一般口演（Oral）

Oral Vision-Ⅱ 一般口演視覚-Ⅱ

7月25日（木）10:00～10:15　第7会場（朱鷺メッセ　2F　201B） 1O-07m2-1 ラット視覚領域の視覚検出課題時における神経表象 Yuma Osako（大迫優真），Tomoya Ohnuki（大貫朋哉），Yoshio Sakurai（櫻井芳雄），Junya Hirokawa（廣川純也）同志社大学大学院脳科学研究科神経回路情報伝達機構部門 Previous studies demonstrated that primary visual cortex (V1) and posterior parietal cortex (PPC) are involved in visually-guided decision making, including processing of visual sensory inputs, accumulation of evidences and transformation of visual sensory signals into accurate motor plans. However, since there were no behavioral paradigms in rodents that distinguish visual detection performance from decision performances, specific roles of these visual cortical areas in visually-guided decision making remain unclear. In our previous study, we established a quantitative behavioral paradigm in rodent to evaluate subjective decision threshold for accurate visual detection performance. Rats were trained to detect identical sets of spatial visual stimuli with and without a third choice option, which allows us to distinguish behavioral responses with and without subjective visual guidance. Here, we recorded neural activity in primary visual cortex (V1) and posterior parietal cortex (PPC) during visually-guided decision making. While V1 population was predominantly selective for visual stimulus rather than visual detection performance (43 % for visual stimulus, 11 % for visual detection performance), PPC population was selective for both visual stimulus and visual detection performance in equal proportion (26 % for visual stimulus, 23 % for visual detection performance). In PPC subpopulation, visual responses were systematically delayed with reaction timing but it was not aligned to movement. This suggests that PPC subpopulation is responsible for subjective detection threshold rather than decision threshold.		7月25日（木）10:15～10:30　第7会場（朱鷺メッセ　2F　201B） 1O-07m2-2 方位弁別時に機能する低コントラスト優位な一次視覚野細胞の生成メカニズム Rie Kimura（木村梨絵）^1,2，Yumiko Yoshimura（吉村由美子）^1,2 ¹生理研基盤神経視覚情報処理 ²総研大院生命科学生理 Animals can often perceive vague visual stimuli. To explore the neural bases, we analyzed neural activities in rat primary visual cortex (V1) during an orientation discrimination task using high-contrast or low-contrast visual stimuli. We performed multiple single-unit recordings from deep layers of V1. We observed high contrast-preferring neurons in which the firing rates during visual stimuli decreased with a reduction of the contrast. In addition, we found low contrast-preferring neurons in which the firing rates increased with a reduction of the contrast. The low-contrast preference was commonly observed in both wide-spiking (putative excitatory) and narrow-spiking (inhibitory) neurons. And, the firing rates in low contrast-preferring neurons increased more in correct-choice trials than in incorrect trials during discrimination at low contrast. We explored the mechanisms for generating low-contrast preference. The low-contrast preference was rarely observed in awake rats before learning with passive viewing, demonstrating that learning is required for the generation of low-contrast preference. The amplitude of visually-evoked field potentials for high- or low-contrast stimuli increased after learning, suggesting that bottom-up inputs were enhanced after learning irrespective of stimulus contrast. Spike synchronization during high-contrast stimuli, but not during low-contrast stimuli, was enhanced after learning, and this enhancement was more remarkable in narrow-spiking neurons than in wide-spiking neurons, suggesting that strong inhibition is driven by high-contrast stimuli after learning. Thus, after learning, net excitation during low-contrast stimuli increased more than that during high-contrast stimuli, leading to reliable signal transmission with low-contrast stimuli. In addition, the basal firing rate in low contrast-preferring neurons increased after learning, whereas the increase did not occur in high contrast-preferring neurons. Therefore, low-contrast preference may be explained by an increase in basal activity working together with reliable transmission with low-contrast stimuli observed after learning. Previous reports demonstrate that top-down inputs remarkably decrease under anesthesia. We did not detect low contrast-preferring neurons in anesthetized rats after learning. Thus, the top-down inputs to V1 might be one of the possible input sources for the generation of low-contrast preference in V1.

7月25日（木）10:30～10:45　第7会場（朱鷺メッセ　2F　201B） 1O-07m2-3 マカカ属サル視覚野におけるGCaMP6sを利用したカルシウムイメージングおよび内因性光学信号計測の適用 Gaku Hatanaka（畑中岳）¹，Yang Fang（Fang Yang）^1,2，Mikio Inagaki（稲垣未来男）^1,2，Ryosuke F Takeuchi（竹内遼介）¹，Ken-ichi Inoue（井上謙一）³，Masahiko Takada（高田昌彦）³，Ichiro Fujita（藤田一郎）^1,2 ¹大阪大院生命機能 ²脳情報通信融合研究センター ³京都大霊長研 Distribution patterns of neural preferences for visual stimulus features determine functional architecture of a cortical area. Elucidation of the architecture of a cortical area provides a critical clue to understanding of the neural processing and representation there. Previous studies investigated functional architectures with several imaging techniques such as two-photon calcium imaging, intrinsic optical signal imaging, and functional magnetic resonance imaging. These techniques probe functional structures in different scales. It was challenging to relate the structures determined with the different techniques. Recent development of virus vectors designed for calcium-sensitive proteins makes it possible to examine the same region with both wide-field imaging and focused, cellular-resolution imaging as well as to monitor neural activity over a prolonged period. These advantages are not yet fully enjoyed in the studies of the macaque cortex. Here we injected adeno-associated virus vector expressing the gene for a calcium indicator (GCaMP6s) under the regulation of a promoter specific for cortical excitatory neurons, CaMKIIα, or a pan-neuronal promoter, Syn, into V1 and V2 of a Japanese monkey (Macaca fuscata). We conducted wide-field imaging and two-photon imaging in the same local regions of areas V1 and V2. In the experiments, the monkey was anesthetized with Propofol and immobilized with vecuronium bromide. We first imaged large-scale functional architectures in V1 and V2 by using intrinsic optical signal imaging and wide-field calcium imaging. Visual responses recorded with wide-field calcium imaging locally (˜2 mm) emerged only at the virus injection sites. Orientation maps and color blob patterns in V1 and orientation specific stripes in V2 in these vector-labeled regions were spatially matched to those visualized with intrinsic optical signal imaging. By referring to the wide-field functional maps and to the vascular pattern over the cortex, we aimed the field of view of our two-photon microscope to a specific target region such as a pinwheel center in an orientation map. We successfully recorded robust (ΔF/F: ˜200%) calcium responses of individual neurons in these targeted regions. Combined application of multiscale calcium imaging with GCaMP6s and intrinsic optical signal imaging opened a way to relate the cellular arrangements at different spatial scales in the visual cortical areas of the macaque monkey.		7月25日（木）10:45～11:00　第7会場（朱鷺メッセ　2F　201B） 1O-07m2-4 Characterization of Stereoscopic Information Processing in Human Brain using Reversed-Depth Perception Stimulus Bayu Gautama Wundari（Wundari Bayu Gautama）¹，Hiroshi Ban（Ban Hiroshi）^1,2，Ichiro Fujita（Fujita Ichiro）^1,2 ¹Graduate School of Frontier Biosciences, Osaka University, Osaka, Japan ²Center for Information and Neural Networks (CiNet), NICT, Osaka, Japan Our visual system extracts depth information by finding the correct match between images projected onto the two retinas. A good foundation has been established regarding how the brain handles this so-called binocular correspondence problem in the primary visual cortex (V1), yet, its hierarchical mechanism for solving this problem remains elusive. An engineered random dot stereogram (RDS) that has anti-correlated RDS (aRDS) surrounded by correlated RDS (cRDS) can generate reversed-depth perception, in which observers perceive "near" for uncrossed disparities and "far" for crossed disparity. Reversed-depth is supposed to be mediated by the neural system that is sensitive to binocular disparity in aRDSs with inverted tuning. Thus, aRDSs can be employed to reveal the neural substrates that retain binocular correlation signals and support reversed-depth perception. We performed visual psychophysics to measure the performance of near-far judgment of 23 human observers using the same stimuli. The results suggested that they can be grouped based on their reversed-depth discernability. We further measured the brain activities of the same subjects with functional magnetic resonance imaging (fMRI, voxel size = 2 mm isotropic, TR = 2 s, horizontal (slightly oblique along the AC-PC line) 78 slices with multi-band factor = 3). We used a 16-s block design scheme for the fMRI experiments (stimulus duration = 1 s, followed by gray background screen for another 1 s, repeated 8 times in a block). Throughout the fMRI scanning, the subjects observed RDSs that had center aRDS with varied dot-contrast match levels (anti-correlated, half-correlated, and full-correlated) surrounded by cRDS. Regions of interest (ROIs) were defined based on the standard phase-encoded retinotopy method. Voxels in each ROIs were sorted in a descending order based on general linear model t-values (fixation vs. all) for further analysis with multi-voxel pattern analysis. A linear support vector machine (SVM) was used to classify near-far disparity of the viewed stimulus from the sorted voxels. Representational similarity analysis with the previously proposed model (Doi et al., 2011) revealed that responses in early visual areas (V1, V2, and V3) and dorsal brain areas such as V3A and V7 were mostly accounted for by correlation computation. In contrast, matching and correlation computation comparably contributed to responses in a higher brain area in the ventral stream, hV4.