感覚運動系の学習・可塑性 Sensorimotor Learning/Plasticity
P3-1-107 Training-specific differences in cortical plasticity of musicians ○Jessica Grahn1, Jonathan Peelle University of Western Ontario1, Department of Otolaryngology, Washington University, St. Louis, USA2 An enduring question in cognitive neuroscience is the degree to which structural differences in the brain reflect individual differences in experience-dependent plasticity. Here we investigate this issue by using voxel-based morphometry (VBM) to compare regional gray matter (GM) volume between musicians who play different instruments, and thus have consistently different sensorimotor experiences from their peers. Past VBM studies of musicians have shown changes in GM volume relative to individuals with no musical training. However, often players of only one type of musical instrument have been tested, or players of many different instruments are analyzed as a single group (Gaser and Schlaug, 2002; Schneider et al., 2002; Sluming et al., 2002). In the current study we compare regional GM volume in keyboard, string, and woodwind/brass musicians to controls. The control group consisted of dancers, who have regular exposure to music but do not play an instrument. We predicted that the regular sensorimotor experience associated with different types of musical training would result in dissociable patterns of structural change in cortical GM. We found GM volume increases that were generic to musical training, as well as increases that were specific to different instrument types. Thus, not only does musical training generically affect local cortical structure, but the structural differences depend on the specific type of sensorimotor stimulation experienced by participants. These findings suggest that the brain networks supporting musical performance are heterogeneous, and allow the possibility that some types of training may generalize to non-music tasks better than others (Patel, 2011).	P3-1-108 経頭蓋直流刺激後のヒト一次運動皮質（M1）活動の可塑的変化：脳磁図研究 Plastic change of the human primary motor cortex (M1) activation after transcranial DC stimulation (tDCS): a neuroagnetic study ○三上佑介1, 松橋眞生1, 福山秀直1, 美馬達哉1 ○Yuusuke Mikami1, Masao Matsuhashi1, Hidenao Fukuyama1, Tatsuya Mima1 京都大学大学院　医学研究科附属　脳機能総合研究センター1 Human Brain Research Center, Kyoto University Graduate School of Medicine, Kyoto, Japan1 Background tDCS is a noninvasive technique that can modulate the human brain activity underneath the electrode, and is widely used in the fields of neuroscience and neurorehabilitation to induce neural plasticity. Especially, the anodal stimulation of the primary motor cortex (M1) can increase the motor excitability beyond the time of stimulation, while the cathodal one can cause suppression. Recent studies using electro- or magneto-encephalogram (EEG/MEG) suggested that the 20-Hz oscillation at M1 induced by median nerve stimulation is the cortical idling rhythm, which decreases its amplitude when M1 is activated. Hypothesis and Objective To further understand the neural mechanism of tDCS, we tested whether the tDCS-induced M1 plasticity is associated with the change in 20-Hz neuromagnetic osicllation or not. Methods We recorded the MEG activities from 4 right-handed normal volunteers using a 306-channel whole-head neuromagnetometer (Vectorview) before and after anodal tDCS intervention (2 mA, 15 min) over the left M1. To measure the left M1 excitability, the electric median nerve stimulation (MNS) above the motor threshold was given at the right wrist, and the MNS-induced 20-Hz rhythmic activity was evaluated (before tDCS, and immediately, 20 min and 40 min after the end of tDCS). Results We found the increase of the MNS-induced 20-Hz oscillation following tDCS. Discussion and Conclusion Our results showed that the anodal tDCS enhance the 20-Hz oscillation of M1 induced by MNS. Since the apparent relative increase in the amount of the induced oscillation can be caused by the decrease of the baseline magnitude of 20-Hz osicllation at rest, we are currently investigating this point to clarify the significance of our findings. In addition, further studies are necessary to elucidate whether these findings are directly related to the tDCS or not by using sham stimulation.
P3-1-109 マウスナー細胞の活動に伴うゼブラフィッシュ仔魚の逃避運動の可塑的変化 Mauthner cell activity associated with initiation and modification of fast escape behavior in zebrafish larva ○高橋めぐみ1, 谷本昌志1, 小田洋一1 ○Megumi Takahashi1, Masashi Tanimoto1, Yoichi Oda1 名古屋大院・理・生命理学1 Dept Biosci, Grad Sch Science, Univ of Nagoya, Nagoya1 Fish startle response is a typical simple behavior and can be modified by learning. It has been shown that activity of the hindbrain reticulospinal neurons correlates with startle behavior in zebrafish. Especially, a paired giant reticulospinal neurons, Mauthner (M) cells, are known to initiate fast escapes in response to sound (Zottoli, 1977; Kohashi and Oda, 2008). A previous study showed that long-term potentiation of the inhibitory synaptic strength in the auditory pathway to M-cell closely correlates with behavioral suppression of acoustically evoked fast escapes in goldfish (Oda et al., 1998). There was, however, no direct evidence for changes in M-cell activity during the learning. Here we performed calcium imaging of M-cell in behaving larval zebrafish to monitor the M-cell firing during fast escapes and studied the modification of the activity during learning. M-cells were retrogradely labeled with a calcium indicator, Oregon Green BAPTA-1, injected into the spinal cord. To simultaneously monitor M-cell response and escape behavior, the head of the larval zebrafish was embedded in agar with the tail allowed to move freely. Sound stimulus elicited two types of escapes, short- and long-latency C-starts (SLC and LLC), as shown previously (Burgess and Granato, 2007). We found that M-cell firing was associated with initiation of SLC, but not LLC. Repeated weak sound stimuli with subthreshold intensity for fast escapes suppressed the M-cell firing and SLC, but not LLC, without changing behavioral kinetics of the SLC and LLC. These results show that the sound-induced M-cell firing initiates SLC but not LLC, and suggest that decrease in the M-cell responsiveness to the sound stimulus underlies the SLC suppression.	P3-1-110 前向き誤差に基づく運動学習：運動学習の統一理論に向けて Prospective error to determine motor learning: A step toward a unified model of motor learning ○瀧山健1, 平島雅也2, 野崎大地2 ○Ken Takiyama1, Masaya Hirashima2, Daichi Nozaki2 日本学術振興会特別研究員/ 玉川大学1, 東京大学・教育・身体教育2 Brain Sci Inst, Tamagawa Univ, Tokyo1, Depart Education, The Univ Tokyo, Tokyo2 How the brain learns to control limb movement in variable environments is still a fundamental question in motor neuroscience. Recent studies have suggested that the brain generates appropriate motor commands depending on movement context by flexibly combining motor primitives (elements of motor control and learning). However, the manner by which the activity of motor primitives is determined is still controversial. One naive idea is that the desired movement determines the activity; a recent study by Gonzalez-Castro et al. (PLoS Comp Biol 2011) showed the possible involvement of the executed movement in determining the activity. Here, we propose an alternative model, in which the predicted error of an upcoming movement, called the prospective error (PE), determines the recruitment pattern of the primitives. This assumption is based on recent findings that showed some motor-related neurons encode PEs rather than the desired or executed movements (Popa et al., JNS 2012). Our model can account for several phenomena in motor learning such as savings (Zarahn et al., JNP 2008), anterograde interference, spontaneous recovery (Smith et al., PLoS Biol 2006), the relevance of error (Wei & Kording, JNP 2009), and structural learning (Braun et al., Curr Biol 2009). Although these phenomena have been conventionally explained by different models, our model can explain them in a single framework. In addition, this model predicts the following novel phenomenon that other conventional models were unable to predict. Motor adaptation to a constant visual perturbation (e.g., 30 ° rotation) is faster after experiencing the randomly changing errors (e.g., from -45 ° to 45 ° rotation) in every 2 trials than after experiencing the errors in a purely random manner. We validated this prediction by conducting a behavioral experiment. These results suggested that motor primitives are recruited according to the PE, and that the PE is a key to a unifying theory for motor learning.
P3-1-111 仮想現実空間における肢刺激呈示システムを導入したマウス用行動課題の確立 New behavior training system combined with tactile stimulation device in virtual world ○本間千尋1, 鴨志田敦史1,2, 山田一之1, 織田充1, 山川宏1, 村山正宜1 ○Chihiro Homma1, Atsushi Kamoshida1,2, Kazuyuki Yamada1, Mitsuru Oda1, Hiroshi Yamakawa1, Masanori Murayama1 理研・BSI・行動神経生理1, 日本ナショナルインスツルメンツ2 Lab for Behav Neurophysiol, BSI, RIKEN, Saitama, Japan1, National Instruments Japan Corporation, Tokyo, Japan2 Cutaneous sensations are thought to be represented in primary somatosensory cortex (S1) as a result of neural activity, which can then be used for higher brain functions such as decision-making, attention and memory. However it is still unclear, 1) how neural activity represents these sensations and 2) how the neural activity contributes to those functions. In previous study we showed that intensity of limb stimulation in rodents can be coded by dendritic activity of layer 5 pyramidal neurons in S1. This implies that somatosensation of the tactile stimuli can also be used as a primary mode of input that is sent to higher brain areas. To further test this hypothesis, we developed new experimental systems in which mice discriminated texture in a modified Y-maze in real world. With this maze, mice were successfully trained to discriminate a specific tactile stimulus on the maze floor at the area just before branch point. From these results, we constructed this task in virtual world situations. This system can present tactile stimulation arbitrarily while mice are running maze in virtual world. We are now using this system to electrophysiological and optogenetical experiments, and tying to provide new knowledge in neurobehavioral researches.	P3-1-112 ヒト間接皮質脊髄路における可塑的変化の活動およびタイミング依存性 Activity- and timing-dependent plasticity in indirect cortico-motoneuronal pathway of humans ○中島剛1, 小宮山伴与志2,4, 大塚裕之3, 鈴木伸弥4, 二橋元紀4, 大木紫1 ○Tsuyoshi Nakajima1, Tomoyoshi Komiyama2,4, Hiroyuki Ohtsuka3, Shinya Suzuki4, Genki Futatsubashi4, Yukari Ohki1 杏林大学　医学部　統合生理学1, 千葉大学教育学部2, 千葉大学大学院医統合生理3, 東京学芸大学連合大学院4 Dept Physiol, Kyorin Univ, Sch. of Med, Tokyo Japan1, Dept Health and Sports Sci, Chiba Univ, Chiba, Japan2, Dept Physiol, Chiba Univ, Sch. of Med, Chiba, Japan3, Dept Health and Sports Sci, Tokyo Gakugei Univ, Tokyo, Japan4 We previously reported that repetitive combined stimulation (RCS) of pyramidal tract and peripheral nerve induces the long-term potentiation (LTP) of indirect cortico-motoneuronal (C-M) excitations, probably via. cervical propriospinal neurons (PNs) in humans. However, mechanisms of the effects are still not clear. For the purpose, we further examined characteristics of the modulation of C-M excitation after RCS. Healthy volunteers (n=12), who all gave written informed consent, participated in experiments. Electromyograms were recorded from right biceps brachii (BB) and first dorsal interosseous muscles. RCS intervention (0.2 Hz) was performed for 10 min, in which transcranial magnetic stimulation (TMS) to the contralateral motor cortex was delivered with right ulnar nerve stimulation, with or without weak contraction of BB. Interstimulus interval (ISI) for the combined stimulation was set to -10, 2 or 10 ms (negative value: TMS ahead), while simultaneous inputs by both stimuli on PNs will be obtaied under ISI 7 ms. Stimulus strengths were determined to observe the maximum spatial facilitation effect in BB before the intervention, by converging inputs on PNs under ISI 10 ms. To observe LTP in C-M excitations in BB, motor evoked potentials (MEPs) were evoked by TMS at ~0-70 min after the intervention. As reported previously, MEP amplitudes were enhanced after RCS under ISI 10 ms and with contraction, which lasted for 20-60 min. However, long-term depression (LTD), instead of LTP, was observed after the RCS without contraction. Furthermore, even with contraction, LTP (ISI 10 ms) and LTD (ISI -10 and 2 ms) could be observed under different ISIs after the RCS. These results suggest that plastic changes induced by RCS are timing- and activity- dependent, which resembled the characteristics of plastic changes on synapses in CNS. The timing to observe LTP indicates that the changes took place in synapses from the pyramidal tract on cervical PNs.
P3-1-113 2つの運動技能間に生じる干渉はランダム練習によって軽減する Random practice of two motor skills increases robustness against behavioral interference from one to the other ○上原信太郎1,2, 水口暢章3, 山本真史1,2, 廣瀬智士2,4, 内藤栄一3,5 ○Shintaro Uehara1,2, Nobuaki Mizuguchi3, Shinji Yamamoto1,2, Satoshi Hirose2,4, Eiichi Naito3,5 京都大学大学院 人間・環境学研究科1, 日本学術振興会2, 独立行政法人情報通信研究機構　脳情報通信融合研究センター3, ATR　認知機構研究所4, 大阪大学大学院　医学研究科5 Human and Environmental studies, Kyoto Univ, Kyoto1, JSPS2, NICT CiNet3, ATR-CMC4, Graduate School of Medicine, Osaka Univ5 Behavioral interference is a well-known behavioral phenomenon, where execution of one motor task deteriorates following performance of another task. Here we show that random practice could be an effective solution to avoid the occurrence of behavioral interference.A total of 32 healthy volunteers practiced two different types of sequential finger tapping tasks. They were subdivided into two groups; one group (random group; n = 16) practiced the two tasks in a randomized order, another group (block group) practiced one task first and then practiced the other task next. In the next day, both groups performed the two tasks in a blocked order (= performed one task first and the other second). We evaluated the number of correct responses in the first and the second tasks and compared these performances in this day.We found that the performance in the second task was lower than that in the first task in the block group. As the performances in the first and the second tasks were the same in the random group, the finding was interpreted as the first task interfered with the second task only in the block group.These results indicate that motor memories for two tasks are differently formed in the central nervous system (CNS) between block and random groups, and the memories for two motor skills seem to be independently stored in the CNS through random practice so as to increase the robustness against interference from one to the other.	P3-1-114 腹側被蓋野の電気刺激により両側一次運動野皮質に生じる神経活動の光学イメージング Optical imaging of neuronal activity in bilateral motor cortex of rats after unilateral ventral tegmental area stimulation ○九里信夫1,2, 梶原利一3, 高島一郎2 ○Nobuo Kunori1,2, Riichi Kajiwara3, Ichiro Takashima2 筑波大院・人間総合・感性認知脳科学1, 産総研　ヒューマンライフテクノロジー研究部門　システム脳科学研究グループ2, 産総研　バイオメディカル研究部門　脳遺伝子研究グループ3 Comp. Human sci., Univ of Tsukuba, Japan.1, Human Tech. Res. Inst., AIST, Tsukuba, Japan2, Biomed. Res. Inst., AIST, Tsukuba, Japan3 Dopamine neurons in the ventral tegmental area (VTA) project to the frontal lobe including the primary motor cortex (M1) (Hosp et al., 2011). In our previous reports, we used voltage-sensitive dye (VSD) imaging and demonstrated that single pulse electrical stimulation of VTA induced transient depolarization followed by hyperpolarization in M1 neurons, producing the strongest response in the forelimb area. In the present study, we recorded local field potentials (LFPs) throughout the cortical layers of M1 using a 16-site multiprobe electrode in order to verify the cortical layers that receive synaptic input from VTA. The recorded LFPs were subjected to current source density (CSD) analysis, which revealed three prominent current sinks, i.e., putative sites of excitatory synaptic inputs, in layer 2/3 and 5. An initial current sink appeared in upper layer 5 together with a strong sink in layer 2/3, which was followed by another massive current sink in deep layer 5 at an interval of ~5 ms. Next, we investigated whether the M1 response elicited by ipsilateral VTA stimulation propagates to the contralateral motor cortex via corpus callosum (CC) which constructs inter-hemispheric connections. To confirm this possibility, VSD imaging was performed on bilateral M1 before and after CC transection. Before the CC transection, unilateral VTA stimulation evoked neuronal activities in both the bilateral M1, and the contralateral response was delayed about 15 ms compared to that of ipsilateral. Whereas, after the CC transection, the contralateral M1 response completely disappeared, but the ipsilateral response remained intact. In summary, it was indicated that information from the VTA was first transmitted to the ipsilateral M1 neurons of layers 2/3 and 5, and then relayed the contralateral M1 through the CC. Further research is needed to understand the functional roles of this projection from the VTA to the bilateral M1.
P3-1-115 ヒト感覚運動皮質における競合する視覚運動変換の神経表象 Human sensorimotor cortex represents conflicting visuomotor mappings ○今水寛1,2, 小川健二2,3 ○Hiroshi Imamizu1,2, Kenji Ogawa2,3 情報通信研・脳情報通信融合研究センター1, ATR認知機構研2, 日本学術振興会3 NICT Center for Information & Neural Networks, Osaka, Japan1, ATR Cognitive Mechanisms Labs., Kyoto, Japan2, Japan Society for the Promotion of Science, Tokyo, Japan3 Behavioral studies have shown that humans can adapt to conflicting sensorimotor mappings that cause interference after intensive training. While previous research works indicate the involvement of distinct brain regions for different types of motor learning (e.g. kinematics vs. dynamics), the neural mechanisms underlying joint adaptation to conflicting mappings within the same type of perturbation (e.g. different angles of visuomotor rotation) remain unclear. To reveal the neural substrates that represent multiple sensorimotor mappings, we examined whether different mappings could be classified with multi-voxel activity patterns of functional magnetic resonance imaging data. Participants simultaneously adapted to opposite rotational perturbations (+90 and -90 degrees) during visuomotor tracking. To dissociate differences in movement kinematics with rotation-types, we employed two distinct patterns of target motion and tested generalization of the classifier between different combinations of rotation- and motion-types. Results showed that the rotation-types were classified significantly above chance using activities in the primary sensorimotor cortex as well as the supplementary motor area, despite no significant difference in averaged signal amplitudes within the region. In contrast, low-level sensorimotor components, including tracking error and movement speed, were best classified using activities of the early visual cortex. A univariate analysis of single voxels could not find a significant difference in activity between the rotation-types. Our results reveal that the sensorimotor cortex represents different visuomotor mappings at the sub-voxel level, which permits joint learning and switching between conflicting sensorimotor skills.	P3-1-116 乱雑な環境に対する運動学習は感覚運動変換における不確実性の表現を更新する Adaptation for uncertain environment updates trail-to-trail variability of gain in sensorimotor map ○井澤淳1, 吉岡利福2, 古川友香2, 大須理英子1 ○Jun Izawa1, Toshinori Yoshioka2, Yuka Furukawa2, Rieko Osu1 ATR脳情報通信総合研究所1, 独立行政法人 情報通信研究機構2 ATR Computational Neuroscience Laboratories1, National Institute of Information and Communications Technology2 When we make a reaching movement, our brain predicts sensory consequences of the action building an association between the efference copy of the motor commands and the sensory feedback. This internal model of the sensory prediction also plays a central role in generating motor commands. Implicit in motor control study is that the representation of internal model is deterministic where the memory is updated to predict the expectation of trajectories. However, in daily life the external world interacting with the nervous system is continuously changing and the sensory information is often influenced by noises. How does the brain deal with this uncertainty? We hypothesized that the brain builds internal model employing probabilistic parameters to represent the uncertainty. To examine this idea, we asked the subjects to reach in channel force and measured the perpendicular forces after they adapted to reach in the zero-mean stochastic force field in which either the elasticity or the viscosity parameters very every trial. We found that the trial-to-trial motor variability that appeared in the channel trials significantly increased after they experienced a stochastic environment. The temporal profile of motor variability in the channel generated by the subjects who experienced the position dependent noise field increased monotonically over time whereas it generated from the subjects who experienced the velocity dependent noise field showed bell-shaped profile. To understand a computational mechanism, we fitted the gain tuning model of the position and the velocity dependent force field to the temporal pattern of the perpendicular forces in the channel. We found that the gains of two internal models are significantly different between these two conditions. This suggests that experience of stochastic environments forms uncertainty of internal model that generates state dependent motor variabilities.
P3-1-117 ソングバード歌発達における学習臨界期中の脳内遺伝子発現誘導率制御 Dynamics of singing-driven genes in a vocal premotor area during critical period for vocal learning ○早瀬晋1, 大串恵理1, 小林雅比古1, 和多和宏1,2 ○Shin Hayase1, Eri Ohgushi1, Masahiko Kobayashi1, Kazuhiro Wada1,2 北海道大学大学院　生命科学院1, 北海道大学大学院理学研究院2 Graduate School of Life Science, Hokkaido Univ, Sapporo, Japan1, Faculty of Science, Hokkaido Univ, Sapporo, Japan2 Vocal learning in human and other species requires integration of auditory information and one's own vocalizations. Some of these species develop own vocalizations only within a time called critical period. To investigate the molecular mechanisms for regulation of the critical period for vocal learning, we focused on zebra finch, a songbird which is one of a few vocal learning groups. A juvenile bird develops its song from highly variable vocalizations to the crystallized form as an adult in the critical period. Even in a day, a juvenile bird develops its song; songs in the morning are variable and become relatively structured in the evening. For vocal learning and production, songbirds possess specific neural pathways, consisting of song nuclei. In the song nuclei, a variety of genes are known to be induced by singing. However, it is not well known the relationship between the singing-induced genes and song development. These genes might also be differentially regulated by ages and time points in a day to regulate the critical period. In this study, we examine dynamics of singing-driven genes expression in a day of the juvenile and adult stages. As the result, a singing-induced gene, Arc was strongly induced in the pallial premotor nucleus RA by singing, specifically in the morning of juvenile stage, compared in the evening of juvenile stage and all the day of adult stage. In the juvenile stage, Arc induction was still high in the evening if the bird sang the first series of song in a day by prohibition of singing, suggesting circadian regulation was not involved in the change of the gene induction rate. Induction rate of c-fos and Egr1, which locate on different chromosomal loci, were similar to Arc in same neurons in the RA. In contrast, in the HVC, these genes induction was robust but not changed all the day in both juvenile and adult stages. These data suggests a strong relationship between song variability and induction rate of singing driven genes in the RA.	P3-1-118 動力学的環境が到達運動時の使用手選択に及ぼす影響 Effect of dynamic environment on hand choice in arm reaching ○幅岸千恵1, 春日翔子2, 大高洋平3, 里宇明元3, 牛場潤一2,3 ○Chie Habagishi1, Shoko Kasuga2, Yohei Otaka3, Meigen Liu3, Junichi Ushiba2,3 慶應義塾大学大学院理工学研究科 基礎理工学専攻1, 慶應義塾大学理工学部 生命情報学科2, 慶應義塾大学医学部 リハビリテーション医学教室3 School of Fundamental Science and Technology, Graduate School of Science and Technology, Keio University, Kanagawa, Japan1, Department of Biosciences and Informatics, Faculty of Science and Technology, Keio University, Kanagawa, Japan2, Department of Rehabilitation Medicine, Keio University School of Medicine, Tokyo, Japan3 In daily life, we frequently face the situation in which we have to make a decision as to which hand to use for some motion such as reaching a coffee cup or opening the door. A recent study demonstrated that implicit virtual reward/cost manipulation by visual feedback in a short time scale could change hand choice to reach a target (Stoloff et al., 2011). In this study, we investigated whether hand choice of reaching was also modified by changes in dynamic environment, as a notion expansion for more natural situation. We developed a reaching task with an exoskeleton robotic system in which subjects moved the indicated hand (i.e., left or right) to a target which appeared at random locations on a display with a regulated speed. During training (330 trials), a resistive viscous force was abruptly applied to the dominant hand and remained constant through the trials to change dynamic environment. The choice test where subjects freely decided which hand to reach the target was interleaved every 110 training trials to evaluate hand choice. The result showed that the use frequency of the loaded hand significantly decreased when the force was introduced, and then recovered during the training in 9 out of 17 subjects. Next, to investigate the effect of temporal change of dynamic environment on hand choice, the same amount of force was applied but was increased gradually through the trials. The result also showed a rapid decrease of the use frequency of the loaded hand at the beginning of the training. The loaded hand was then significantly less used by the end of the training (p = 0.031). It is suggested that the uncertainty in the body state resulting from time-varying force perturbation rather than the amount of force applied may have stronger influence on hand choice. These findings implicate a good model of stroke hemiparesis to understand how functional asymmetry and its temporal fluctuation in arm dynamics influence on the recovery from the learned non-use.

上部に戻る

前に戻る