演題詳細

New registration

Poster Sessions

神経情報学
Neuroinformatics

print

開催日
2010年09月04日(土)
時 間
13:00 - 14:00
会 場
Poster Room 2

Back

視覚的注意の神経機構を考慮した視覚探索モデル
A visual search simulation by using a top-down feature feedback model

演題番号 : P3-q14

杉本 圭佑 / Keisuke Sugimoto:1 小濱 剛 / Takeshi Kohama:2 

1:近畿大院・生物理工・電子システム情報工 / Grad Sch Biology-Oriented Science and Technology, Kinki Univ, Wakayama, Japan 2:近畿大・生物理工・電子システム情報工 / Fac Biology-Oriented Science and Technology, Kinki Univ, Wakayama, Japan 

 

The activity of the visual nervous system is modulated by top-down attention for visual features such as color or shape. In this phenomenon, Hamker (1994) suggested that the attentional modulation for visual features is generated by a feedback information network of V4, IT, frontal eye field (FEF), and prefrontal cortex by using a model simulation (Hamker F.H.,Vision Res.,44,501-521,2004). However, the simulation condition of this model was assumed to maintain fixation; therefore, this model was unable to reproduce the eye movements required for target selection in the visual search task, which requires attentional modulation for visual features. In this study, we propose a modified version of the Hamker model to reproduce the scan paths in visual search tasks by expanding the model to the peripheral area and considering the inhibition of return (IOR) corresponding to the output of the FEF cells. In particular, we scaled up the Hamker model by 13-15 units so that the eye movements in the peripheral visual area could be reproduced. A maximal response unit in the FEF cells is selected as the location of the gaze point at approximately every 300 ms. A Gaussian weight function is applied to the responses of the FEF cells around the selected fixation point because the spatial resolution deteriorates with increasing retinal eccentricity. After a gaze shift is carried out, recurring selection of the same location as a target point is avoided by the IOR mechanism. The IOR is modeled as a simple first-order system, and it temporarily inhibits the activity of the fixation region units in FEF cells. The searching profiles are modeled by repeating this gaze point decision process with the IOR. The simulated scan paths obtained using our model closely resemble the experimental data obtained in monkeys and show the dynamic processes that occur in the visual nervous system to generate eye movements with attentional modulation for visual features.

Related Link

Back