2006年11月16日−11月17日
代表・世話人：本田 学（国立精神・神経センター）
所内対応者：定藤規弘（生理学研究所心理生理学）
- （１）
- Mapping cortical activity using fMRI in awake behaving monkeys
Naokazu Goda, Takuya Harada , Minami Ito, Hiroshi Toyoda, Norihiro Sadato, Hidehiko Komatsu(NIPS)
Tadashi Ogawa(Kyoto Univ.)
- （２）
- Network extraction using exploratory multivariate analysis of intersubject variability in task-related activation
Motoaki Sugiura(Dept. Cerebral Research, NIPS, Okazaki, Japan)
- （３）
- A Toolbox for NN-ARX Modeling of fMRI Time Series
Bosch,J., Riera,J., Biscay,R., Galka,A., Valdes,P., Sadato,N. and Ozaki,T.
（情報システム研究機構，統計数理研究所）
- （４）
- Motion correction in k-space for fMRI
Costagli M, Waggoner RA, Ueno K, Tanaka K, Cheng K(RIKEN Brain Science Institute)
- （５）
- Electromagnetic Source Imaging: Resolution Spread Function-Constrained and
Functional MRI-Guided Spatial Filtering
Xiaohong Wan(RIKEN Brain Science Institute)
- （６）
- Hemodynamic correlates of human sleep spindles: A simultaneous EEG-fMRI study
Takeyuki Mori, Noriko Sato(National Center of Neurology and Psychiatry)
Kimitaka Anami(Takatsuki Hospital)
- （７）
- Sex differences in automatic emotion processing: Association with harm avoidance
Shigeru Toki, Yasumasa Okamoto, Keiichi Onoda, Akiko Kinoshita, Shigeto Yamawaki
(Department of Psychiatry and Neurosciences, Graduate School of Biomedical Sciences, Hiroshima University)
Hiroshi Yoshida(Department of Social and Clinical Psychology,
Faculty of Contemporary Culture, Hijiyama University)
- （８）
- 脳機能計測を目指したMRSによる非侵襲的脳内温度測定
吉岡芳親1,5，高濱祥子1,5，及川浩2,5，神原芳行1,5
松村豊1,5，江原茂1，関淳二3,5，精山明敏4,5
（1岩手医科大学先端医療研究センター，2岩手県立二戸病院放射線科，
3国立循環器病センター研究所 生体工学部，4大阪大学大学院生命機能研究科，
5科学技術振興機構 CREST）
- （９）
- 赤外線カメラによるラット神経活動に伴う脳表温度変化の計測
駒井豊1，大井康浩2，精山明敏3，関淳二1
（1国立循環器病センター研究所生体工学部，2大阪大学歯学系研究科，3大阪大学生命機能研究科）
- （10）
- The role of the human left ventral premotor cortex for exact enumeration of successive stimuli
Kenji Kansaku 1,2,3
(1 Cognitive Functions Section, Dept of Rehabilitation for Sensory Functions, Res Inst of National Rehabilitation Center for Persons with Disabilities (NRCD), Tokorozawa, Japan
2 Division of Cerebral Integration, Dept of Cerebral Research, NIPS, Okazaki, Japan
3 Human Motor Control Section, Medical Neurology Branch, NINDS, Bethesda, USA)
- （11）
- A motor network study by concurrent fMRI and TMS technique
Takashi Hanakawa, Tastuya Mima, Mitsunari Abe, Kimitaka Anami,
Shin-ichi Urayama, Manabu Honda, Hidenao Fukuyama
（国立精神・神経センター 神経研究所 疾病研究第七部
京都大学医学研究科附属高次脳機能総合研究センター）
【参加者名】
上田一貴（広島大学大学院教育学研究科），天野 薫（NTTコミュニケーション科学基礎研究所），加美由紀子（九州大学大学院歯学研究院），尾崎 統，Bosch,Jorge，Biscay,Rolando，Kevin,Won（情報システム研究機構統計数理研究所），山村裕美（筑波大学大学院人間総合科学研究科），松田佳尚，小川昭利，COSTAGLI,Mauro，万 小紅（理化学研究所脳科学総合研究センター），作田由衣子，熊谷智博（東北大学大学院文学研究科），土岐 茂（広島大学大学院精神神経医学講座），平野好幸（放射線医学総合研究所分子イメージング研究センター），森 健之（国立精神・神経センター武蔵病院放射線診療部），兒玉直樹（産業医科大学神経内科），吉岡芳親（岩手医科大学先端医療研究センター），金 完哉，渡邊和子（岐阜大学医学系研究科生理学分野），加藤 誠（情報通信研究機構未来ICT研究センター），本田 学，花川 隆（国立精神・神経センター神経研究所），関 淳二，駒井 豊（国立循環器病センター研究所生体工学部），橘 篤導（神奈川歯科大学生体機能学講座），郷田直一（生理学研究所感覚認知情報），神作憲司（国立身体障害者リハビリテーションセンター研究所），水谷 勉（茨城大学大学院），田中靖人，劉 国相（情報通信研究機構バイオICTグループ），定藤規弘，杉浦元亮，田邊宏樹，豊田浩士，齋藤大輔，原田宗子，大塩立華，中下 悟，村瀬未花，内山仁志，米田英嗣，森戸勇介，牧 陽子，間野陽子，出馬圭世，谷中久和，酒井朋子，林 正道，佐々木章宏，藤井 猛（生理学研究所心理生理学）
【概要】
1990年に発見されたblood oxygen level dependent (BOLD)効果を主な原理とする機能的磁気共鳴画像法（以下fMRI）は近年著しく普及し，ヒトを含む霊長類の脳機能を非侵襲的に探る上で卓越した可能性をもつことが明らかとなってきた。その反面，いまだ手法自体の歴史が浅いため，適切な撮像方法や統計的解析法といった技術面のみならず，記録される信号変化のもつ生理学的意味（即ち局所脳血流あるいは電気的活動との関係など）を理解する上でも検討すべき課題が山積している。本研究会では，このように萌芽的側面を有するfMRI手法を，技術的ならびに生理学的な諸課題について活発な議論をおこなうとともに，情報交換の場を提供することを目的とする。具体的には，現在，fMRI研究を盛んに実施している研究施設や装置開発など関連領域の研究者並びに技術者から，fMRIの記録・解析における技術的側面，ならびにBOLD効果などの生理学的側面に関連する研究成果を募集し，発表ならびに討議をおこなう。こうした活動を通して，fMRIを神経科学の確固たる道具として育て上げることを目標とする。平成18年11月16日−11月17日の2日間に渡り，所内外あわせて59名の参加者を得て，第5回fMRI研究会を行った。fMRIに関する技術的ならびに生理学的な諸課題について活発な議論をおこなうとともに，情報交換の場として有効に機能した。具体的には，現在，fMRI研究を盛んに実施している研究施設や装置開発など関連領域の研究者並びに技術者による，fMRIの記録・解析における技術的側面，ならびにBOLD効果などの生理学的側面に関連する研究成果11演題について，口頭発表ならびに討議を活発に行った。例年同様，国内研究機関に在籍する外国人研究者の参加が多かったため，出来るだけ英語による発表・討論を行うことに留意した。fMRIの技術的，生理学的側面に焦点を絞った研究会は国内に類をみず，会終了後のアンケートでも継続開催の希望が多数出ていた。
Naokazu Goda, Takuya Harada, Minami Ito, Hiroshi Toyoda, Norihiro Sadato, Hidehiko Komatsu(NIPS)
Tadashi Ogawa(Kyoto Univ.)
FMRI study in awake behaving monkey is expected to be a new valuable experimental approach for investigating large-scale brain function, which is complementary to electrophysiological recordings from relatively small number of neurons. Here, we report the results of preliminary fMRI in the awake monkeys aimed to map retinotopy and motion- as well as form-selective regions in the visual cortex, with our experimental setups developed for fMRI in the awake monkeys.
Two monkeys were trained to fixate for 1-4s at a central fixation spot. The monkey viewed a screen in the 'sphinx' position inside a custom monkey chair. The head was fixed rigidly with the chair. Eye position was measured by tracking the pupil with an IR video camera. We presented stimuli of the different conditions (checkerboard wedges for retinotopic mapping, moving/stationary dots and object/noise pictures for measuring motion- and form-selective activities, respectively) in a block-design; each block consisted of 4 trials, in each of which the stimulus was presented for 0.5-3.5s during the fixation. BOLD responses during the blocks were measured using a 3T scanner and a surface coil with EPI sequences optimized to reduce susceptibility artifact. The data with large head/body motion were excluded from the analysis.
We obtained the retinotopic maps in areas V1/V2/ V3/V3A/V4 in good agreement with the maps reported by previous electrophysiological studies. We found robust motion-selective activities within superior temporal sulcus (STS) including area MT, and form-selective activities as patches within STS and gyrus in the inferior temporal cortex, largely consistent with previous electrophysiological studies. Interestingly, some of these were observed in the regions where electrophysiological recordings were not explored. These results demonstrate the feasibility and the utility of fMRI for functional mapping in the monkey visual cortex.
Motoaki Sugiura(Dept. Cerebral Research, NIPS, Okazaki, Japan)
In complex environment in the daily life, perception and behavior differ among individuals. Each brain may respond incidentally or engage complementary processes to different extents for a single behavioral goal. This entails intersubject variability in activation and reduces the sensitivity of the conventional intersubject analysis of the functional imaging study, which tests the average between-condition difference in activation, treating the between-subject effect as noise. A novel approach to identification of brain networks using this intersubject variably has been recently proposed (Sugiura et al., Human Brain Mapping 2006, in press). The assumption is simply that a network for a specific cognitive mechanism will be activated differently in different subjects and that this will be expressed as a systematic pattern of intersubject variability. A principal component analysis (PCA) is used to summarize patterns of variability, over subjects, in regions of interest (ROIs) showing task-specific effects. Inclusion of multiple ROIs and task conditions enables to extract patterns of intersubject variability in distributed networks induced by different conditions. The principal component scores enters a voxel-by-voxel multiple regression analysis of subject-specific activations (i.e., contrast images). The resulting statistical parametric maps identify cortical networks that participate in the principal modes of intersubject variability. The functional attribution of these principal components is based on the functional selectivity of the ROIs used in the PCA. Examples of the successful applications of this approach will be presented and the limitation of the technique will be discussed.
Bosch,J., Riera,J., Biscay,R., Galka,A., Valdes,P., Sadato,N. and Ozaki,T.
（情報システム研究機構，統計数理研究所）
A Matlab software is presented to analyze fMRI data on the basis of the NN-ARX approach[1]. NN-ARX is a voxel-wise methodology for estimating both the brain responses to stimuli contained in fMRI data, as well as gathering areas of the brain highly correlated. In contrast with other models, NN-ARX takes also into consideration dynamic aspects of consecutive fMRI scans interpreted as a multiple time series. The software includes facilities for data preprocessing (SPM like) but can also accepts the output from SPM preprocessing. Automated brain segmentations as well as coregistration to Talairach space are provided. The software offers options for the estimation of the optimal model parameters. To estimate the correlated areas, interesting voxels are selected as reference. The voxels of interest can be taken either from the voxels that show higher stimulus activation values or the voxels with higher innovations (the residual after model estimation). The HRF is also estimated for each experimental condition and shown as a 3D spatial map. A Hotelling T2 is used to assess the significance of the stimuli related activations as well as to assess differences between conditions. By pinpointing with the mouse over the spatial maps, a specific voxel can be picked up and its specific HRF is displayed for each condition. Information about the reference voxels is provided and the user can study them separately. Also, for a selected voxel, the optimal model selection, the parameters, the original and the estimated signals are shown, as well as the HRF for each condition, the residuals and some statistical tests about the residuals.
[1]fMRI activation maps based on the NN-ARx model, Riera-Diaz, J.J.; Bosch-Bayard, J.; Yamashita, O.; Kawashima, R.; Sadato, N.; Okada, T.; Ozaki, T. Neuroimage, 23(2), (2004), 680-697
Costagli M, Waggoner RA, Ueno K, Tanaka K, Cheng K(RIKEN Brain Science Institute)
Unlike the most widely used applications such as AFNI and SPM that operate in image domain, the algorithm presented here corrects 3D rigid-body motion by operating in k-space, thus decoupling and independently correcting rotation and translation effects.
Rotations between two volumes to be aligned are visible in their 3D-FFT magnitude data, where no contribution from translational effects is present. Since rotations in k-space are always about the origin of the spatial frequency axes, we estimate them by aligning spherical shells of equal radius, centered at the origin, encoding magnitude Fourier data of the volumes to be registered. After correcting rotations, we exploit the Fourier shift theorem for treating translations: a shift along any spatial axis introduces a linear phase change along the corresponding frequency axis, which is estimated by considering the cross power spectrum (CPS) of the original and the translated volumes. Translation correction is achieved by fitting a linear function (a hyperplane in 3D) to the phase of the CPS, adding it to the Fourier transform of the test volume, and finally inverse Fourier-transform the result. The method is non-iterative, and can accurately detect subvoxel translations.
The algorithm was tested on both synthetic and real EPI time series, and compared to the registered data obtained with AFNI and SPM2. Better volume registration was achieved by serial application of motion correction in k-space and then in image-space, the former step being more robust to local minima and correcting for larger motions, and the latter seeming more suitable for fine corrections.
Xiaohong Wan(RIKEN Brain Science Institute)
The problem of spatial ambiguity in electromagnetic source imaging has been unresolved. Our heuristic analysis points out the discrepancy between the estimated source and the real source arises from the other sources’ interference. We suggest a novel and simple solution based on the principle of spatial filtering combined with the Backus-Gilbert method by minimizing the estimated source power and effectively suppressing the other sources’ interference. Within this framework, the functional MRI information can be also effectively integrated into the inverse solution and the source localization is not biased by incompatible knowledge. Our Monte Carlo simulations of EEG system based on a realistic head model show that the resolution spread function-constrained and functional MRI-guided spatial filtering provides precise source localization, even in the cases of multiple simultaneously active sources.
Takeyuki Mori, Noriko Sato(National Center of Neurology and Psychiatry)
Kimitaka Anami (Takatsuki Hospital)
The EEG-fMRI is being applied to the examinations of the distributed neural networks related to spontaneous neural activities such as awake (alpha rhythms etc.) and sleep (sleep spindles, K complexes etc.) electrophysiological phenomena as well as epileptiform activities. In the present study, we tried to utilize the EEG-fMRI for mapping hemodynamic correlates of sleep spindles.
The sleep spindle is a distinct EEG event observed in non-REM sleep, especially in Stage II of the sleep cycle. Although sleep spindles have been considered to play a role in sleep maintenance, their physiological implication still remains to be clarified. We investigated the hemodynamic correlates of spindles using an EEG-fMRI measurement in 43 normal volunteers. Twelve out of the 43 subjects succeeded in reaching Stage II sleep during fMRI acquisition. We found positive spindle-related BOLD signal changes in the early sensorimotor areas, auditory areas, visual areas, anterior cingulate cortex, posterior cingulate cortex and the thalami. The results indicate that extensive cortical and subcortical areas are involved in the formation of spindles. We suggest that such spindle-related neural substrates play important roles in sensory processing and sleep protection during non-REM sleep.
Shigeru Toki1, Yasumasa Okamoto1, Keiichi Onoda1,
Akiko Kinoshita1, Shigeto Yamawaki1, Hiroshi Yoshida2
（1Department of Psychiatry and Neurosciences,
Graduate School of Biomedical Sciences, Hiroshima University
2 Department of Social and Clinical Psychology,
Faculty of Contemporary Culture, Hijiyama University)
Background: Neural science revealed that females appear to be more sensitive and responsive to social, emotional information, including facial expression, than are males. Although subcortical, automatic, face processing has been considered to take an important role in face perception, no neuroimaging studies have been done to elucidate sex differences in subconscious emotion processing and its relationship with personality variables.
Methods: Using fMRI, we investigated the neural responses associated with the subconscious (backwardly-masked) perception of fearful, happy, disgust and neutral faces in 32 healthy volunteers (16 males, 16 females) who varied in harm avoidance, associated with trait anxiety and also threat processing.
Results: When subtracting the activation values of males from those of females, suprathreshold positive signal changes were detected in the right medial frontal gyrus during masked happy presentation, and in the left superior parietal lobule and the right parahippocampal gyrus during masked fear presentation. When subtracting the activation values of females from those of males, suprathreshold positive signal changes were detected in the right anterior cingulated and the right caudate during the masked disgust presentation. While, in females, the harm avoidance positively correlated both, with happy-related neural activation in the right medial frontal gyrus, and with fear-related activation in the right parahippocampal gyrus, in males, there was no such correlation.
Conclusions: Our findings suggest sex-related neural responses to emotional stimuli and could contribute to the understanding of mechanisms underlying sex-related vulnerability of the prevalence and severity of anxiety and mood disorders.
吉岡芳親1,5，高濱祥子1,5，及川浩2,5，神原芳行1,5，
松村豊1,5，江原茂1，関淳二3,5，精山明敏4,5
（1岩手医科大学先端医療研究センター，2岩手県立二戸病院放射線科，
3国立循環器病センター研究所生体工学部，4大阪大学大学院生命機能研究科，
5科学技術振興機構 CREST）
体内温度は生命活動の基盤となる物理量であり，組織の活動度，血流，環境に影響されながら統合的・有機的に制御されていると考えられ，生理的条件下の脳においても部位差があるし活動により変動していると考えられる。安静時の脳は基礎代謝の約20%を占めており（約20ワット），同等の産熱が行われている。また，活動時には代謝が10%前後変化すると考えられており，温度は脳活動を評価する一手段となると思われる。しかし，精度の高い非侵襲的な脳内温度計測は容易では無く，脳内温度に関する情報は非常に乏しいのが現状である。温度は定量的に取り扱う事が可能であり，精度良く温度計測ができるようになれば，安静時の脳活動レベル，個体間の比較，脳活動の絶対量を評価できる可能性が有る。私たちは，磁気共鳴法で得られる温度情報を用いた脳機能評価方法を模索しており，これまでに行ってきた磁気共鳴法を用いた脳内温度計測について発表する。
駒井豊1，大井康浩2，精山明敏3，関淳二1
（1国立循環器病センター研究所生体工学部，2大阪大学歯学系研究科，3大阪大学生命機能研究科）
脳機能を可視化する新しい手法として，脳局所温度を計測する方法が提案されている。しかし，その基礎となる生理学的データは限られている。近年，高感度の赤外線カメラ（インジウムアンチモン）が実用化されたことにより，脳の局所的な活動に伴う温度変化を脳表で探知することが可能となった。脳表の温度変化は，脳神経活，及び，それとカップルする脳血流変化によってもたらされると考えられることより，脳神経活動の起きた領域，及び，強度を推測する有効な情報と考えられる。
本研究では，麻酔下のラットの下肢に電気刺激(3mAx5)を与え，誘起される脳感覚野での温度変化を経頭蓋，及び，頭蓋骨除去後の硬膜表面で計測した。経頭蓋での計測では，最も変化の顕著な部位で，刺激開始から2.9秒後に0.065度の上昇，頭蓋除去後には2.0秒後に0.087度の温度上昇が観察され，後者の方がより鋭い空間分布を示した。本計測系は，局所的な温度変化を探知するのに十分な感度と再現性があることが示された。
Kenji Kansaku 1,2,3
(1Cognitive Functions Section, Dept of Rehabilitation for Sensory Functions,
Res Inst of National Rehabilitation Center for Persons with Disabilities (NRCD), Tokorozawa, Japan
2Division of Cerebral Integration, Dept of Cerebral Research, NIPS, Okazaki, Japan
3Human Motor Control Section, Medical Neurology Branch, NINDS, Bethesda, USA)
Adult humans have the ability to count successive stimuli exactly. What brain areas underlie this uniquely human process? To investigate the neural basis, we first used functional magnetic resonance imaging (fMRI) and investigated brain areas involved in counting small numbers of successive stimuli up to 4, and demonstrated activations in the bilateral lateral premotor cortex1. We further applied fMRI and found that the upper part of the left ventral premotor cortex was preferentially activated during counting of successive sensory stimuli presented 10 to 22 times, while the area was not activated during small number counting up to 4. We then used transcranial magnetic stimulation (TMS) to assess the necessity of this area, and found that stimulation of this area preferentially disrupted subjects’ exact large number enumeration (10-22). Stimulation to the area affected neither subjects’ number word perception nor their ability to perform a non-numerical sequential letter task2. The results suggest that the left ventral premotor cortex is indispensably involved for exact counting of large numbers of successive stimuli.
1. Kansaku, K. et al. Neural correlates of counting of sequential sensory and motor events in the human brain. Neuroimage 31, 649-60 (2006).
2. Kansaku, K, et al. The role of the human ventral premotor cortex in counting successive stimuli. Exp Brain Res 178, 339-50 (2007).
Takashi Hanakawa, Tastuya Mima, Mitsunari Abe, Kimitaka Anami,
Shin-ichi Urayama, Manabu Honda, Hidenao Fukuyama
（国立精神・神経センター 神経研究所 疾病研究第七部
京都大学医学研究科附属高次脳機能総合研究センター）
The concurrent functional MRI and transcranial magnetic stimulation technique can potentially visualize the whole brain networks connected to the stimulated site. However, the quantitative relationship between TMS intensity and BOLD signals in the stimulated and remote areas is yet to be elucidated. To answer this question, fifteen healthy subjects were scanned on a 3-T MRI scanner. An MRI-compatible figure-of-eight TMS coil was attached to the subject’s scalp site adequate for eliciting right thumb movement. BOLD measurement was performed with the gradient-echo EPI (TR=2.7 s/TE=30 ms) modified for “stepping-stone”sampling, which enabled online monitoring and recording of motor evoked potentials during scanning. A single TMS pulse was delivered during the delay periods (200 ms) between each volume acquisition. The intensity of TMS pulses was varied from 30% to 95% of machine output (110% with a booster) at a 5% step (mean frequency at ~0.15 Hz). SPM analyses revealed activation of motor and sensory networks as well as auditory areas during TMS stimulation with higher intensities. BOLD signal changes as a function of TMS intensities were assessed in the directly stimulated left primary motor cortex (M1) and other remote areas including supplementary motor areas (SMA). A sharp increase in BOLD signals was observed in M1 only above the 80% stimulation corresponding to the resting motor threshold. The remote motor areas, SMA, for example, showed an increase in signals even below the motor threshold. Further investigation will be necessary to clarify the energy requirement in the direct and remote areas influenced by TMS.
