感覚運動制御 Sensorimotor control
P2-1-117 ゼブラフィッシュのマウスナー回路／非マウスナー回路によって駆動される音/振動誘発性のすばやい逃避運動 Sound/vibration-evoked fast escapes triggered by Mauthner circuits or non-Mauthner circuits in zebrafish ○谷本昌志1, 横道聖奈1, 杉本温子1, 高橋めぐみ1, 小田洋一1 ○Masashi Tanimoto1, Sena Yokomichi1, Atsuko Sugimoto1, Megumi Takahashi1, Yoichi Oda1 名古屋大学大学院 理学研究科 生命理学専攻1 Div. Biol. Sci., Grad. Sch. Sci., Nagoya Univ.1 To execute an escape response, a single neuron in at least some species produces a decisive command to rapidly flex the body [e.g. crayfish giant fibers (Wiersma, 1947)]. The Mauthner (M-) cell, a pair of giant reticulospinal neurons in the teleost fish hindbrain, plays a central role on the initiation of the fast escapes (Zottoli, 1977). The M-cell receives acousticovestibular inputs and activates the contralateral spinal motor circuits, thereby triggering the C-shaped body bending. An ablation study suggested that the M-cell is required for sound/vibration-evoked short-latency C-starts (SLCs) in zebrafish larvae (Burgess and Granato, 2007), whereas another work indicated inconsistent results (Kimmel et al., 1980). To reexamine the functional role of the M-cell on the fast escapes, we ablated the M-cell at 1 or 5 days postfertilization (dpf) using a fine glass microprobe in Tg(α-actin:GFP) or Tol-056 transgenic zebrafish which express green fluorescent protein (GFP) in the M-cells (Miyashita et al., 2004; Satou et al., 2009), and analyzed sound/vibration-evoked escapes at 6 dpf. Control larvae showed SLCs with a peak onset latency of 5 ms. In contrast, the escapes of larvae with bilateral M-cell ablation started with a delayed onset (with a peak of 11 ms), but most of them were classified as SLCs. Juveniles and adults also exhibited the delayed SLCs. To examine whether intact fish show SLCs without M-cell activity, we recorded M-cell firing by calcium imaging during fast escapes (Kohashi and Oda, 2008). Most of SLCs occurred with M-cell firing, but a small fraction of SLCs with onset latency of 6-14 ms did not accompany M-cell activity. These data show that the M-cell is necessary for the initiation of sound/vibration-evoked escapes with the shortest onset latency in zebrafish, and suggest the presence of alternative (non-M) circuits to initiate fast escapes in zebrafish, as indicated in adult goldfish (Eaton et al., 1982; Zottoli et al., 1999).	P2-1-118 遠心性信号が重さの知覚に与える影響についての研究 A study of efferent signal's influence on weight perception ○小川展夢1, 川瀬利弘4, 神原裕行2, 辛徳2, 吉村奈津江2, 小池康晴2,3 ○Hiromu Ogawa1, Toshihiro Kawase4, Hiroyuki Kambara2, Duk Shin2, Natsue Yoshimura2, Yasuharu Koike2,3 東工大院・総合理工・物情1, 東工大　精研2, JST CREST3, 東工大・総合理工・知シス4 Dept Info Processing ,Titech, Kanagawa1, P & I Lab., Titech, Kanagawa2, JST CREST, Tokyo3, Grad Sch of Sci and Eng, Titech, Kanagawa4 When we compare the heaviness of two objects of equal weight, we feel the smaller as the heavier. This phenomenon, called size-weight illusion, shows that weight perception is influenced not only by somatosensory inputs, but also by the prediction of object's weight derived from its physical appearance.Our group had revealed that the prediction of object's weight reflects on the stiffness of the wrist just before objects are placed on the hand; when we predict object is heavy, the stiffness becomes high. Consequently, it is expected that more study about relationship between stiffness and perception may reveal the influence of a prediction on a perception.However, it was difficult to study, based on the stiffness, how the prediction influences the perception. Since the stiffness is also influenced by somatosensory inputs, it is difficult to pick out only prediction's influence from recorded stiffness.To solve this problem, we used a robotic arm, which could reproduce the movement of subject's wrist by his EMG. Subjects held two objects with the robotic arm, and judged which was heavier. In this situation, they could not get any somatosensory inputs. Therefore, we can assess the influence of the prediction on the perception separately from the sensory inputs.As the results of the experiment, we got following two facts.1. Subjects judged the order of two object's weight as precisely as when the objects were placed on their own hand. Furthermore, they experienced size-weight illusion as well.2. The subjects' judgments could be estimated accurately from the four valuables; stiffness levels just before and after object's placement, angles of the wrist just before and after object's placement. The result of the experiment showed that both stiffness, an efferent signal of prediction of object's weight, and visual information of the wrist play important rolls in a perception of weight.
P2-1-119 目標追従課題におけるヒト運動制御の間欠性は目標軌道の予測可能性に依存する Intermittency of human motor behavior in target tracking depends on predictability of target motion ○井上康之1, 阪口豊1 ○Yasuyuki Inoue1, Yutaka Sakaguchi1 電通大院情報システム情報メディアシステム1 Grad. School of IS, Univ of Electro-Communications, Tokyo, Japan1 Discontinuous velocity patterns (known as "intermittency") are often observed in human motor behaviors, even in continuous motor tasks (such as target tracking). Although the underlying mechanism of this phenomenon is still unclear, it seems broadly accepted that non-negligible delay in sensorimotor system is as essential factor, and several motor control frameworks have been proposed to explain its nature and functional meaning. We have also proposed a computational framework based on the concept of "model predictive control (MPC)" (Tanaka, Inoue and Sakaguchi, 2012), where a continuous motor task is executed as a series of movement segments each which is executed by feed-forward (or predictive) control. Because the essence of this framework is prediction of the target and self-movements, behavior of the control model naturally depends on their predictability. Especially, time lengths of the movement segments plausibly depend on the predictability. We hypothesized that this may explain why the interval of the intermittent discontinuities varied according to the task condition. To test this view, we conducted a behavioral experiment using a visuo-manual tracking task with manipulating the predictability of the target motion. Because we presumed that transition pattern and velocity profile of target motion could affect the predictability, we prepared sinusoidal, minimum-jerk, and constant-acceleration trajectories as unit patterns of a reciprocal target motion between two endpoints. Although tracking trace was similar among these conditions, timing of segment boundaries tended to be concentrated at specific phases of the target motion (just after leaving the endpoint) in the minimum-jerk and constant-acceleration conditions, compared to the sinusoidal condition. This suggests that predictability of target motion affects segmentation of motor control, supporting the predictive control framework.	P2-1-120 実験的咬合干渉の有無による脳賦活状態の検討 Effects of experimental occlusal interferences on brain activation during gum chewing ○林勇大1, 大塚剛郎1, 丹羽政美2, 笹栗健一1 ○Yuta Hayashi1, Takero Otsuka1, Masami Niwa2, Kenichi Sasaguri1 神奈川歯大　成長発達1, 揖斐厚生病院2 Dept Cranio Growth & Develo Dent, Kanagawa Dent Col, Kanagawa1, Ibi Kosei Hospital, Gifu2 Objective: Modern prosthetic restorations are fabricated, fitted and adjusted with the aim of restoring occlusal function. In most cases, the ultimate goal of occlusal therapy is comfort as determined by the subjective assessment of the patient. Our previous study reported that occlusal interference of posterior region induced unpleasant feeling during gum chewing. However, the effect of each tooth type has not been determined. Methods: In this study, using functional magnetic resonance imaging in eight healthy human subjects, measured blood oxygenation level-dependent (BOLD) signals during gum chewing in interference models, using each tooth type overlays (1st premolar, 1st molar, 2nd molar) of thickness about 200μmResults: Group analysis showed that gum chewing with all overlays was associated with significant increases in the BOLD signal in the sensory and motor cortexes of oral region. In addition, activation area with overlay at molar region (1st molar or 2nd molar) was not only sensorimotor cortex, but also limbic area (amygdala, anterior cingulate cortex). Conclusion: These results suggested that specific contact of molar region may prevent healthy occlusal functions.
P2-1-121 推定された腕の状態を反映した力学的外乱への短潜時応答 Fast corrective responses to perturbations applied during reaching reflect estimated limb state: Evidence for optimal feedback control in the motor system ○林拓志1, 横井惇1, 平島雅也1, 野崎大地1 ○Takuji Hayashi1, Atsushi Yokoi1, Masaya Hirashima1, Daichi Nozaki1 東京大院 教育学 身体教育学1 Dept of Physical and Health Edu, Grad of Edu, Univ of Tokyo1 The accurate estimation of limb state during movement is accomplished by integrating sensory information with the sensory consequences predicted by the forward model. According to the optimal feedback control theory, the estimated state should be used to generate motor commands rather than the actual state. However, there is little experimental evidence to support this notion. Thus, in this study we created novel situations in which the predicted state was altered while the actual state remained unchanged, and we assessed whether alterations in the predicted state influence fast responses of the arm which, like long-latency reflex, reflect how the brain transforms sensory feedback signals into motor commands. Participants performed planar reaching movements by moving a handle towards a target that was gradually shifted with trials in a certain direction. The visual rotation was simultaneously applied to the cursor representing the handle position in the opposite direction. In this way, only their predicted movement was altered while their physical movement was unchanged without being recognized by participants. Fast force response (FFR) was induced to oppose a brief assistive mechanical perturbation, which was applied to the handle during movement. Notably, the perturbation and the required response were identical despite alterations of predicted state resulting from the target shift. If FFR is determined by the actual state of the limb, then it should not be influenced by the predicted state. However, the FFR changed its direction significantly in accordance with the alterations of target shift. Thus, fast response to perturbation was not solely determined by the actual state of the limb. Rather, it reflected the estimated state, whereby it could be flexibly modulated according to the estimated states. Assuming that the estimated state was used to generate feedback motor commands, these results are consistent with the optimal feedback control theory.	P2-1-122 運動スキルのバラツキは広範な運動関連領野の活動変化に由来する Variability of skillful motor performance is stemmed from variability of neuronal activity in wider range of brain regions recruited during motor execution ○水口暢章1, 上原信太郎2,3, 廣瀬智士3,4, 山本真史2,3, 内藤栄一1,5 ○Nobuaki Mizuguchi1, Shintaro Uehara2,3, Satoshi Hirose3,4, Shinji Yamamoto2,3, Eiichi Naito1,5 情報通信研究機構・脳情報通信融合研究センター1, 京都大院・人環2, 日本学術振興会3, ATR・認知機構研究所4, 大阪大院・医学系5 CiNet, NICT, Kyoto1, Grad sch of Human and Environmental Studies, Kyoto Univ, Kyoto2, JSPS3, DCN, ATR, Kyoto4, Grad sch of Med, Osaka Univ, Osaka5 Human motor performances vary across trials even when people perform well-trained motor skills. In non-human primate study, it is shown that the variability of reaching movements is originated from the variability of spikes in the premotor cortex (Churchland et al. 2006). However, we hypothesized that variability of motor performance is also derived from variability of neuronal activity in other brain regions recruited during motor execution. In the present study, we focused on a skillful motor task and investigated brain activity that generates variability of its motor performance using functional magnetic resonance imaging (fMRI). 15 right-handed participants repeated a sequence of 5 button-presses (one sequence = ring–middle–little–index–ring) with their right hand for 10 seconds (= epoch) as fast and accurate as possible. Ten epochs were completed with inter-epoch-intervals of 12 seconds in each fMRI scan, and the participants completed 15 fMRI scans in the experiment. We calculated the number of correct sequences for each epoch as an index for the variability of motor performance, and performed parametric modulation analysis. As the participants well-trained the task before the experiment, no consistent performance improvement was observed in the scanner. But we found that the performance (the number of correct sequences) varied across epochs. In the fMRI analysis, we found that the lower performance was associated with the higher activities in bilateral premotor cortices, dorsolateral prefrontal cortices, cingulate cortices, insular cortices and parietal cortices. Thus, the present findings suggest that variability of motor performance is stemmed not merely from variability of premotor activity but from the activity in wider range of brain regions that are recruited during motor execution.
P2-1-123 環境の不確実性に応じてスティフネスレベルを調節する物体受け取りタスクのための運動制御・学習モデルの提案 A motor control-learning model enabling stiffness adjustment according to uncertainty of object's weight during load-on task ○神原裕行1, 小川展夢2, 辛徳1, 小池康晴1,3 ○Hiroyuki Kambara1, Hiromu Ogawa2, Duk Shin1, Yasuharu Koike1,3 東工大・精密工学研究所1, 東工大・総理工・物理情報システム2, 科技構・CREST3 P&I Lab, Tokyo Tech, Yokohama1, Dept Information Processing, Tokyo Tech, Yokohama2, CREST, JST, Tokyo3 When we try to receive an object, our brain predicts weight of the object from its physical appearance. Motor command signals, according to the predicted weight, is then sent to muscles in advance of the object's placement on the hand. Prediction of external world dynamics and anticipatory control based on the prediction are the keys to generate stable movements actuated by the muscles of which force cannot be changed instantaneously. Previously, we had proposed a motor control-learning model for load-on task in which an object is placed on the hand (Kim et al., 2009). We had confirmed that the model can reproduce subject's motion if the correct dynamics, the weight of the object in this case, is given to the model as a context cue. In our real world, however, it is difficult to predict exact dynamics of the world. In case the dynamics cannot be predicted accurately, we often stiffen our body and prepare for unpredictable external force. When we ride a bike on a dirt road, for example, we often stiffen our arm to stable the handle against unexpected force disturbance exerted by stones on the road. In this study, we investigated whether our motor control-learning model can learn to adjust the stiffness of the arm according to the uncertainty of the prediction of object's weight.We run computational simulations of load-on task in which actual object's weight is determined randomly with Gaussian distribution in each load-on trial but the predicted weight was kept constant in all trials at the mean of the distribution. As the results, we confirmed that stiffness level of the arm became higher as the variance of Gaussian distribution became larger. This result shows the ability of our model to adjust the stiffness level according to probabilistic features of the external world.

上部に戻る

前に戻る