Simulations with a range stimuli, including real images, show qualitative and quantitative agreement with human perception.