演題詳細
Symposium
注意の脳内ネットワーク
Network of attention in human and macaque
開催日 | 2014/9/13 |
---|---|
時間 | 17:10 - 19:10 |
会場 | Room C(502) |
Chairperson(s) | 吉田 正俊 / Masatoshi Yoshida (自然科学研究機構 生理学研究所 / Department of Developmental Physiology, National Institute for Physiological Sciences, Japan) Ziad Hafed (Physiology of Active Vision, Centre for Integrative Neuroscience, University of T2014010063ebingen, Germany) |
視覚サリエンシーに関わる脳内ネットワーク
Brain network for visual saliency
- S3-C-3-4
- 吉田 正俊 / Masatoshi Yoshida:1,2
- 1:自然科学研究機構 生理学研究所 / Dept.of Developmental Physiol., Natl Inst.for Physiol.Sci., Okazaki, Japan 2:総合研究大学院大学 生命科学研究科 / Sch. Life Sci., Grad. Univ. Adv. Stud., Hayama, Japan
Visual saliency is a stimulus-driven property that drives attention by announcing that certain location is different from its surroundings. Several studies reported neurons encoding saliency in various brain regions but it remain elusive which part of the brain is necessary for processing saliency. Here we examined this question using an animal model of blindsight.
Blindsight occurs in patients with damage to primary visual cortex (V1) who demonstrate residual performance on visual tasks despite denial of conscious seeing. We examined whether the monkeys with damage in V1 retain computation of visual saliency. More specifically, we examined whether orienting attention toward salient stimuli during free viewing is still possible. Despite general deficits in gaze allocation, monkeys were significantly attracted to salient stimuli. Our results show that attention guidance over complex natural scenes is preserved in the absence of V1, thereby directly challenging theories and models that crucially depend on V1 to compute visual saliency.
One candidate for saliency computation in the absence of V1 is the superior colliculus (SC). A recent electrophysiological study using acute slices of mouse SC (Phongphanphanee et al, 2014) suggest that the visual layers of SC (SGS) had local excitation and distal inhibition, thus comprising a Mexican-hat response function which is ideal for edge detection and saliency computation. We used a computational method to examine how such saliency detector is implemented in SGS. We constructed a full-scale spiking neural circuit of SGS horizontal slice and estimated the anatomical spreads and synaptic dynamics parameters of the neural populations in SGS. Despite the high-dimensional space of the parameters, we were able to find good fits for reconstructing the Mexican-hat function in SGS. These results suggest that the horizontal interaction is the key factor to implement saliency detector in SC.
Altogether, these results suggest that combination of behavioral or electrophysiological data with computational neuroscience is a promising strategy to understand how visual saliency is computed in a brain network.