視覚 Visual System
P1-1-95 哺乳類のUV光センサー蛋白質OPN5の機能解析 Functional analysis of mammalian UV photoreceptor OPN5 ○小島大輔1,2, 森卓1, 鳥居雅樹1, 深田吉孝1 ○Daisuke Kojima1,2, Suguru Mori1, Masaki Torii1, Yoshitaka Fukada1 東京大・院理・生物化学1, 科学技術振興機構・さきがけ2 Dept Biophys Biochem, Grad Sch Sci, Univ Tokyo, Tokyo, Japan1, PRESTO, Japan Science and Technology Agency, Japan2 The ultraviolet (UV) component of sunlight is utilized as the environmental cues by a variety of animal species such as insects and birds, whereas physiological roles of UV photoreception in mammals, especially in humans, remain open questions. We have previously reported that neuropsin (OPN5) is the hitherto unknown UV photoreceptor protein in humans and mice [Kojima et al. (2011) PLoS ONE 6, e26388]: The mammalian OPN5 protein exhibits an absorption maximum at 380 nm when reconstituted with 11-cis-retinal. Upon UV illumination, the UV photopigment is converted to a longer-wavelength-absorbing photoproduct, which is stable in the dark and reverted to the UV-absorbing state by the subsequent orange light illumination, indicating its bistable nature. OPN5 is capable of activating heterotrimeric G protein Gi in a UV-dependent manner. In mice, OPN5 is expressed in the retina, brain and outer ears as the major expression sites. In these tissues, OPN5 is localized in a subset of non-rod/non-cone retinal neurons as well as in the epidermal and muscle cells of the outer ears, most of which also express the alpha-subunit of Gi. It is strongly suggested that OPN5 triggers a UV-sensitive Gi-mediated signaling pathway in the mammalian tissues. In this study, we further investigated the OPN5 signaling pathway by establishing real-time monitoring systems for OPN5-expressing cells. The results demonstrate that OPN5 can regulate intracellular concentration of Ca++ as well as cAMP in UV-dependent manners.	P1-1-96 ラット一次視覚野におけるコリン作動性システムの機能的役割 Functional role of cholinergic system in the rat primary visual cortex ○相馬祥吾1, 七五三木聡1,2, 末松尚史1, 佐藤宏道1,2 ○Shogo Soma1, Satoshi Shimegi1,2, Naofumi Suematsu1, Hiromichi Sato1,2 大阪大院・生命機能・認知行動1, 大阪大院・医・認知行動2 Grad Sch Front Biosci, Osaka Univ, Osaka1, Grad Sch Med, Osaka Univ, Osaka2 Acetylcholine (ACh) is known to modulate neuronal activity in the rodent primary visual cortex (V1). We recently examined effects of a microiontophoretic administration of ACh in V1 of anesthetized rats, finding that ACh facilitated or suppressed visual responses to varying stimulus contrasts by multiplying the control responses, i.e. response gain control. This result raise important questions: 1) why opposing effects are concomitantly observed in a cortical area, 2) how interlaminar circuitry is modulated by ACh, and 3) whether ACh improves the visual contrast detection ability. To examine these points, we combined extracellular multi-unit recordings and topical administration of ACh, and measured V1 neuronal responses to drifting sinusoidal grating stimuli in anesthetized rats. We confirmed that ACh changes the response gain upward or downward in facilitated or suppressed cells, respectively. These ACh effects showed a laminar bias, where the response suppression and facilitation prevailed in layers 2/3 and layer 5, respectively. Next, we examined ACh effects on the signal-to-noise (S/N) ratio and the grating-phase information calculated as F1/F0 ratio. In facilitated cells, ACh improved the S/N ratio, while in suppressed cells it enhanced the F1/F0 ratio without any concurrent reduction in the S/N ratio. These effects on S/N and F1/F0 ratios were observed in regular-spiking cells, but not in fast-spiking cells. Our findings suggest that ACh promotes the signaling of grating-phase information from supragranular cells to higher-order areas by the suppressive modulation, and enhances feedback signals with a high S/N ratio from infragranular cells to subcortical areas by the facilitatory modulation. Thus, ACh distinctly and finely controls visual information processing in a manner that is specific for the modulation and cell types and is also laminar dependent. We will present the net effect of ACh on contrast detection ability in behaving rats.
P1-1-97 ヒト背側経路における道具刺激の無意識的処理回路 Neural circuitry to process invisible tool images in the human dorsal pathway ○野口泰基1, 鈴木恵美1, 柿木隆介1,2 ○Yasuki Noguchi1, Megumi Suzuki1, Ryusuke Kakigi1,2 神戸大学大学院　人文学研究科　心理学1, 生理学研究所　統合生理　感覚運動調節2 Department of Psychology, Kobe University, Kobe1, Department of Integrative Physiology, National Institute for Physiological Sciences2 The primate visual system is assumed to comprise two main pathways: a ventral pathway for shape and color perception and a dorsal pathway for spatial processing and visuomotor control. Previous studies consistently reported strong activation in the dorsal pathway induced by visible images of manipulable object such as tools. However, it is controversial whether the dorsal pathway retains this preferential activity to tool images under unconscious perception. In the present study, we used magnetoencephalography (MEG) and investigated spatio-temporal dynamics of neural responses to visible and invisible tool images. A presentation of visible tool images elicited neural responses in the visual cortex mainly in the right hemisphere, followed by larger activities over the posterior parietal regions in the left hemisphere. This sequence of the neural processing from the right to the left hemispheres was still observed when conscious perception of tool images was inhibited by interocular suppression. Those results demonstrated that the neural circuit to process the tool information was preserved under unconscious perception, highlighting an implicit aspect of the dorsal pathway.	P1-1-98 サル網膜において杆体・錐体双方から基底型シナプス結合を介して入力する双極細胞型とマウス網膜における相当する細胞型との比較 A type of bipolar cell receives rod as well as cone input by basal contacts in the monkey retina and its comparison to the corresponding types in the mouse retina ○塚本吉彦1,2, 臣尚子1 ○Yoshihiko Tsukamoto1,2, Naoko Omi1 スタジオレチナ1, 兵庫医科大学2 Studio Retina1, Hyogo College of Medicine, Nishinomiya, Japan2 Whether there are OFF cone bipolar cells that directly receive chemical synaptic input from rod photoreceptors in the primate retina has remained unsolved since such bipolar cell types were discovered in mouse, rat, squirrel, and rabbit retinas. We identified at least one type of OFF cone bipolar cell that made basal contacts to rods, by using serial electron microscopy at the rod-richest area (Eccentricity: 3.06-3.22mm temporal) in the Japanese monkey (Macaca fuscata) retina. In the mouse retina, five types (1, 2, 3a, 3b, and 4) of OFF bipolar cell were identified (Wassle, et al., 2009). Three (3a, 3b, and 4) of them were immunocytochemically (Mataruga et al, 2007; Haverkamp et al., 2008) revealed to make basal contacts to rods as well as cones. We reconstructed three cells for each type in 3D space and examined their synaptic connectivity. In the monkey retina, four types (DB1, FMB, DB2, and DB3) of OFF bipolar cell were known (Boycott and Wassle, 1991). We further classified DB2 into DB2a and DB2b, because DB2a dendrites had basal contacts to rods but DB2b did not, and the axon terminal of DB2a stratified slightly upper than that of DB2b. This DB2a type had basal contacts to both rods and cones. The basal contacts of DB2a were less dense than rods and often spaced with one another by a few unconnected rods, thus they covered an array of rods intermittently. Outputs of DB2a were mainly directed at parasol ganglion cells. A comprehensive catalog of monkey bipolar cells has remained elucidated, but at least type DB2a bipolar cells are involved in carrying OFF rod signals. This newly-discovered direct pathway from rods to OFF cone bipolar cells in the primate retina provides a possibility to explain a duality between the sensitive-slow and the insensitive-fast rod signals in the human visual system (Sharpe and Stockman, 1999)
P1-1-99 定常状態視覚誘発電位による物体認識の調査 Investigation of object recognition using steady state visually evoked potential ○東和樹1, 南哲人2, 中内茂樹1 ○Kazuki Azuma1, Tetsuto Minami2, Shigeki Nakauchi1 豊橋技科大情報知能工学1, 豊橋技科大エレクトロニクス先端融合研2 Department of Computer Science and Engineering, Toyohashi University of Technology, Toyohashi, Japan1, EIIRIS, Toyohashi Univ of Tech, Toyohashi, Japan2 Recent researches revealed that the EEG component caused by flickering visual stimulus, which is called steady state visually evoked potential (SSVEP), might be a potential index for object recognition. In this study, We examined whether SSVEP reflects different states during object recognition. In one trial, binary image (BI) which is difficult to recognize was followed by a gray scale image (AI) of the same object as the answer. Both BI and AI were presented in flickering manner at frequency of 7.5 Hz. Participants were first asked to answer whether they could recognize BI. Then, after AI was presented, participants had to answer whether they had correctly recognized BI. EEG recorded during BI and AI presentation was classified into 2 conditions (recognized/unrecognized after BI presentation and recognized before/after AI presentation) according to participants' behavioral responses. SSVEP from the two recognition conditions was compared to see whether SSVEP could reflect different states of object recognition. SSVEP for unrecognized BI was stronger and the middle frontal gyrus responsible for working memory access was more activated. This might result from activation of theta activity during memory search which subsequently amplified the 7.5 Hz activity in SSVEP. On the other hand, when participants could correctly recognize objects in the images after AI presentation, brain activity was suppressed and the activities were localized in the middle temporal gyrus and middle occipital gyrus. A previous study suggested that suppression of the 6-8 Hz activity in the occipito-parietal area reflects the semantic processing. Therefore, the decrement of SSVEP found in this study might be a result semantic memory retrieval. In conclusion, this study suggests that SSVEP reflects the different recognition states and shows the possibility that SSVEP might be a potential tool for extraction of human's introspective information.	P1-1-100 マウス一次視覚野において方位選択性のバイアスはトポグラフィックに構成される Biases of orientation preference are topographically organized in mouse primary visual cortex ○村上知成1, 吉田盛史1, 大木研一1 ○Tomonari Murakami1, Takashi Yoshida1, Kenichi Ohki1 九州大学大学院　医学系研究科　分子生理学1 Dept Molecular Physiology, Kyushu Univ, Fukuoka, Japan1 Orientation selectivity and receptive fields are fundamental features of neurons in visual cortex, and receptive fields are organized into retinotopic maps of the visual field. Visual cortex in higher mammals such as carnivores and primates has the functional orientation columns, and represents biased sensitivity to radial orientations depending on retinotopic positions in visual fields. Thus, there is systematic relationship between orientation bias and retinotopic map in visual cortex with functional columns. In rodent, their neurons in V1 are organized in a salt-and-pepper fashion for orientation selectivity, and there is no functional column. Is there any relationship between orientation bias and retinotopic map in mouse visual cortex? To examine this possibility, we observed the activity in the entire mouse V1 with intrinsic optical imaging. We identified the area of V1 by horizontal and vertical retinotopic mapping and measured orientation selectivity driven by grating stimulation drifting to 8 directions. We found that there was a robust topographic map of orientation biases in V1. Although rodent V1 does not have orientation column, when viewed on the whole, there is a characteristic structure in mouse V1; the orientation preference in peripheral visual fields was biased toward particular orientation, depending on the position in V1. This finding suggests the possibility that the organization of orientation bias in mouse V1 is similar to that in higher mammals.
P1-1-101 2光子イメージングを用いたマウス上丘の側方抑制を示す細胞集団活動の記録 Two photon population imaging for surround suppression im mouse superior colliculus ○笠井昌俊1, 伊佐正1 ○Masatoshi Kasai1, Tadashi Isa1 生理学研究所　発達生理学研究系　認知行動発達機構研究部門1 Dept Developmental Physiol, NIPS, Okazaki1 The superior colliculus (SC) is a brainstem center which plays a key role in mediating the signal for sensory-motor translation. The superficial layer of the SC is innervated by the optic tract and visual space is represented in the retinotopic coordinates. In the visual pathway as well as other sensory system, firing activity of neurons in response to the stimuli presented in their response field are often inhibited by stimuli presented outside their receptive field. These effects are known as "surround suppression" or "lateral inhibition" and thought to enhance the contrast of sensory information. In the SC, horizontal inhibitory connection which could be related to the lateral inhibition has been studied in individual cell level; however, it has not been studied in the larger cell population. Therefore, we examined how the lateral inhibition in the SC is organized in the neuronal population level and tried to elucidate its micro-circuit mechanism, by applying an in vivo two-photon Ca2+ imaging technique in anesthetized mice. We could recorded from more than one hundred of neurons simultaneously during presentation of light spots with different sizese. We confirmed that in the superficial layer of the SC, if the visual stimuli were applied in the center of receptive field of several cells, the larger the stimuli size, the less the neural activity of these cells. Thus, we succeeded in establishing the experimental preparation in which we can analyze the neural mechanis of lateral inhibition in the neuronal population of the SC superficialy layer.	P1-1-102 均一な方位選択性分布を示すニューロン集団の記録に適した多電極アレイのデザイン Design of multi-electrode array for homogeneous samplings of differently orientation tuned unit population ○圓山由子1, 伊藤浩之1 ○Yoshiko Maruyama1, Hiroyuki Ito1 京都産業大学大学院工学研究科情報通信工学専攻1 Dept.of Inf.& Commun.Sci., Fac.of Eng., Kyoto Sangyo Univ., Kyoto, Japan1 We are recording multiple single units simultaneously within a small region as a single hyper column in cat visual cortex. We attempt to predict an orientation of light bar stimulus based on multineuron data of a single trial. For a better prediction to any orientation, we need reasonably homogeneous samplings of differently orientation tuned units in entire orientation range. In this study, we used two kinds of multiple electrodes array (4-tetrodes array & 8-micro electrodes array) and examined what kinds of electrodes and their arrangements provide more homogeneous distribution of units' optimal orientations. We quantified homogeneity in the distribution of optimal orientations over all units by circular variance (CV), which takes 1 in case of completely homogeneous distribution on a circle and 0 when all units has the same orientation. We found that unit population isolated by 4-tetrodes array was likely to show larger CV (N =25, median 0.65) than those by 8-micro electrodes array (N =22, median 0.50). However the difference between the two distributions did not reach to significance (P >0.1) due to small sample size. To confirm the difference, we simulated in-virtu recordings on the optical imaging data (Tani & Tanaka, 2008). We chose 2000 recording sites randomly, set positions of electrodes in each array, assigned optimal orientations of units according to the orientation map, and then compute CV. The number of units isolated by each electrode was set following to the probability distribution obtained by experiment. As a result, CVs of 4-tetrodes array were significantly larger (median 0.58) than those of 8-micro electrodes array (median 0.52) (P <0.0001), which were consistent with our experimental result. We conclude that since tetrode samples units in relatively wide area, heterogeneity of orientation map (pin-wheel) leads to large CV. Furthermore, a lower probability of recording a unit by micro electrode (0.5) leads to smaller CV.
P1-1-103 大脳皮質第5層錐体細胞のユニット状活動を担う神経回路 Neuronal circuit underlying unitary activity of pyramidal neurons in neocortical layer V ○鶴野瞬1, 丸岡久人1, 黒川留美1, 松本直実1, 木曽かおり1, 細谷俊彦1 ○Shun Tsuruno1, Hisato Maruoka1, Rumi Kurokawa1, Naomi Matsumoto1, Kaori Kiso1, Toshihiko Hosoya1 （独）理研 BSI1 RIKEN BSI, Wako1 A major question in neocortical research is the extent to which neuronal organization is stereotyped. We found that the tangential arrangement of subcerebral projection neurons (SCPNs), which are the major pyramidal neuron subtype in mouse layer 5, was not random but significantly periodic (Maruoka et al., 2011). This periodicity was observed in multiple cortical areas and had a typical wavelength of 30 μm. Under specific visual stimulation, neurons in single repeating units exhibited strongly correlated expression of c-Fos, a marker of activated neurons. Therefore, SCPNs have a periodic arrangement, and neuronal activity leading to c-Fos expression is similar among neurons in the same repeating units. Here, we studied the neuronal mechanisms underlying the unitary activity of SCPNs. We focused on two types of network structures: mutual connections between SCPNs and common inputs to SCPNs. To characterize these structures we performed whole-cell patch clamp recordings from multiple SCPNs in acute brain slices prepared from the mouse visual cortex. To investigate mutual connections between SCPNs we calculated the probability and strength of monosynaptic excitatory and disynaptic inhibitory connections. To characterize common inputs to SCPNs, coincident synaptic inputs were analyzed. These structures were then compared between intra-units and inter-units to investigate whether neurons within single repeating units have specific circuits. Our results showed that strength of mutual connection was not significantly different between intra-units and inter-units. On the other hand, large coincident inputs were more often found for SCPN pairs located in the same units than those located in adjacent units, suggesting SCPNs in the same unit tend to receive large common inputs. These results suggest that large common inputs to single repeating units might underlie the unitary activity of SCPNs. We are investigating the origin of large common inputs using optogenetics.	P1-1-104 下側頭皮質に表現されている知覚光沢パラメータ：正準相関解析による神経データ解析 Perceptual gloss parameters represented in inferior temporal cortex: a canonical correlation analysis of neuronal data ○下川丈明1, 西尾亜希子2, 郷田直一2,3, 小松英彦2,3 ○Takeaki Shimokawa1, Akiko Nishio2, Naokazu Goda2,3, Hidehiko Komatsu2,3 , 生理学研究所2, 総研大・生命科学・生理3 ATR-NIA, CBI, Kyoto, Japan1, National Institute for Physiological Sciences, Okazaki, Japan2, The Graduate University for Advanced Studies (SOKENDAI), Okazaki, Japan3 A part of optical reflectance properties of material surface is recognized as gloss. The following three physical parameters are known to be the main parameters relevant to the gloss: the diffuse reflectance, the specular reflectance, and the spread of specular reflection. It is recently found that there are neurons in the monkey inferior temporal cortex that selectively respond to these parameters [1]. In this study, we analyzed the nonlinear relationship between the reflectance parameters and the responses of gloss selective neurons [1] by canonical correlation analysis (CCA). CCA identifies the relationship between two populations of variables. Here, one population is the set of the three reflectance parameters and the other population is the set of neuronal responses.First, we found that the perceptual gloss parameters proposed by earlier psychophysical study [2] need to be revised to explain the neuronal data. The contrast gloss space was enlarged in dark condition and shrunk in bright condition. Second, the correlation coefficient between the parameters of the physical reflectance and those of neural responses was very high. It exceeds 0.95 in all three dimensions. This indicates that the gloss space is definitely represented in the cortex like the color space.[1] A. Nishio et al, J Neurosci 32 (2012) 10780; A. Nishio et al, the 34th Annual Meeting of the Japan Neuroscience Society, O3-D-3-1.[2] J. Ferwerda et al, Proc. SPIE Human Vision and Electronic Imaging (2001) 291.
P1-1-105 サル視覚野における素材情報処理過程：fMRI研究 Material information processing in monkey visual cortex: a functional MRI study ○郷田直一1,2, 橘篤導1, 岡澤剛起1,2, 小松英彦1,2 ○Naokazu Goda1,2, Atsumichi Tachibana1, Gouki Okazawa1,2, Hidehiko Komatsu1,2 生理研1, 総研大2 National Institute for Physiology, Okazaki, Japan1, SOKENDAI, Okazaki, Japan2 We can effortlessly identify and categorize materials at a glance. Little is known, however, about how visual information about materials is processed in the brain. We recently found that the way visual information is coded in human visual cortex changes from an image-based representation in early visual areas to a representation reflecting material perception in the ventral higher-order visual areas (Hiramatsu, Goda & Komatsu, 2011). Here we extend our study to explore how various brain regions in monkeys represent information about materials. Two awake macaque monkeys were scanned with 3T MRI while viewing images of materials from nine categories (metal, wood, fur etc.) that were identical to those used in human study. Neural dissimilarity between categories was computed from the activities in various brain regions by using multi-voxel pattern analysis, and was compared with image-based and perceptual measures of dissimilarities between categories. Image-based dissimilarity was obtained from low-level image features such as spatial frequency and color histogram, and perceptual dissimilarity was obtained from human psychological experiments. We found that, as in humans, neural dissimilarities in monkey V1 and V2 were highly correlated with the image-based dissimilarity. Interestingly, neural dissimilarities in monkey V3, V4 and posterior inferior temporal cortex (IT) were correlated with perceptual dissimilarity, although there were some differences in the representational structures between the activities in these monkey areas and human material perception. Our results suggest that visual information about materials is processed along the ventral visual pathway, and V3, V4 and posterior IT play important roles for constructing meaningful information of materials. Our data also give new insight into functional homology between human and monkey visual areas.	P1-1-106 ショウジョウバエにおける低レベルノイズが負荷された動き刺激に対する視運動反応の増強 Enhancement of optmotor response to visual stimulus including low level of noise in Drosophila melanogaster ○鈴木力憲1, 永浦翔大2, 青西亨1, 宮川博義2, 森本高子2 ○Yoshinori Suzuki1, Shota Nagaura2, Toru Aonishi1, Hiroyoshi Miyakawa2, Takako Morimoto2 東工大院総合理工1, 東薬大生命2 Interdisciplinary Graduate School of Science and Engineering, Tokyo Tech, Kanagawa, Japan1, School of Science, Tokyo Univ of Pharmacy and Life Science, Tokyo, Japan2 Animals perform various behaviors according to the integration of information from the outside world. Information from the world is often masked with noise, therefore brain must process such unreliable information. In general, noise interferes with the signal flow, and as a noise level increases, the precision of signal detection decreases. Recently, however, it has been reported that the addition of an appropriate level of noise improves the ability to detect signals in some biological systems, including sensory and motor systems in animals including humans.In our previous study, we have investigated the effects of noise in visual stimulus on motion detection of Drosophila by measuring the optomotor response to moving stripes. We found that in some cases examined in wild type CantonS (CS), the response was enhanced by the addition of low level of noise. Moreover, this enhancement was absent in mutants which were defective in attention-like behavior. These results may provide the hypothesis that the enhancement of the response by noise addition is associated with the brain states, that is, changes in global brain activity such as sleep-wake transitions. To prove this hypothesis and further uncover the neural mechanism underlying the enhancement of optomotor response, we have focused on dopamine, which is well known as one of neuromodulators that is effective in extensive regions of brain. It has been also reported that dopamine regulates sleep-wake transition in Drosophila. In this study, we used CS flies feeding on food with L-Dopa after eclosion, which increased the dopamine level in flies (L-Dopa CS). We found that the enhancement of the optomotor response by the addition of noise was absent in L-Dopa CS. This result suggests that dopamine system, a potential modulators of brain states, underlies the process regulating the optomotor response to stimulus including noise.
P1-1-107 後外側核ニューロンの視床皮質投射：ウイルスベクターを用いた単一ニューロンの形態学的解析 Thalamocortical projections of the lateral posterior nucleus: a single-neuron tracing study using viral vectors ○中村悠1, 日置寛之1, 古田貴寛1, 金子武嗣1 ○Hisashi Nakamura1, Hiroyuki Hioki1, Takahiro Furuta1, Takeshi Kaneko1 京大・医・高次脳形態1 Dept Morphol Brain Sci, Kyoto Univ, Kyoto1 The dorsal lateral geniculate (LGd) and lateral posterior thalamic nuclei (LP) are known to receive massive afferent inputs from retinal ganglion cells (RGC) and the superficial layer of the superior colliculus, which is also the target of the RGC. The LGd sends fibers predominantly to the primary visual cortex (V1), whereas projections of the LP terminate in the secondary visual cortex (V2), and other cortical areas. Although thalamocortical axons of LGd neurons have been studied at a single neuron level, those of LP neurons have not been elucidated yet. In the present study, we investigated complete axonal arborizations of single LP neurons in the rat brain. We first divided the LP into the caudal (LPc) and rostral (LPr) portions according to immunoreactivity for the calbindin D-28k. We then labeled thalamocortical neurons in the LP by injecting the Sindbis virus expressing membrane-targeted green or red fluorescent proteins. The single neurons in the lateral part of the LPr densely projected to layer 2 - 5 of the V2. Most of them (four of five) formed 2 or 3 axonal clusters within the V2. The neurons in the central part of the LPr mainly terminated in both the V1 and V2. Two of three neurons in this part projected with wider axonal arborization than that of the other LP neurons examined in the present study. In contrast to the lateral and central parts of LPr, LPc and medial LPr neurons sent fewer axon collaterals to the V1 and V2. The LPc neurons mainly projected to the temporal association (TeA) and ectorhinal (Ect) areas, whereas the medial LPr neurons gave off axon collaterals preferentially to the posterior parietal, cingulate, secondary motor, TeA, and Ect areas. These results suggest that the lateral and central parts of the LPr are primarily involved in the visual information processing, and the LPc and medial LPr may be important for higher-order function such as attention and learning.	P1-1-108 物体認知における速い視覚経路により生じる予測の機能的役割 A functional role of prediction generated by fast visual pathway in object recognition ○会田昇1, 樫森与志喜1,2 ○Noboru Aita1, Yoshiki Kashimori1,2 電気通信大学大学院情報システム学研究科情報メディアシステム学専攻1, 電気通信大学大学院情報理工学研究科2 Graduate School of Information Systems, Univ. of Electro-Communications, Chofu, Tokyo Japan1, Dept. of Engineering Science, Univ. of Electro-Communications, Chofu, Tokyo Japan2 We can recognize rapidly and effortlessly complex visual scenes. Such amazing ability in visual recognition needs the effective processing of visual information along the multiple stages of visual pathways. Neurophysiological experiments have provided evidence for a "simple-to-complex" processing model based on a hierarchy of increasing complex image features, performed along the feedforward pathway of the ventral visual system. On the other hand, visual system has abundant feedback connections, whose number is even larger than the feedforward ones. It has been proposed in predictive coding that in early visual areas prediction signal generated in higher visual areas is compared with feedforward signal coming from retina. However, it remains unclear how the prediction signal is generated in higher visual areas. To address this issue, we present a model of visual system that consists of the ventral visual pathway from retina to inferior temporal cortex (IT) and the memory system including perirhinal cortex and hippocampus. The model has two pathways: one is a slow pathway, in which visual information is processed along the ventral pathway from retina to IT via V1 and V4, and another is a fast pathway from lateral geniculate body to the memory system via superior colliculus. We show here the neural mechanism by which prediction signal emerges in the fast pathway to perform rapid object recognition. We also show that the prediction signal causes synchronous firing of neurons in the areas involved in the ventral pathway, playing a crucial role in feature conjunction of object recognition. These results suggest that the prediction generated by the fast pathway may contribute to rapid object recognition by producing the synergetic activity of broad brain areas.
P1-1-109 マウス大脳皮質第5層における単一細胞レベル多層構造の発生機構 Developmental mechanisms of single-cell-level multilayer organization in mouse neocortical layer 5 ○佐伯麻衣1, 黒川留美2, 丸岡久人1, 鶴野瞬1, 松本直実1, 小川正晴3, 細谷俊彦1 ○Mai Saeki1, Rumi Kurokawa2, Hisato Maruoka1, Shun Tsuruno1, Naomi Matsumoto1, Masaharu Ogawa3, Toshihiko Hosoya1 理研BSI局所神経回路1, 理研ASI吉田化学遺伝研2, 理研BSI視床発生3 Laboratory for Local Neuronal Circuits, BSI, RIKEN, Wako, Japan1, Chemical Genetics Laboratory, ASI, RIKEN, Wako, Japan2, Molecular Mechanisms of Thalamus Development, BSI, RIKEN, Wako, Japan3 The mammalian neocortex is composed of a large number of neuronal types. We have found that in mouse neocortical layer 5 specific neuronal types show highly-ordered three-dimensional microorganization. In the radial orientation, pyramidal neurons in layer 5 show multilayer organization. A number of layer 5 marker genes showed graded expression along the radial axis, and different genes showed different gradients. Axonal projection specificity also showed multilayer organization, and was well correlated with the graded gene expression. These multilayer organizations were built with single-cell precision. These results suggest that layer 5 is composed of multiple sublayers of functionally different neuronal types. It is not known how this multilayer organization is generated during the cortical development. One possibility is that the graded expression is established by positional information provided to neurons depending of their radial positions in layer V. To test this hypothesis, we analyzed the reelin(-/-) mice, in which the radial positions of cortical pyramidal neurons are severely disorganized. We analyzed the correlation of the expression of two genes that showed graded expression in wild type. In the reelin(-/-) cortex, although the radial cell positions were disorganized, the correlation of gene expression was not affected: For example, if two genes showed the same gradients in the wild type, their expression was positively correlated in the reelin(-/-) cortex. This result suggests that positional information is dispensable to establish the correlated expression, and that mechanisms independent of positional information play essential roles in the construction of the multilayer organization. Possible developmental mechanisms will be discussed.	P1-1-110 BGMとともに3D映像を視聴すると腹側皮質視覚路の活動が促進される Ventral processing stream activity is facilitated by watching 3D movie with BGM ○田口太郎1, 小野弓絵1, 遠藤聡人2, 小阪将也2, 藤井透2, 市川哲哉2 ○Taro Taguchi1, Yumie Ono1, Akito Endo2, Masaya Kosaka2, Tooru Fuzii2, Tetuya Itikawa2 明治大学院・理工・電気工学1, オリンパスメモリーワークス株式会社2 Grad Sch of sci & Tech, Meiji Univ, Kanagawa1, OLYMPUS MEMORY WARKS CORP., Tokyo, Japan2 Three-dimensional movies cause high realistic sensation and therefore have been widely adopted for the purpose of entertainment, although the neuronal and physiological effects of these movies have not yet quantitatively evaluated. We compared neuronal oscillatory activity among different watching conditions to investigate the effects of stereopsis and background music using full-CG 3D movies. Ten subjects with normal ability of stereopsis (aged 23.4 ± 1.4, 8 male and 2 female) watched the movies with three conditions of 2D with background music (2D), 3D with background music (3D), and 3D without background music (3DNS). Subjects watched five 3-min movies with one of the above conditions every single experimental day. They watched a fixation point shown on a screen as a baseline for 3 min in addition to watch the movies every experimental day. We recorded EEG at F3, F4, T3, C3, Cz, C4, T4, P3, P4, O1, and O2 in the international 10-20 electrode system and the EEG spectrum power in theta, alpha, and beta bands was compared with those during baseline. We found that the EEG spectrum power in beta frequency band was significantly increased during 3D condition in the T4 area compared to those during 2D and 3DNS conditions. These results suggest that the combination of background music and stereopsis facilitate the ventral processing stream of the visual system ("what" pathway) to increase the attention to recognize the movie screen, which may modulate the perception of visual stimuli even if we watch the same 3D movies.
P1-1-111 マウス視覚野での色選択性の機能的微小構造 Functional micro-architecture of color selectivity in mouse primary visual cortex ○藍原周平1, 吉田盛史1, 大木研一1 ○Shuhei Aihara1, Takashi Yoshida1, Kenichi Ohki1 九州大院・医・分子生理1 Mol Physiol, Kyushu Univ, Fukuoka, Japan1 Color perception systems have been studied in humans and monkeys that possess trichromatic color vision. Unlike these primates, rodents have only two types of color cones, UV- and middle-wavelength, in retina. Despite of the behavioral evidence that mice can distinguish different colors (Yacobs et al., 2004), color processing in rodents is not understood well. To understand how color information is processed in mice, we performed in vivo two-photon calcium imaging in mouse primary visual cortex (V1) and examined functional architecture of color responsive cells.We recorded responses of V1 cells to color stimulation. For color stimulation, we used two different LED lights whose spectral peaks are at 370nm (UV) and 510nm (green), corresponding to two types of mouse opsins. We found that some cells respond to both UV and green lights, but other cells respond to either UV or green light. We categorized these cells into UV-responsive, green-responsive, and both-responsive cells. We analyzed topographical arrangement of these differently categorized cells and found that they seem to be spatially intermingled in V1.It has been known that relative amount of green opsin is higher in dorsal and that of UV opsin is higher in ventral (Applebury et al., 2000). Therefore, we examined whether a global gradient map of color preference also exists in mouse V1. We recorded responses of cells from anterior and posterior part of V1 and obtained color response index (CRI) of individual cells as (R(UV)-R(green)) / (R(UV)+R(green)). We found that the distribution of CRI is biased to negative (green) in the anterior V1 which receives inputs from the dorsal retina, while it is biased to positive (UV) in the posterior V1 which receives inputs from the ventral retina. These results indicate that individual V1 cells have selectivity to colors with locally intermingled distributions of different color preferences, yet preserving the global topographical map of color preference in the retina.	P1-1-112 マウス一次視覚皮質における遅延性応答 Long-delayed responses of mouse primary visual cortex ○舟山健太1, 南澤玄樹1, 松本信圭1, 松木則夫1, 池谷裕二1 ○Kenta Funayama1, Genki Minamisawa1, Nobuyoshi Matsumoto1, Norio Matsuki1, Yuji Ikegaya1 東京大院・薬・薬品作用学1 Dept Chemical Pharmacol, Univ of Tokyo, Tokyo1 The primary visual cortex (V1) has long been used as a model to study cortical responses to sensory inputs. This cortical region receives direct projections from the lateral geniculate nucleus (LGN) of the thalamus and provides computational outputs to higher-order cortical areas. This route, commonly known as the feedforward (bottom-up) pathway, contributes to hierarchical neural processing of specific visual features, such as orientation, direction, color, and motion. Furthermore, V1 also receives projections from higher-order cortices (top-down). Thus, complex visual processing in V1 may allow cortical interplays between bottom-up and top-down information pathway. To comprehend its activity patterns, we patch-clamped layer 2/3 neurons in mouse V1 in vivo and monitored their voltage responses to visual stimuli. Previously, we found out a simple brief flash stimulus elicited a complex response consisting of a well-known transient depolarization with a latency of approximately 90 ms and thereafter a prolonged, larger depolarization (delayed response) that persisted for a few seconds. To comprehend response characteristics to further extent, we applied a brief orientated grating (flash grating). This stimulus also induced a delayed response which exhibited weak orientation selectivity. In vivo Ca2+ imaging of neuronal population suggested that despite weak selectivity of individual neurons, visual information were expressed by distinct neuron populations. We investigated the impact of delayed responses on visual function of behaving mice. We found that behavioral responses to prolonged grating stimuli were modulated by prestimulation with flash grating. Thus, contrary to the passive information processing model, V1 mingles instantaneously ongoing inputs with immediately preceding information and thereby may modify visual perception.
P1-1-113 二重正弦波縞の動きに対する運動視知覚と眼球運動反応 Motion perception and ocular responses to apparent motion of dual-grating stimuli ○野原静華1, 河野憲二1, 三浦健一郎1 ○Shizuka Nohara1, Kenji Kawano1, Kenichiro Miura1 京都大学　医学研究科　認知行動脳科学1 Dept. Integ. Brain Sci., Grad. Schl. Med, Kyoto Univ, Kyoto, Japan1 The illusory motion stimuli that cause reverse motion percepts are perceived as veridical if the successive motion frames are presented with inter-stimulus intervals (ISIs) of mean luminance. Some researchers interpreted these phenomena as evidence for the feature tracking mechanism of motion detection, while the others evidence for biphasic impulse response functions involved in the energy-based motion sensing. This study was designed to understand which of these mechanisms underlies visual processing for the motion perception and oculomotor control in humans. We used the 3f5f stimulus which was composed of two sinusoids of spatial frequency in the ratio of 3:5 (3f and 5f) creating a pattern with beat frequency, f. When this stimulus was stepped by one-fourth of the wavelength of the beat, the two components shifted one-fourth of their wavelengths and had opposite directions, the 5f forward and the 3f. backward. Two image frames that differ in phase (90°, 1/4 wavelength) were presented with various ISIs (0-853 ms). The total contrast of the patterns was fixed at 32%, while the contrast ratio between the 3f and 5f was 3/7 or 7/3. The subjects' task was to report the perceived direction of motion. Eye movements of the right eyes were monitored during the experiments. Perceived directions of the motion stimuli presented with the ISIs were veridical when the 3f component had higher contrast, while they were reversed when the 5f had higher contrast. These results are evidence for the energy-based mechanism with biphasic impulse response functions. The ocular responses to the same motion stimuli showed a similar dependence on ISIs when the ISIs were shorter, but the responses were almost negligible for longer ISIs despite that the perceived directions were significantly biased, suggesting a difference in mechanisms underlying the motion perception and oculomotor control.

上部に戻る

前に戻る