その他 Others
P1-2-251 ラット社会行動及び性行動の3次元ビデオ解析システム A system for 3D video analysis of social and sexual interactions in rats ○松本惇平1, 浦川将2, 堀悦郎1, 小野武年2, 西条寿夫1 ○Jumpei Matsumoto1, Susumu Urakawa2, Etsuro Hori1, Taketoshi Ono2, Hisao Nishijo1 富山大学大学院医学薬学研究部システム情動科学1, 富山大学大学院医学薬学研究部神経・整復学講座2 System Emotional Science, Grad. Sch. Of Med. Pharma. Sci., Univ. of Toyama, Toyama1, Dept. Judo Physiother., Grad. Sch. Of Med. Pharma. Sci., Univ. of Toyama, Toyama2 Extensive studies have applied to social and sexual interaction between two rodents to analyze neural functions. Previous computerized video analysis systems using 2D video had difficulties in estimating 3D positions of animals. Furthermore, once animals overlapped (e.g. mounting), these systems easily lost tracking targets and needed manual intervention. In this study, we developed a novel system that recorded and analyzed 3D video of two rats during social and sexual interactions. In the proposed system, the images of 4 depth cameras were integrated into 3D video. Then, rat skeleton models consisting of head, neck, trunk, hip and joints between them were fitted to the 3D video, to estimate 3D position of each body part of each rat. Behaviors were defined as spatio-temporal patterns of these positions. The validity of the system was tested in male-male social interaction and male-female sexual interaction. First, the videos were divided into serial 20 or 4 s periods around copulatory behaviors in social or sexual interaction, respectively. In each period, we checked whether the system could keep tracking target. Second, the positions of nose and origin of tails were marked by two blind experimenters and compared with the outputs of the systems. Finally, the moments when the system detected mountings and when an experimenter detected by inspection were compared. The results indicated that the system could keep the tracking targets in 94 and 90 % of the periods in social and sexual interaction, respectively. The errors between the positions estimated by the system and by the experimenters were 2.1±1.4 cm and 3.1±2.1 cm in social and sexual interaction, respectively. Furthermore, among 42 moments that the system detected, 41 were the same as ones that the experimenter judged. These results demonstrated that the proposed system can estimate the 3D positions of body parts of interacting two rats and detect behaviors correctly and robustly even if those rats overlapped.	P1-2-252 MRIを用いて脳神経活動を直接的に計測できるか？ – 拡散MRI法をもちいた新規計測法の開発 Can we detect neural activation with MRI in the presence of neurovascular coupling inhibiters? – Direct detection of neural activation by diffusion fMRI ○釣木澤朋和1,2 ○Tomokazu Tsurugizawa1,2, Luisa Ciobanu2, Denis Le Bihan2 味の素（株）イノベ－ション研究所1 Institute for Innovation, Ajinomoto Co., Inc.1, Neurospin/CEA, Gif-sur-Yvette, France2 Functional magnetic resonance imaging (fMRI) using the blood oxygen level-dependent (BOLD) method relies on the neurovascular coupling and indirectly detects neuronal activation by measuring local hemodynamics changes. As a result, BOLD fMRI responses can be significantly altered or even suppressed in situations interfering with the neurovascular coupling mechanisms by specific drugs such as alcohol, isoflurane, and hypotension drugs. Another candidate to investigate the neural activity is diffusion fMRI (DfMRI) that can visualize brain activation based on changes in water diffusion in the neurons or astrocytes. Here we assessed whether BOLD or diffusion fMRI could detect the neural activity under specific drugs disrupting the neurovascular coupling. First, we compared BOLD and diffusion fMRI signals acquired at 7T with in vivo electrophysiological local field potentials (LFPs) evoked in the primary somatosensory cortex of rats during forepaw electrical stimulation. During the protocol, the rats were anesthetized with isoflurane and subject to a continuous infusion of nitroprussiate, a neurovascular-coupling inhibitor which acts by elongating the blood vessels. Second, we compared BOLD and diffusion fMRI signal changes after intraperitoneally infusion of 10% ethanol solution. Under nitroprussiate infusion, the BOLD response was abolished while the DfMRI response and LFPs were maintained. After ethanol infusion, BOLD fMRI signals were significantly decreased in several brain regions including the cortex and nucleus accumbens, which is related to the reward and addiction. However, DfMRI signals were significantly increased in the nucleus accumbens, consistent with previous results. These results demonstrate that water diffusion in the neuropile could more directly reflect dynamic tissue structural changes in activated regions and enable the understanding of brain function even under the action of drugs disrupting the neurovascular coupling.
P1-2-253 単一言語課題時の酸素化Hbの変動 Single event-related changes in cerebral oxy-Hb during verbal tasks ○石井洋平1, 森田喜一郎1,2, 小路純央1,2, 藤木僚1,2, 森圭一郎1,2, 内村直尚1,2 ○Youhei Ishii1, Kiichiro Morita1,2, Yoshihisa Shouji1,2, Ryo Fujiki1,2, Keiichiro Mori1,2, Naohisa Uchimura1,2 久留米大学　高次脳疾患研究所1, 久留米大学　医学部　神経精神医学講座2 Cognitive and Molecular Research Institute of Brain Diseases, Kurume University, Fukuoka, Japan1, Dept Neuropsychiatry, Kurume Univ, Fukuoka, Japan2 Near infrared spectroscopy (NIRS) is a noninvasive method to measure brain function. Changes in regional blood flow associated with neuronal activities of the brain can be evaluated according to changes in the Oxy-Hb level. In this study, we investigated changes in Oxy-Hb during single-verbal tasks on the basis of a method to measure event-related potentials in brain waves.The subjects were 20 healthy native Japanese speakers. The measurements were performed after explaining this study in writing and obtaining consent from all subjects.Changes in cerebral blood flow during the tasks were measured at bilateral recording points (22 on each side) using a multi-channel NIRS system. The subjects were instructed to phonate "a-i-u-e-o" for 12 seconds as a resting condition, and performed a verbal production task (VPT) or shiritori task (to say one word that starts with the last sound of the preceding word). The shiritori task involved two task patterns, i.e., standard shiritori (VST) and living creatures shiritori (VLT). From the data obtained at each measurement point during the 20 trials, an averaged waveform was prepared, and peak values of waveform were calculated and analyzed. The analyzed areas were middle prefrontal areas (left 11th channel, right 12th channel), prefrontal areas (left 19th channel, right 22nd channel) and association area (left 9th channel, right 5th channel).We found that the Oxy-Hb amplitude in left 11th channel significantly greater than in right 12th channel in the VLT. Furthermore, the Oxy-Hb amplitude during the VST and VLT was significantly increased in comparison with VPT. The Oxy-Hb latency during the VST and VLT was significantly slower than VPT in prefrontal areas.These findings suggest that single event related NIRS measurement using shiritori tasks is considered to be a useful examination as it shows high site specificity and provides event related potential like wave forms.	P1-2-254 マルチチャンネル・アレイコイルを用いたfMRIデータに対して領域間結合分析を行う際の注意点 A remark on dynamic causal modeling analysis with a multi-channel array head coil ○河内山隆紀1,2, 佐藤渉1, 島田育廣2, 藤本一郎2, 正木信夫2 ○Takanori Kochiyama1,2, Wataru Sato1, Yasuhiro Shimada2, Ichiro Fujimoto2, Shinobu Masaki2 京都大・霊長研・白眉1 Hakubi Proj, Kyoto Univ, Kyoto1, ATR-Promotions, BAIC, Kyoto2 A multi-channel array head coil is mainly used in recent fMRI experiments. It can perform better in signal to noise ratio terms than single channel coils. However, on the other hand, previous studies reported negative side effects associated with the spatial intensity inhomogeneities and the increase in the physiological noise. In this work, we address the problem of applying dynamic causal modeling (DCM) to the data collected using a multi-channel array head coil. The coupling parameters of DCM may be affected by the spatial intensity inhomogeneities since DCM estimates them based on fMRI signal amplitude at each location. We constructed the network model including the regions with the signal intensity bias and investigated whether the coupling parameters were changed with and without inhomogeneity correction.Four volunteers participated in this study. A simple visual stimulation task was conducted. Data acquisition was performed on a 3.0 Tesla scanner with a 32-channel head coil. The EPI images were reconstructed with or without a prescan normalization filter. First, task-related activation was evaluated using SPM8. DCM was then performed to estimate the coupling parameters between V1 and fusiform gyrus (FG). The same analysis was applied to the original and prescan normalized dataset.Although the results of the statistical map were almost the same between the original and prescan normalized dataset, the coupling parameters of DCM analysis were considerably different. In the original data, the coupling parameter from V1 (the signal is biased toward high intensity) to FG (low intensity) was decreased. When researchers plan to conduct DCM analysis on fMRI data collected using a multi-channel array head coil, it is important to correct the spatial intensity inhomogeneities in the data by a prescan normalization filter or other bias correction methods.

上部に戻る

前に戻る