2018/10/05から2018/10/08に東京大学にて開催された fNIRS2018に研究室から4件の発表を行いました。
- A fNIRS-based hyperscanning study of inter-brain neural synchronization during a cooperative task Megumi Mizuno, Sho Taniguchi, Satoru Hiwa,Tomo Hiroyasu
- Detecting attentional and inattentional brain metastates based on dynamic functional connectivity analysis, Miyu Nishizawa , Satoru Hiwa,Tomo Hiroyasu
学会参加報告書
| 報告者氏名 | 水野めぐみ |
| 発表論文タイトル | 協調課題時の脳間神経同期のfNIRSを用いたhyperscanningの研究 |
| 発表論文英タイトル | A fNIRS-based hyperscanning study of inter-brain neural synchronization during a cooperative task |
| 著者 | Megumi Mizuno, Sho Taniguchi, Satoru Hiwa,Tomo Hiroyasu |
| 主催 | The Society for fNIRS |
| 講演会名 | fNIRS2018 |
| 会場 | 東京大学 本郷キャンパス |
| 開催日程 | 2018/10/05-2018/10/08 |
- 講演会の詳細
2018/10/05から2018/10/08にかけて，東京大学 本郷キャンパスにて開催されましたfNIRS2018に参加いたしました．この学会は，光学的方法を用いて生物組織、特に脳機能を理解しようとする基礎科学者および臨床科学者の専門家が意見交換をすることを目的に開催されています．私は，2018/10/05から2018/10/08の全日参加致しました．本研究室からは他に廣安先生，日和先生，西澤美結，池田幸樹，山本渉子，谷口尚が参加しました．
- 研究発表
- 発表概要
私は2018/10/07のPoster Session 「PosterⅡ」にて谷口尚と共に発表致しました．発表の形式はポスター発表で，午前,午後1時間ずつの計2時間，自由に発表および質疑応答を行う時間となっておりました．
今回の発表は，「A fNIRS-based hyperscanning study of inter-brain neural synchronization during a cooperative task」です．以下に抄録を記載致します．
| Introduction: Social interaction is a dynamic behavior between individuals who modify their actions and reactions depending on the actions of their partner. In this study, we investigate the relationship between social interaction and brain functions. Cui et al. analyzed the neural synchronization between two subjects who played a cooperation game involving synchronizing each other’s response timing and revealed that the interpersonal brain coherence of the subjects increased during the cooperative task [1]. In this study, to easily facilitate the cooperative behavior of the participants, we improved the experimental design and examined the inter-brain neural synchronization during the cooperative task. Methods: Twenty-two healthy adult males (11 pairs, age: 22.7 ± 1.0 years old, right-handed) participated in this experiment. The experimental environment is shown in Fig 1. The brains of the two participants during their social interaction were simultaneously measured by hyperscanning using a single functional near-infrared spectroscopy (fNIRS) device (ETG-7100, Hitachi, Ltd.). A 3 × 10 probe consisting of 47 measurement channels was attached to the forehead of each subject. Participants were instructed: (1) to synchronize the response timing of their partner after the cue (synchronization task), (2) to respond faster than their partner (competition task), and (3) to respond quickly to the cue (single A/B task) [1]. In the synchronization task, the time difference between the responses of the two participants were fed back to each of them. Each participant predicted his partner’s behavior based on the time difference, and was asked to synchronize his responses at the next cue. Wavelet transform coherence was calculated from two sets of time-series data of cerebral blood flow changes. Then we performed a one-sample t-test (p < 0.05) of the coherence increase of each task and investigated the brain regions where the coherence increased. Results & Discussion: The difference in response time between the two participants in the synchronization task was significantly larger than that of the single A and B task (p < 0.05). This result indicates that the participants not only responded quickly to the cue, but also reacted by anticipating the behavior of their opponents. A t-score map of coherence increase is shown in Fig 2. The coherence within the left superior frontal gyrus (SFG) and left middle frontal gyrus (MFG) increased significantly in the synchronization task. The left SFG is involved in building a trust relationship [2]. Hence, it is assumed that this region affects cooperative behavior. Furthermore, the coherence in these two regions did not increase significantly in either the competitive or single task. Therefore, we suggest that the inter-brain synchronization in these brain regions can be utilized as a metric of cooperativeness. |
- 質疑応答
今回の講演発表では，以下のような質疑を受けました．
・質問内容1
質問者の氏名を控え損ねてしまいました．
協調と競争を比較してwavelet coherenceで評価している先行研究が既に報告されているが、新規性は何か。この質問に対して、先行研究の実験を模倣しているが、より多くの領域を計測していますと回答しました．
・質問内容2
質問者の氏名を控え損ねてしまいました．
coherenceは行動のアクティベーションを検出してしまわないか．競争はより行動が同期しているため比較できないのではないかとご意見を頂きました．Coherenceによる評価の有用性を検討している人もいるため，評価方法の検討も行っていく必要があると思いました．
・質問内容3
質問者の氏名を控え損ねてしまいました．
Single課題はどうなったのか。なぜsingle課題を行ったのか。この質問に対しては、ただ単に合図に対して応答するという要素を比較するために行った．Singel課題の結果は、singleAとBのコンディションは同じなのに，t-mapがかなり異なったため，状態を推定することが出来なかった．これは被験者の個人のEQSスコアの違いに現れていて，この違いが脳状態に影響を与えたと考えていると回答しました．
・質問内容4
ヤクルトの研究所の方からご意見を頂きました．
競争状態が好きな人と協調状態が好きな人の主観的な状態などが結果に影響したりしないのか．実験実施後にアンケートを心理評価のアンケートを行ってみてはどうか．確かに協調状態だけが2者の関係性の中で良い状態であるかは分からないため，アンケートに関して今後の参考にしますと回答しました．
・質問内容6
質問者の氏名を控え損ねてしまいました．
なぜ協調時にのみMFG.Lの有意なコヒーレンス増加が見られたと考えているかとご質問頂きました．この質問に対しては，MFG.Rは協調時と競争時の両方に見られているため，他者のことを考えることに関与しており，MFG.Lはその中でも協力することに関わる領域であった可能性があると考えていると回答しました．一方で，心の理論に関わるMFGの左右差について調査していく必要があると考えています．
- 感想
2回目の国際学会，3回目のポスター発表ということもあり，これまでの経験を生かして落ち着いて発表することができました．質問に対しても何かしらの回答をすることができたのではないかと思います．多くの人に興味を持って頂き、hyperscanningの研究を行っている方々と交流することができ、大変有意義な発表となりました．この発表を通じて、自分の研究がhyperscanning研究の標準のレベルにまで持っていくことが出来たのではないかと、この1年間の取り組みの成果を実感することができました．
- 聴講
今回の講演会では，下記の4件の発表を聴講しました．
| 発表タイトル ： Verifying Wavelet coherence analysis on fNIRS data using pseudorandom visual stimulation sequence 著者 ： X. Zhang, J. A. Noah, S. Dravida, Y. Ono and J. Hirsch セッション名 ：PosterⅠ Abstruct ： Neural mechanisms that underlie dynamic interpersonal interactions are thought to include temporally synchronous signals that reflect coupled processes between two brains [1]. Wavelet coherence analysis of hemodynamic signals acquired simultaneously during hyper-scanning experiments has been proposed for analysis of these neural processes [2,3]. However, this computational approach has not been validated against known markers. Here we generate a set of known neural coherences using pseudo random sequences of reversing checkerboard patterns. It is expected that input sequences that were more highly correlated generated higher cross-brain output correlations of wavelets in the visual cortex across pairs of subjects. Each visual stimulus event was a 2-second full-field reversing checkerboard pattern. Three random sequences, called sequence A, B and C (inset in figure), of such visual stimuli were presented for 2 minutes and repeated twice. The sequences were generated to have varying levels of correlation (e g: A-B more correlated than A-C). The expected coherence (left panel) was obtained by coherence analysis on the modeled fNIRS waveform, which was generated by convolving the stimulus sequence and the hemodynamic response function (HRF).The overall expected coherence is summarized as the average coherence within 0-30 second wavelength range (horizontal lines and numbers in left panel). Ten subjects participated. Subject’s data were randomly paired with every other subject’s data, resulting in ninety possible pairs. The combinations included A-A, B-B, or C-C, where both subjects viewed identical sequences of the stimuli (green). The other pairings were AB(red), A-C (blue) and B-C (cyan). The wavelet coherence results from measured fNIRS deOxyHB data (right panel) confirmed the expected coherence. Our results validate wavelet coherence as a technique for quanitfying the coherence of brain signals across participants during hyperscanning to study social interaction in ecologically valid paradigms. |
Wavelet coherenceが偽陽性を検出しているのではないかとレビューでの指摘もあったため、とても興味深い研究でした．指タップの同期とコヒーレンスの同期が類似していたという，いい結果が示されていました．
| 発表タイトル ： Interpersonal brain synchronization during social cooperation in children with Autism: a hyperscanning study using fNIRS 著者 ： Vanessa Reindl, Jana A. Kruppa, Julia Prinz, Eileen Weiß, Christian Gerloff, Wolfgang Scharke, Beate Herpertz-Dahlmann, Kerstin Konrad and Martin Schulte-Rüther セッション名 ： PosterⅡ Abstruct ： Background: Interpersonal brain synchronization has repeatedly been demonstrated during joint social tasks for healthy adults and recently also for parent-child dyads [1]. Pioneering studies have demonstrated diminished brain-to-brain synchrony in adults with autism spectrum disorder (ASD) during such tasks [2]. To date, no study has investigated this in children with ASD and has examined whether the familiarity of the interaction partner modulates interpersonal brain synchronization. Methods: Using functional near-infrared spectroscopy (fNIRS) hyperscanning, we assessed brain-to-brain synchrony in prefrontal brain regions during a cooperative and a competitive computer task in 43 typically developing (TD) children and 15 children with ASD (8-18 years) while playing with one of their parents. Adult strangers performed the identical tasks with each child. Participants were instructed to either respond jointly via button press in response to a target (cooperation) or to respond faster than the other player (competition). Within each dyad, wavelet coherence was calculated for oxy-hemoglobin brain signals of corresponding channels as a measure of brain-to-brain synchrony. Results: On the behavioral level, preliminary results showed that the dyad’s cooperative performance was neither influenced by interaction partner nor by group. However, during competition, the child won more often against the parent than against the stranger, and children with ASD won more often against parent/stranger than TD children. On the neural level, preliminary results revealed a significant interaction of partner and group for coherence in two channels located in Brodmann areas 8 and 9: coherence in the ASD group was significantly smaller when playing with the parent compared to a stranger. No significant effect of partner was observed in the TD group. Conclusion: Data collection in both samples is ongoing. Preliminary results suggest differential coherence in ASD with respect to the interaction partner’s familiarity. In a larger sample it remains to be seen, whether fNIRS hyperscanning represents a valuable tool for investigating brain-to-brain synchrony during social tasks as a proxy for typical and atypical social interaction. |
学会を通してfNIRSの利点を生かして子供を対象にしている研究が多いと感じました．中でもこの発表は，協調時と競争時の親子と知らない大人と子供の神経同期をCuiの実験を子供でも分かるゲームにして実験を行い比較していました．
| 発表タイトル ： THE NEW NEUROSCIENCE OF TWO: HYPERSCANNING WITH fNIRS TO UNDERSTAND COMMUNICATING BRAINS 著者 ： Joy Hirsch, Ph.D. セッション名 ： Cognitive & social neuroscience [invited talk] Abstruct ： Humans are a profoundly social species. However, little is known about the most fundamental brain functions that mediate live social interactions. This knowledge gap between single and social brain processes, is, in part, a consequence of conventional neural imaging methods that are generally restricted to single individuals, static tasks, and nonverbal responses. However, recent developments of fNIRS hyperscanning (imaging two individuals simultaneously) enables rigorous investigation of this unexplored neural domain of human social behavior. Two existing theoretical frameworks converge as a foundation for this new “neuroscience of two”. The first is the Interactive Brain Hypothesis (De Jaegher, et al, 2016) which proposes that social interactions are mediated by dedicated brain processes. This hypothesis provides a general framework for the investigation of neural mechanisms underlying social interaction. The second is a quantifiable model of dynamic neural coupling, i.e. the correlation between temporal patterns of neural signals across two interacting partners (Hasson and Frith, 2016). This model proposes that dynamic coupling is the mechanism by which information is shared between the sender and the receiver across brains. Hyperscanning with fNIRS, together with these emerging theoretical frameworks, establishes fNIRS as the neuroimaging method of choice for investigations of live human-to-human interactions. Paradigms for investigation of two individuals engaged in live and spontaneous communications include developments for imaging in natural environments as well as the computational methods necessary to quantify live cross-brain interactions. We have applied these tools to test specific cases of the interactive brain hypothesis using fNIRS hyperscanning and primary social cues such as real eye-to-eye contact compared to eye gaze at a picture-face (Hirsch, et al., 2017), a video-face (Noah, et al.,2018), and talking and listening (Hirsch, et al., 2018) with and without social interaction. Consistent with the Interactive Brain Hypothesis cross-brain coherence measured by Wavelet Coherence Analysis (MATLAB Wavelet Toolbox) between local brain areas is greater for interactive than non-interactive conditions. Specifically, dynamic neural coupling increases during interactions such as eye-to-eye contact and interactive speaking and listening. Further, the dynamic neural coupling is associated with neural activity in temporal-parietal regions of the brain consistent with specialized cross-brain neural mechanisms for live social interaction regardless of the modality or task. These, and other similar findings recently reported using fNIRS hyperscanning, contribute to an expanding experimental and theoretical foundation for a new neuroscience where the aim is to understand the dynamic signal exchanges and neural mechanisms within a dyad that underlie live episodes of social interaction in natural conditions. |
fNIRS hyperscanningの動向を知ることができて興味深い講演でした．CooperationとCompetitionの比較の研究は最も多いテーマであることを知ることができました．一方でアイコンタクトの研究はほぼなく，昨年MISLで取り組まれていたアイコンタクトの研究はいいテーマだったのではないかと思いました．使用装置を研究室にある装置ばかりで実際に同じような実験を行うことも環境的には可能であることも分かりました．
| 発表タイトル ： Extraction of Synchronizing Cortical Activities between Mother and Infant 著者 ： Satoshi Morimoto, Eiichi Hoshino, Masahiro Hata, Michiko Asano and Yasuyo Minagawa セッション名 ： Poster Ⅲ Abstruct ：Background: Hyperscanning studies using functional near-infrared spectrography (fNIRS) have revealed synchronized brain activities between participants who interact each other. However, synchronizing activities between mother and infant could not be evaluated precisely by analysis methods used previously (e.g., wavelet coherence), because those methods assume that two brains have similar hemodynamics, while this is not the case for adult and infant brains. The present study investigated how mother and 3- to 4-month-old infant spontaneously interact each other at brain hemodynamic levels by employing a new analysis method (Fig.1) that we developed. Methods: 55 mother-infant dyads were participated in fNIRS recording under three conditions; breast feeding (the feeding condition), resting state during mother holding her sleeping infant (the holding condition), and resting state during separation (the separation condition) [1]. For the following analysis, we utilized data of dyads who completed each condition. After removing artifacts, we obtained power spectrograms of OxyHb signals and combined them into a large matrix. The matrix was decomposed by non-negative matrix factorization, which lead two smaller matrices; a basis matrix and a coefficient matrix. We calculated the mean of the coefficients for each condition and compared these values between the conditions. Results and discussion: We found several synchronized components which significantly increased in the feeding condition compared to those in other conditions. These components were shared by mother and infant channels, which related to the social brain network. These results suggested that we successfully extracted synchronized brain activities between mother and infant, and such synchronizations may be a fundamental of social attachment between them. |
異なる周波数成分の類似度を算出するために，Non-negative matrix functionを用いていました．Hyperscanningの研究では，解析方法に工夫が見られる研究があまりなかったため，一番面白い内容だと感じました．
参考文献
- fNIRS2018 abstract
http://fnirs2018.org/wp-content/uploads/2018/09/fNIRS2018_abstract.pdf
学会参加報告書
| 報告者氏名 |
西澤美結 |
| 発表論文タイトル | 動的機能的接続分析に基づく注意と不注意メタ状態の検出 |
| 発表論文英タイトル | Detecting attentional and inattentional brain metastates based on dynamic functional connectivity analysis |
| 著者 | 西澤美結, 日和悟, 廣安知之 |
| 主催 | Society for functional Near Infrared Spectroscopy |
| 講演会名 | fNIRS2018 |
| 会場 | the University of Tokyo |
| 開催日程 | 2018/10/5-2018/10/8 |
- 講演会の詳細
2018/10/5から2018/10/8にかけて，東京大学にて開催されましたfNIRS2018(http://fnirs2018.org/)に参加いたしました．この学会は，Society for functional Near Infrared Spectroscopyによって主催された研究会で，機能的近赤外分光法（fNIRS）の分野におけるあらゆるトピックに関する研究に関心のある研究者・医師・学生が参加し，fNIRSの可能性や方向性についての情報交換や議論の場になることを目的に開催されています．
私は5-8日の全日参加いたしました．本研究室からは他に廣安先生，日和先生，池田，水野，谷口，山本が参加しました．
- 研究発表
- 発表概要
私は7日のセッション「PosterⅡ」に参加いたしました．発表の形式はポスター発表で，午前と午後1時間の計2時間のポスター発表となっておりました．
今回の発表は，「Detecting attentional and inattentional brain metastates based on dynamic functional connectivity analysis」という題目で，発表を行いました．以下に抄録を記載致します．
| Introduction: Recent studies have revealed changes in intrinsic brain organization using dynamic functional connectivity (dFC) analysis [1]. However, temporal changes in functional network topology owing to extrinsic factors have not been carefully examined. In addition, a lack of extrinsic attention sometimes leads to reduced work productivity or car accidents. It is important to quantify one’s level of attention and distraction toward external stimuli in such situations. In this study, we propose a method for detecting changes in extrinsic attention using a dFC analysis and an evolutionary optimization algorithm. Methods: Healthy adults (18 males, aged 22.4 ± 1.0 years) participated in the study. The participants were asked to perform a psychomotor vigilance task (PVT) to induce sustained attention to external stimuli. Their brain activity was measured using a 116-channel functional near-infrared spectroscopy covering the entire brain. The measured cerebral blood flow changes were bandpass filtered, and time-varying FC matrices were calculated using a sliding-window approach and then binarized. Here, we assumed that there is a metastate that represents the characteristic network organization for each attentional and inattentional state. The metastate was modelled as a symmetric binary matrix, and its optimum structure was determined using a genetic algorithm. Metastate optimization was performed to maximize the similarity between the optimized metastate and the measured FC matrices obtained from the periods where the fastest or slowest 10% of the response times (RTs) were achieved. We assumed that the attentional and inattentional states could be observed in the fastest 10% of RTs and slowest 10% of RTs, respectively. Results: The derived attentional and inattentional metastates are shown in Fig. 1 (a). In the attentional metastate, the connections around the left and right midfrontal gyrus (L/R-MFG) were extracted as characteristic connections because MFG is involved in an interaction between the dorsal and ventral attention networks; it also plays a central role in the attention function. In the inattentional metastate, the right superior occipital gyrus (R-SOG) and the medial part of the left superior frontal gyrus (L-SFGmed) connection was extracted. The R-SOG is associated with the visual network, and the L-SFGmed is a component of the default mode network (DMN). Furthermore, a connection between the right supplementary motor area (R-SMA) and R-MFG was found in the attentional metastate，whereas a connection between the R-SMA and L-SFGmed was found in the inattentional metastate. SMA is associated with motion control based on memory．This suggests that the connection with SMA was altered depending on the level of attention. SMA was connected to the attention network in the attentional state and to the DMN in the inattentional state. The temporal changes in RT and the similarity between the metastate and time-varying FC matrices are shown in Figure 1(b). The similarity of the attentional state increased with the fastest 10% of RTs, and the similarity of the inattentional state was high in the case of the slowest 10% of RTs. We suggest that our proposed method can detect the attentional and inattentional state and evaluate the attention level. Even in the absence of behavioral data, such as RT, it was verified that the state of the human brain can be visualized by the occurrence rate of the metastate. [1] Honey, C. J., et al. “Predicting human resting-state functional connectivity from structural connectivity.” Proceedings of the National Academy of Sciences 106.6 (2009): 2035-2040. |
- 質疑応答
今回のポスター発表では，以下のような質疑を受けました．
・質問内容1
質問者の氏名を控え損ねてしまいました．こちらの質問は，この研究の展望は何かというものでした．この質問に対して私はこのメタ状態が定義できたことで，外因性注意における動的な脳状態のレベルを評価することができるため，交通事故の防止や作業効率の向上にもつなげることが可能であると回答しました．
・質問内容２
質問者の氏名を控え損ねてしまいました．こちらの質問は，何CH計測しているのかというものでした．この質問に対して私は116CHと回答しました．
・質問内容3
質問者の氏名を控え損ねてしまいました．こちらの質問は，functional connectivityとは何かというものでした．この質問に対してオキシヘモグロビンの時間的相関と答えましたが，シンクロナイゼーションの図との関わりが理解されませんでした．
・質問内容4
質問者の氏名を控え損ねてしまいました．こちらの質問は，注意と不注意の区間の当てはまる行列は何枚かというものでした．この質問に対して2枚と回答しました．
・質問内容5
質問者の氏名を控え損ねてしまいました．こちらの質問は，前処理は何を行っているのかというものでした．この質問に対して，バンドパスフィルタを用いてノイズ除去を行っていると回答しました．
・質問内容6
質問者の氏名を控え損ねてしまいました．こちらの質問は，RTとwindowsizeに差があるが問題ないのかというものでした．この質問に対して，刺激呈示時間の状態を最も反映する行列を採用し，また，SWCを計算するのに必要なサンプル数を考慮したwindowsizeになっていると回答しました．毎学会でwindowsizeについては質問され，そのたび解決できていないことにもどかしさを感じるので修士論文ではwindowsizeのサイズの検討まで行いたいと思っています．
・質問内容7
質問者の氏名を控え損ねてしまいました．こちらの質問は，RTfast10%とRTslow10%の値は具体的にどれくらいかというものでした．この質問に対して，被験者によってばらつくが，ポスターで発表しているfast10%は0.6-0.8[s],slow10%は0.1-0.3[s]と答えました.被験者で値が異なることに対して指摘を受けたため，この研究では，各被験者に注意不注意の2状態があると仮定しているため問題ないと答えました．
・質問内容8
質問者の氏名を控え損ねてしまいました．こちらの質問は，attention metastateとinattention metastateで検定を行っているかというものでした．この質問に対して，行っていないと答えました．
・質問内容9
質問者の氏名を控え損ねてしまいました．こちらの質問は，PVTはどのような状態を計測したいときに用いるのかというものでした．この質問に対して，注意の持続を促したいときに行う課題であると回答しました．
・質問内容10
質問者の氏名を控え損ねてしまいました．こちらの質問は，attention metastateとinattention metastateをどのように作成しているのかというものでした．この質問に対して，各被験者の注意，不注意区間での行列と最適化された行列の類似度が最大になるように最適なメタ状態を検出していると回答しました．
・質問内容11
質問者の氏名を控え損ねてしまいました．こちらの質問は，刺激は1つかというものでした．この質問に対して，1つであると答えました．
・質問内容12
質問者の氏名を控え損ねてしまいました．こちらの質問は，「r」は何かというものでした．この質問に対して，RTとoccurrenceの相関と答えました．
・質問内容13
質問者の氏名を控え損ねてしまいました．こちらの質問は，PVTは何分間行っているのかというものでした．この質問に対して10分間と答えました．
- 感想
educational courseに参加し，NIRSの基礎的な部分を学ぶことができました．基礎的な内容をもう一度聞くことによって，何があいまいになっているのかを理解することができました．確率的レジストレーション法の原理や計測方法に関しては見直す必要があると感じました．また，ポスター発表では，様々なバックグラウンドを持った方々に研究を説明する機会があり大変有意義な時間になりました．GAの説明部分を増やし，原理を理解できたことが前回からの成長だと思っています．ただ，英語で質問された時に英語で答えることが出来なかったので修士論文までに説明できるようにしたいと思います．また，修士論文に向けて，再度課題を洗い出し，目標を持って研究を進めていきたいと思います．
- 聴講
今回の講演会では，下記の4件の発表を聴講しました．
| 発表タイトル ： Dynamics of Functional Networks in the Developing Brain 著者 ： Fumitaka Homae セッション名 ： Neonatal, pediatric & developmental neuroscience I Abstruct ： Introduction: Structural and functional organization of the brain occurs rapidly in early infancy. Previous studies report that in neonates and young infants, this neural organization is sensitive to speech sounds, and related functions in distinct brain regions have been partially elucidated (Homae et al., 2014). It is known that short speech sounds induce eventrelated activation in the temporal brain regions in young infants; increases in oxygenated hemoglobin (oxy-Hb) signals are followed by signal decreases to the onset level within a 10- to 20-second time scale. We can also visualize the static aspects of functional relationships, such as correlations between spontaneous activations in homologous regions (Homae et al.,2010). However, we have limited information on the dynamics of cortico-cortical interactions during the presentation of speech sounds to infants. In the present study, we calculated the dynamic functional connectivity (dFC) of cortical activation in neonates and 3- and 6-monthold infants in response to short sentences in Japanese. We hypothesized that functional relationships will not be constant, but will instead be modulated, during typical cortical hemodynamic responses to speech sounds. We further predicted that functional connectivity between distant cortical regions would change from a local to a global state depending on the stage of development. Methods: We analyzed data obtained from quietly sleeping neonates (N = 28) and 3- and 6- month-old infants (N = 26 and 27, respectively). We presented auditory sentences while measuring brain activation using 94-channel fNIRS (ETG-7000, Hitachi). The continuous oxy-Hb signals were band-pass filtered from 0.01 to 0.2 Hz. The signal changes in response to speech sounds were examined by averaging the signals over segmented data blocks. We applied the Hilbert transform to the continuous data to estimate the phase of the signals in all channels. We defined dFC using the following equation (Cabral et al., 2017): Results and Discussion: We found that the frontal and temporal regions in all participant groups demonstrated increases in oxy-Hb signals when the sentences were presented. In addition, 3- and 6-month-old infants exhibited similar changes in the occipital regions, as in our previous studies (Homae et al., 2011; Taga et al., 2018). The dFC between homologous regions increased with age, consistent with our previous findings (Homae et al., 2010). Although frontal regions exhibit dense clusters in all groups, clusters over the temporal, parietal, and occipital regions were observed only in 3- and 6-month-old infants. Overall, we found that the dFC between the frontal and temporal regions changed with the presentation of speech sounds, and the organization of functional networks depended on age. These results support our hypotheses and suggest that the dynamics of functional networks will help in revealing how infants hear speech sounds and acquire their native language. |
新生児と乳幼児における動的機能的結合解析の研究でした．dFCの研究で，SWCではない手法を用いており，dFCの研究として押さえておかなければいけないと感じました．
| 発表タイトル ：Functional Connectivity Patterns in Monolingual and Bilingual Infants 著者 ： B. Blanco, M. Molnar, E. Amico, M. Carreiras and C. Caballero-Gaudes セッション名 ： Neonatal, pediatric & developmental neuroscience I Abstruct ：Introduction. In this work we evaluated whether brain adaptations are induced by the effect of an early and continued exposure to a bilingual vs. a monolingual environment by testing between group differences in resting state functional connectivity (RSFC). We also assessed the reliability of our previous results by testing sleeping, as opposed to awake infants, in order to maximize signal quality. Methods. Spontaneous hemodynamic activity was recorded (9 min.) using nearinfrared spectroscopy (NIRS) in 4-month-old infants (n=27 bilinguals, n=25 Spanish and n=26 Basque monolinguals). Within each group we measured the antiphase relationship between deoxy- (HbR) and oxyhemoglobin (HbO2), and perform a hierarchical spatio-temporal clustering. Network based statistics (NBS) (Zalesky et al., 2010) and connICA (Amico et al., 2017) procedures were also employed to explore differences in functional connectivity patterns between groups. Results and Discussion. In our three experimental groups we observed an antiphase relationship between HbR and HbO2 (Fig. 1A) which resembles the results of Watanabe et al., (2017). The spatial configuration of clusters (Fig. 1B) across groups also demonstrates a high degree of consistency with previously reported results (Homae et al., 2010). These results were not replicated in our previous study in awake infants. We assume that our previously reported effects were probably caused by motion artifacts in the signal, and recognize the importance of correct data quality assessment in NIRS studies with infants to avoid this type of spurious results. Pairwise comparisons with NBS revealed a network (Fig. 1C) involving spatially homologous channels of both hemispheres showing stronger synchronization in Spanish monolingual infants than in Basque monolingual infants (p=0.04). The same difference between groups is also observed in HbO2 (p=0.04) in a network showing a similar spatial disposition. ConnICA revealed a FC pattern (Fig.1C) showing a significantly larger presence in Spanish than in Basque monolingual infants, in HbR (p=0.04) and HbO2 (p=0.04), which resembles the results obtained with NBS. Despite their small effect, the observed between group differences are consistent across HbR and HbO2, and show a similar spatial pattern regardless of the procedure being employed. |
休止状態機能的接続性における群間の差異に関する研究でした．NIRSデータを用いたネットワーク解析で，オキシ・デオキシを用いて解析を行っており，NBS，COnnICAなど新しいことを知ることができました．この研究では，行列が領域で示されていた点，Grobal signal regression，robust independent FC patternなどポイントとなる点は修士論文までにおさえたいです．
| 発表タイトル ：Advanced topics in fNIRS data analysis : Motion Artifact Detection and Correction 著者 ：Meryem A. Yücel セッション名 ： Morning tutorial I Abstruct ： Functional near-infrared spectroscopy (fNIRS) is a relatively new brain imaging technology. fNIRS allows measurements in relatively unrestrained environments. While this allows its adaptation to populations/studies where other modalities are difficult or impossible to implement such as infants, children, speech studies, gait studies, movement disorders or neuro-intensive care, the resultant fNIRS signal often contain motion artifacts. In this lecture we will go over motion artifact detection and correction in detail. Particularly, we will go over several commonly-used motion artifact correction algorithms and will discuss advantages and disadvantages of each comparatively. |
モーションアーチファクトに関する研究でした．PCA,Waveletなど既存の知識の他に，Sokine，CBSIなどの新たな知見を知ることができました．
| 発表タイトル ：NeuroDOT: an extensible Matlab toolbox for streamlined optical brain mapping 著者 ：A.T. Eggebrechta, D. Muccigrosso, J.P. Culver セッション名 ： Data analysis & algorithms Abstruct ： Multiple challenges exist in standardization of data format and processing pipelines for optical neuroimaging. The goal of this project is to provide a fully self-contained and enduser-friendly workflow that seamlessly integrates with a wide variety of other data processing toolboxes and contains dedicated functions and pipelines for preprocessing, anatomical light modeling, reconstruction, post-processing analyses, and visualization. To meet these goals, we present a tool that provides the flexibility to be used by multiple array-based imaging modalities, format compatibility, spatial registration, ease of use, and analytical breadth and sophistication for post-processing. NeuroDOT is written in MATLAB in the style of a conventional MATLAB toolbox. Its functionality is distributed among several pipelines (Fig. 1), with extensive functions for data quality analysis and visualization. To aid in end-user support at multiple levels of familiarity and expertise, beyond the basic functionality, NeuroDOT contains data samples, support files, help sections, appendices, and tutorials. Specifically, a set of anonymized and published data samples have been chosen to reflect common experimental paradigms in neuroimaging (e.g., retinotopy and language based tasks), and are provided in both raw and pre-processed versions to aid in troubleshooting and training for the new user. The extensive support files – all stored as .matfiles – contain geometric information for some of our diffuse optical tomography caps, sensitivity matrices, spectroscopy matrices, and standard atlases. Together with the documentation, these files provide a blueprint for users to create counterparts for their own systems. The NeuroDOT toolbox currently supports a wide variety of standard data file formats (e.g., NIFTI, GIFTI, and others). Help sections exist for each function and are searchable from the MATLAB command line, with a Help Viewer version as well. Both are written and formatted in the style of their native MATLAB counterparts for familiarity and ease of use. Several appendices detail data structures, pipelines and their construction, and select visualizations of our pipelines’ results for multiple data samples. Several tutorials are also included, each of which runs a data sample through a given pipeline to help the user harness the power and flexibility of NeuroDOT. |
NIRSデータを処理するツールボックスの紹介がされていました．HOMER2,AFNI,freesurfer,NueroDOTなど，使用したことのないツールボックスがたくさんあることを知ることができました．NIRSデータの前処理や基本的な考え方をこういったツールボックスでも学べるのではないかと感じました．
参考文献
- fNIRS2018, http://fnirs2018.org/
学会参加報告書
| 報告者氏名 | 池田幸樹 |
| 発表論文タイトル | fNIRSを用いた脳機能ネットワークにおけるワーキングメモリ負荷量依存変化の解析 |
| 発表論文英タイトル | Analysis of working memory-load dependent changes in functional network properties using fNIRS |
| 著者 | 池田幸樹，日和悟，廣安知之 |
| 主催 | 医療情報システム研究室 |
| 講演会名 | Neurophotonics |
| 会場 | Hongo Campus of The University of Tokyo, Tokyo, Japan |
| 開催日程 | 2018/10/05-2018/10/08 |
- 講演会の詳細
2018/10/05から2018/10/08にかけて，東京大学本郷キャンパスにて開催されましたfNIRS2018に参加いたしました．The Society for functional near-infrared spectroscopy （SfNIRS）は，光学的方法を用いて生物組織，特に脳の機能的特性の理解を深める基礎科学者および臨床科学者の組織で， 本学会はアイデアの交換，学際的協力，教育の促進を目的として開催されました．私は全日程参加し，本研究室からは他に廣安先生，日和先生，水野さん，西澤さん，谷口さん，山本さんが参加されました．
- 研究発表
- 発表概要
私は10月08日のポスターセッションに参加いたしました．発表の形式はポスター発表で，貼り出し期間は1日間，セッション時間は午前1時間，午後1時間の計2時間となっておりました．
今回の発表について以下に抄録を記載致します．
| はじめに：ワーキングメモリ（WM）は，情報の同時処理と保存のための認知システムである[1]．本論文では，fNIRSにより測定されたN-backタスク中の脳活動について，グラフ理論解析により機能的コネクレィビティネットワークにおけるWM負荷量依存変化を調べた．方法：N-backタスク（N = 1,2,3）中の6人の健康な被験者（男性6人，23±1.2人）の脳血流変化を，116チャネルfNIRS（ETG-7100，Hitachi，Ltd. ）を用いて計測した． fNIRSの全測定チャネルは，Automated Anatomical Labeling（AAL）に基づく脳領域と関連していた． N-backタスク中の全脳領域間の機能的接続性を示す相関係数行列を計算し，次にエッジ密度を30％に保つように二値化した．グラフ理論特徴量である，degree，clustering coefficient，betweenness centrality，eigenvector centrality，modularity，participation coefficient，within-module degree z-scoreは，Brain Connectivity Toolboxを使用して計算された．各グラフ理論特徴量とWM負荷とのピアソン相関を分析した．結果：左上後頭部回（SOG）のclustering coefficientは，WM負荷と正の相関があった（r = 0.53，p <0.05）．左上前頭回内側部（SFGmed）のparticipation coefficientもまた正の相関があった（r = 0.50，p <0.05）．また，左中心前回（PreCG）のdegree，eigenvector centralityおよびwithin-module degree z-scoreは，WM負荷と負の相関があった（r = -0.59，-0.57，-0.49，p <0.05）．左中心前回周辺のネットワーク構造を図1に示す． 上前頭回内側部と中心前回はWMプロセス[2,3]に関与しているため，これらの領域はWM負荷量変化の特徴量として利用可能である． |
- 質疑応答・意見
今回の講演発表では，以下のような質疑・意見を受けました．
・質疑応答1
Q. 「edge density」とはどういう意味か？
A. エッジの本数を結合可能なエッジの総数で割ったもの，
すべての被験者でエッジの本数が同じになるようにすること
・質疑応答2
Q. 「feature values」はどうやって算出しているのか？
A. MATLABのBrain Connectivity Toolbox（BCT）を用いて算出した．
・質疑応答3
Q. 成績と特徴量の相関はあったのか？
A. なかった
WM-loadは被験者によって異なるため実際の成績では評価できないため
・質疑応答4
Q. なぜ他の特徴量はみていないのか？
A. すべて見て，2-backで違う傾向があり，
その中でも顕著にその傾向が表れていた特徴量2つをピックアップして結果に載せている
・意見1
「quadrant」の意味が伝わらず4つのグループの意味が分かりにくかった
・意見2
N数をもっと増やしたら成績との相関も見られるのでは
・意見3
2群に別れるのは自然なことなので，成績やRTなどと相関があれば面白い
- 感想
本学会での私のポスター発表は，ネットワーク解析自体も少なくグラフ理論に当てはめて解析している人がほとんどいなかったので，そのあたりの質問がとても多かったです．発表前から不安でしたが，4つのグループに分ける話がなかなか理解してもらえず難しかったです．しかし，午前・午後のセッションとも多くの人が発表に興味を持ってくれて常に誰かと話している状況で，有意義な時間になりました．前回の国際学会に参加した際に課題だと感じていた研究への理解度は確実に上がっていたため，自信を持ってポスター発表ができました．英語能力に関しては，質問の意味などは理解できるもののボキャブラリーの少なさもあり思ったように伝わらない場面が多々ありましたが，一生懸命聞いてくださる人が多く頑張って話せたと感じました．
- 聴講
今回の学会では，下記の4件の発表を聴講しました．
| 発表タイトル ： Comparison of source localization techniques in Diffuse Optical Tomography for fNIRS application using a realistic head model 著者 ： J. Tremblaya , E. Martinez-Montesb , P. Vannasinga , D.K. Nguyend , M. Sawane , F. Leporec , M. Lassondec , A. Gallaghera,c セッション名 ： Poster Session Abstruct ： 背景：慢性閉塞性肺疾患（COPD）は死亡原因の第4位であり，2010年には500億米ドルの医療関連負担があった．注意制御，計画，意思決定に関わる前頭前野（PFC）の変化は，COPD患者における認知障害および運動障害が含まれる．fNIRSは，酸素ヘモグロビン（O2Hb）を介して皮質神経活動を定量化することが可能であり，中央処理と運動性能との間のつながりを把握することが可能な新規イメージング技術である． 目的： COPD患者，健常な若年者および高齢者のPFC O2Hbの相対的変化，後方スペリングの正確さ，および単一および二重の作業中の歩行の減少を比較すること． 仮説：H1：PFC O2Hbは，ベースライン課題と比較して後方スペリングの間に増加し，COPD患者および健康な高齢者および若年成人の単一のタスクと比べて二重の仕事の間にさらに増加する． H2：単一のタスクと比較して，3つのグループすべての参加者は二重のタスクの間，後方スペリングの精度が低下し，歩行の減少が大きくなる． 方法：被験者は以下の単一タスクを実行した．（1）認知課題（CT）である後方スペリング（2）30mの好ましい歩行歩行（PPW）（3）30m高速歩行歩行（FPW）．その後，PPWとFPWのそれぞれとCTをペアにした二重のタスクを行った．fNIRSデータをfnirSoftで処理し，生理学的ノイズおよびモーションアーチファクトを減衰させた．次いで，修正されたBeerLambert法を用いてL. DLPFCのO2Hbの変化を計算した．歩数パラメータは5×0.88mの圧力に敏感なゼノマットを用いて測定した． 結果： 20歳の若年成人（男性=10，女性=10，年齢：28±4歳），20人の高齢者（男性=10，女性=10，64±11歳）および6人のCOPD患者（男性=3，女性=3，年齢：74±8歳）が参加した．DLPFCのΔO2Hbは，若年成人のベースライン課題（p = 0.042）と比較してCT期間中に増加し，高齢者およびCOPD患者においてより大きくなる傾向があった．同様に，FPW + CTではFPWと比較して高齢者で有意に高かった（0.27±0.95μM対-0.49±1.01μM，p = 0.007），若年成人およびCOPD患者ではより高い傾向があった．1対2のタスクを比較すると，若い健常者（-16.9±21.4％，p = 0.002）および古い健康グループ（-20.1±22.2％，p = 0.001）において，後方スペリングの精度が低下しCOPD群では（-32.2±45.7％，p = 0.191），FPW + CTではCTと比較して高かった．（1）FPW + CT中の速度が3群すべてにおいてFPWと比較して減少した（p <0.05）．（2）FPW + CT中の高齢者およびCOPD患者においてFPWと比較してスイング時間の変動性が増加した（p = 0.030およびp = 0.038）．（3）FPW患者と比較してCOPD患者におけるFPW + CTの間に立位時間変動が増加した（p = 0.026）． 討論：本結果は，単一のタスクと比較して二重のタスクがPFCのΔO2Hbを増加させ，後方のスペルの正確さを低下させ，より遅くより可変な歩行をもたらすことを示している．より大きなCOPD患者のサンプルは，本データの強さを改善する可能性がある． |
この発表で着目した点は， 図の見せ方です．
発表タイトル ： The Brain at Work: Neural Correlates of Cognitive and Motor Performance著者 ： W. D. Reida , S. A. Hassanb , L. V. Bonettic , K. K. Pattersona , D. S. Beald and A. C. Ruoccoe |
この発表で面白かったところは，被験者が俳優だった点です．NIRSの利点である動作しながら脳活動を計測しておりモーションのデータと脳活動のデータについて考察されていました．課題はN数が少ないことです．
発表タイトル ： Seeing into the brain of an actor with fNIRS and mocap著者 ： A. Hamilton1, P. Pinti1,2, D. Paoletti2, J.A. Ward2 |
この発表で面白かったところは，拡散光トモグラフィ（Diffuse optical tomography，DOT）に着目していた点です．順問題・逆問題を方程式を用いて解いていて，普段あまり聞きなれない話だったので難しかったですが，本ポスター以外にもDOTの発表をしている人が多く見受けられました．
| 発表タイトル ：Comparison of Kernels in Online SVM Classification of fNIRS Data 著者 ： Ruisen Huanga , Ho-Ryong Yoob and Keum-Shik Honga セッション名 ： Poster Session Abstruct ： 目的：SVM（Support Vector Machine）は脳信号分類において広く用いられているが，SVMのための最適化されたカーネルは未知のままである． 方法：8名の健康な被験者から取得した脳信号を調べた．fNIRSを用いてmental arithmetic (AR)課題とmental counting (MC)課題中の前頭前野からの信号を取得した．各チャネルから傾き（最小二乗回帰近似），平均，分散，最大および最小特徴を抽出し，最終的にバイナリおよびマルチクラス分類タスクにSVMを使用した．このプロセスでは線形，ガウス，多項式などのいくつかの異なるカーネルがテストされた． 結果と考察：分類率は，カーネルごとに非常に異なった． ガウシアンカーネルの標準偏差や多項式カーネルの0次項などのパラメータは，分類精度に大きく影響する可能性がある．その結果，これらの調整可能パラメータの最適化された値は被験者固有であることが証明される．これは，同じ値で分類率が85％を超える場合や32％より低くなる場合があるためである． 各カーネルの長所と短所をまとめ議論した． |
この発表で着目したのは， NIRSデータに対してSVMを用いている点です．2つの手法を比較しており，提案手法の結果について熱く語ってくれたのが印象的でした．
学会参加報告書
| 報告者氏名 |
山本渉子 |
| 発表論文タイトル | |
| 発表論文英タイトル | A fNIRS study of attentional state induced by breath-counting meditation |
| 著者 | 山本渉子, 日和悟，廣安知之, |
| 主催 | Society for fNIRS |
| 講演会名 | fNIRS2018 |
| 会場 | 東京大学 |
| 開催日程 | 2018/10/05-2018/10/08 |
- 講演会の詳細
2018/10/05から2018/10/08にかけて，東京大学にて開催されましたfNIRS2018に参加致しました．このfNIRS2018は，Society for fNIRSによって主催された研究会で，Society for fNIRSは，光学的方法を用いて生物，特に脳機能特性を理解しようとする基礎科学者及び臨床科学者の専門組織です．この学会の目的は，アイデアの交換や，学際的協力，教育を促進することです．
私は全ての日程に参加致しました．本研究室からは他に廣安先生，日和先生，M2池田さん，西澤さん，水野さん，M1谷口くんが参加しました．
- 研究発表
- 発表概要
私は8日の10時半から11時半，及び14時から15時のセッション「PosterⅢ」に参加致しました．発表形式はポスター発表で，発表時間は午前と午後各1時間となっておりました．
今回は，A fNIRS study of attentional state induced by breath-counting meditationというタイトルで発表致しました．以下に抄録を記載致します．
| Introduction: Mindfulness, nonjudgmentally paying attention to the present moment, is expected to promote our mental well-being. One of the important components of meditation is attention. In the meditation practice, practitioners try to sustain their attention on their physical sensation so as to improve their interoceptive attention. In this study, we aim to investigate whether meditation induces our interoceptive attention using fNIRS scanning of brain activity during breath-counting meditation. To emphasize the characteristics of the interoceptive state, we also measured the exteroceptive state of practitioners during responding to and counting the external auditory cues. The brain states between two states were compared and discussed using functional connectivity analysis. Methods: Meditation novices (10 males, aged 21.4 ± 1.0 years) who had never experienced meditation practices or retreats participated in the experiment. They were asked to perform two tasks: 1) breath-counting task (BCT) and 2) auditory counting task (ACT). BCT is first proposed by Levinson et al. and shown to be effective as a behavioral measure of mindfulness meditation [1]. In the BCT, participants counted their breaths from 1 to 9 repeatedly, not with voice but just mentally. With breaths 1–8 they pressed one button, and on ninth breath they pressed another. In the ACT, they were asked to quickly press the button responding to the auditory cues and counted them from 1 to 9 as same as the BCT. In both tasks, they restarted counting from one if they got distracted as well as pressing a third button. An fNIRS device (ETG-7100, Hitachi, Ltd.) whose probes were placed on frontal region (47 CH), occipital region (47 CH) and parietal region (22 CH) was used to measure the brain activity during two tasks. The measured data was band-pass filtered with the pass band of 0.008 – 0.09 Hz. Pearson’s correlation coefficient matrix of the cerebral blood flow changes among all the brain regions were calculated for each participant. Each matrix was binarized so as to preserve the edge density of 15%. Graph theoretical metrics, degree centrality and betweenness centrality were calculated for each matrix. Differences in functional connectivity network between two tasks were compared based on these graph metrics. Results: We found that degree centrality of the right dorsal superior frontal gyrus (SFGdor) was significantly higher in the BCT than the ACT, while couldn’t find no significant differences in betweenness centrality. The SFGdor is included in dorsal attention network and involved with top-down attention [2]. We assume that the BCT induced active attention to the participants’ interoceptive sense, while the ACT induced the passive states to the external stimuli. Also, our study has revealed that the difference between the two states was reflected on the nodal degree of the SFGdor. The future study should be performed to further test these assumptions together with the analysis of correlations between neural states and behavioral metrics such as a counting accuracy. [1] D.B. Levinson, E.L. Stoll, S.D. Kindy, H.L. Merry and R.J. Davidson, “A mind you can count on: validating breath counting as a behavioral measure of mindfulness,” Frontiers in psychology, vol. 5, p.1202, 2014. [2] R.N. Spreng, J. Sepulcre, G.R. Turner, W.D. Stevens and D.L. Schacter, “Intrinsic architecture underlying the relations among the default, dorsal attention, and frontoparietal control networks of the human brain,” Journal of cognitive neuroscience, vol. 25, no. 1, pp. 74-86, 2013. |
- 質疑応答
今回の講演発表では，以下のような質疑を受けました．
・質問内容1
質問者の氏名を控え損ねてしまいました．こちらの質問は，2つのタスクの違いは何かというものでした．この質問に対して，ACTは外受容的処理を促すよう注意対象を外部刺激とし，一方でBCTは数息観を行うことにより内発的注意を促していると回答しました．
・質問内容2
質問者の氏名を控え損ねてしまいました．こちらの質問は，Counting accuracyとはなにかというものでした．この質問に対して，実験中どのようにボタンプレスを行っているかを説明し，正しいボタンプレスのセット数を総セット数で割ることで算出していると回答しました．
・質問内容3
質問者の氏名を控え損ねてしまいました．こちらの質問は相関を求めたデータはどのデータか，オキシヘモグロビンのみ解析したのかというものでした．この質問に対して，オキシヘモグロビンの変化量のみ解析したと回答しました．
・質問内容4
質問者の氏名を控え損ねてしまいました．こちらの質問は，CH数はいくつかというものでした．この質問に対して，全脳測定し，116CHであると回答しました．
・質問内容5
質問者の氏名を控え損ねてしまいました．この質問は，測定データに髪の毛の影響はないのかというものでした．この質問に対して，装置によりエラーCHを判定し，そのCHは解析対象から除外しているため問題ないと考えていると回答しました．しかし今後検討する必要があると考えています．
・質問内容6
ヤクルトの研究所の方から質問いただきました．こちらの質問は，このタスクの違いが分かったら何がいいのか，上位目的は何かというものでした．この質問に対して，瞑想中の脳状態を定量化し，フィードバックすることを目的としている回答しました．
・質問内容8
質問者の氏名を控え損ねてしまいました．こちらの質問は，瞑想と言ったら目をつぶっているイメージだがこの実験はどうしているのかというものでした．この質問に対して，閉眼で実験を行っていると回答しました．
・質問内容9
新潟医療福祉大学の小島さんから質問いただきました．こちらの質問は，このタスクは人によって数は違うのかというものでした．この質問に対する回答ですが，両方共のタスクで個人によって回数が異なり，BCTは人によって呼吸のスピードが違い，ACTはランダムな感覚で音を提示しているため異なると回答しました．
・質問内容10
同じく，小島さんから質問いただきました．こちらの質問は，瞑想とはそんなに行われているものなのかというものでした．この質問に対する回答ですが，海外では有名な企業で作業効率を上げるために瞑想する時間を設けるなどと取り入られているが，それに対して日本ではまだそこまで普及していないかもしれないと回答しました．
・質問内容11
質問者の氏名を控え損ねてしまいました．こちらの質問は，ボタンプレスの際，緑のボタンはいつ押すのかというものでした．この質問に対する回答ですが，2つのタスクとも気が逸れたら押すよう指示していると回答しました．しかし，特にACTでは1から9までボタンを押している最中に緑のボタンを押すタイミングが分からないと質問いただき，うまく答えられず廣安先生に助けていただきました．1から9まで数えている間に，音だけに集中するように指示しており，例えばお昼ご飯のことを考えてしまったらその時にボタンをおすように指示している，と回答していただきました．
・質問内容12
三菱ケミカルホールディングスグループの金子さんから質問いただきました．こちらの質問は，瞑想とは無になるものではないのかというものでした．弓道で的を狙うときに無になるようなときと同じような感じかと質問いただきましたが，この質問に対して，マインドフルネスとは「今この瞬間」に意図的な注意を向けることであり，無になることとは異なると回答しました．
・質問内容13
質問者の氏名を控え損ねてしまいました．こちらの質問は，瞑想の研究とはよく行われているものなのかというものでした．この質問に対して，瞑想の研究は行われているが，熟練者における研究に比べて初心者を対象とした研究は少ないと回答しました．
・質問内容14
浜松医科大学の大星先生から質問いただきました．こちらの質問は，数を数えると何か結果が異なるのかというものでした．この質問に対して，2つのタスクどちらとも数は数えていることを説明し，タスクの違いをお伝えしました．そして結果として，賦活では違いが見られなかったが，ネットワークでは有意な差があったこと説明しました．
・質問内容15
質問者の氏名を控え損ねてしまいました．こちらの質問は，どのようにコントラストを立てたのか，2タスクの比較のみを行ったのかというものでした．この質問に対して，ACTのみ，BCTのみ，ACとBCTの比較とすべてのパターンを検討したが，集団解析では賦活領域は見られなかったと回答しました．
- 感想
私たちの研究室としても初めて参加する学会で，どんな学会なのか楽しみにしていました．私としては2度目の国際学会の参加，そして東京で行われたこともあり，前回の国際学会ほど緊張はせず臨むことができました．初日にはeducational courseにも参加させていただき，NIRSの計測や処理方法などを再度確認することができました．2日目からはポスター発表もはじまり，聴講もたくさんさせていただきました．NIRSについての学会のため，今まで知らなかった解析ソフトを知ることができたり，今後自分たちの研究に活かせそうなこともたくさん聞くことができ，とても貴重な経験をさせていただきました．私のポスター発表は学会最終日でしたが，たくさんの方が聞きに来てくださり，質問もたくさんいただきました．前回に比べると緊張もしすぎずに対応できたため，冷静に質問を聞くことができました．しかし，まだ英語力が乏しく議論が深められない場面もありました．また，計画性が足りず準備もかなりぎりぎりとなってしまったところも反省点です．しかし，今回の学会を通して直前までやり続ける粘り強さをさらに身につけ，英語力向上，そして自分の研究に対するモチベーションも上げることができました．今後活かすことができるよう，英語力も高め，自分の研究もさらに深めながら進めていきたいと強く思います．
- 聴講
今回の講演会では，下記の6件の発表を聴講しました．
| 発表タイトル ： fNIRS Technology and Applications 著者 ： Robert J Cooper セッション名 ： fNIRS Training Course，Principles & Methodology In this session, I will outline the past, present and future of fNIRS technologies, with a specific emphasis on what are and what are not appropriate experimental applications of fNIRS methodologies. I will also review a number of key papers that have employed fNIRS technology over the last 25 years. |
この発表は，今まで，そしてこれからのfNIRS技術の説明でした．まずNIRS技術には主なカテゴリーが4つあるというところから始まりましたが，そこからよく知らなかったので勉強になりました．また，普段あまり知ることのないレーザーやファイバなどのハードウェアの話も聞くことができました．さらに，今まで25年間の重要な論文のレビューも多く紹介してくださったので，参考にできるものや興味があるものを自分でも読んでみようと思いました．
| 発表タイトル ： Guide to the probabilistic spatial registration methods for fNIRS : Brain atlas handling 著者 ： Daiske Tsuzuki セッション名 ： fNIRS Training Course，Measurement & Analysis Functional near-infrared spectroscopy (fNIRS) measures regional cerebral blood flow and monitors relative regional changes of hemoglobin concentration noninvasively from the head surface. On the other hand, fNIRS in a standalone setting lacks structural and anatomical information of the underlying brain. In this session, I will introduce both the concept of atlas model based on the standardized brain space such as Montreal Neurogical Institute (MNI) space, and methods to solve the spatial registration issue and realize the functional mapping at the gyrus level through the technical descriptions of Virtual spatial registration, Probabilistic registration, Anchor-based probabilistic registration, and Anatomical labeling. |
この発表では，レジストレーションの方法に関して教えていただきました．私たちの研究室では，NIRSのCH位置をMNI標準脳座標系にレジストレーションする確率的レジストレーションを用いていますが，その他にも仮想レジストレーションという方法が紹介されていました．今まで聞いたことはありましたが，あまり知らなかったのですごく興味を持ちました．詳細はあまり分からなかったので，参考文献なども読んでもっと詳しく知りたいと思いました．
| 発表タイトル ：Functional Connectivity Patterns in Monolingual and Bilingual Infants 著者 ： B. Blanco, M. Molnar, E. Amico, M. Carreiras and C. Caballero-Gaudes セッション名 ： Neonatal, pediatric & developmental neuroscience I Introduction. In this work we evaluated whether brain adaptations are induced by the effect of an early and continued exposure to a bilingual vs. a monolingual environment by testing between group differences in resting state functional connectivity (RSFC). We also assessed the reliability of our previous results by testing sleeping, as opposed to awake infants, in order to maximize signal quality. Methods. Spontaneous hemodynamic activity was recorded (9 min.) using nearinfrared spectroscopy (NIRS) in 4-month-old infants (n=27 bilinguals, n=25 Spanish and n=26 Basque monolinguals). Within each group we measured the antiphase relationship between deoxy- (HbR) and oxyhemoglobin (HbO2), and perform a hierarchical spatio-temporal clustering. Network based statistics (NBS) (Zalesky et al., 2010) and connICA (Amico et al., 2017) procedures were also employed to explore differences in functional connectivity patterns between groups. Results and Discussion. In our three experimental groups we observed an antiphase relationship between HbR and HbO2 (Fig. 1A) which resembles the results of Watanabe et al., (2017). The spatial configuration of clusters (Fig. 1B) across groups also demonstrates a high degree of consistency with previously reported results (Homae et al., 2010). These results were not replicated in our previous study in awake infants. We assume that our previously reported effects were probably caused by motion artifacts in the signal, and recognize the importance of correct data quality assessment in NIRS studies with infants to avoid this type of spurious results. Pairwise comparisons with NBS revealed a network (Fig. 1C) involving spatially homologous channels of both hemispheres showing stronger synchronization in Spanish monolingual infants than in Basque monolingual infants (p=0.04). The same difference between groups is also observed in HbO2 (p=0.04) in a network showing a similar spatial disposition. ConnICA revealed a FC pattern (Fig. 1C) showing a significantly larger presence in Spanish than in Basque monolingual infants, in HbR (p=0.04) and HbO2 (p=0.04), which resembles the results obtained with NBS. Despite their small effect, the observed between group differences are consistent across HbR and HbO2, and show a similar spatial pattern regardless of the procedure being employed. Watanabe H. et al. (2017). Hemoglobin phase of oxygenation and deoxygenation in early brain development measured using fNIRS. Proc. Natl. Acad. Sci. 114(9), E1737–E1744. Homae F. et al. (2010). Development of global cortical networks in early infancy. J. Neurosci. 30(14), 4877– 4882. Zalesky A. et al. (2010). Network-based statistic: identifying differences in brain networks. Neuroimage, 53(4), 1197-1207. Amico E. et al. (2017). Mapping the functional connectome traits of levels of consciousness. NeuroImage, 148, 201-211. |
この研究は，単一言語と二言語の幼児における機能的コネクティビティのパターンについての内容でした．コネクティビティのパターンを調べる際に，オキシヘモグロビンだけではなく，デオキシヘモグロビンも用いて解析されていました．今までデオキシヘモグロビンのデータについては用いていなかったので，今後他の論文でも調査していきたいと思いました．また，オキシヘモグロビンとデオキシヘモグロビンの位相差についても述べられていましたが，はじめて見て理解しきれなかったので学んでいきたいと思います．
| 発表タイトル ： Dynamics of Functional Networks in the Developing Brain 著者 ： Fumitaka Homae セッション名 ： Neonatal, pediatric & developmental neuroscience I [invited talk] Introduction: Structural and functional organization of the brain occurs rapidly in early infancy. Previous studies report that in neonates and young infants, this neural organization is sensitive to speech sounds, and related functions in distinct brain regions have been partially elucidated (Homae et al., 2014). It is known that short speech sounds induce eventrelated activation in the temporal brain regions in young infants; increases in oxygenated hemoglobin (oxy-Hb) signals are followed by signal decreases to the onset level within a 10- to 20-second time scale. We can also visualize the static aspects of functional relationships, such as correlations between spontaneous activations in homologous regions (Homae et al., 2010). However, we have limited information on the dynamics of cortico-cortical interactions during the presentation of speech sounds to infants. In the present study, we calculated the dynamic functional connectivity (dFC) of cortical activation in neonates and 3- and 6-monthold infants in response to short sentences in Japanese. We hypothesized that functional relationships will not be constant, but will instead be modulated, during typical cortical hemodynamic responses to speech sounds. We further predicted that functional connectivity between distant cortical regions would change from a local to a global state depending on the stage of development. Methods: We analyzed data obtained from quietly sleeping neonates (N = 28) and 3- and 6- month-old infants (N = 26 and 27, respectively). We presented auditory sentences while measuring brain activation using 94-channel fNIRS (ETG-7000, Hitachi). The continuous oxy-Hb signals were band-pass filtered from 0.01 to 0.2 Hz. The signal changes in response to speech sounds were examined by averaging the signals over segmented data blocks. We applied the Hilbert transform to the continuous data to estimate the phase of the signals in all channels. We defined dFC using the following equation (Cabral et al., 2017): dFC(p, q, t) = cos(θ(p,t) – θ(q,t)), where p and q are channels and t is a time point. Results and Discussion: We found that the frontal and temporal regions in all participant groups demonstrated increases in oxy-Hb signals when the sentences were presented. In addition, 3- and 6-month-old infants exhibited similar changes in the occipital regions, as in our previous studies (Homae et al., 2011; Taga et al., 2018). The dFC between homologous regions increased with age, consistent with our previous findings (Homae et al., 2010). Although frontal regions exhibit dense clusters in all groups, clusters over the temporal, parietal, and occipital regions were observed only in 3- and 6-month-old infants. Overall, we found that the dFC between the frontal and temporal regions changed with the presentation of speech sounds, and the organization of functional networks depended on age. These results support our hypotheses and suggest that the dynamics of functional networks will help in revealing how infants hear speech sounds and acquire their native language. Acknowledgements: This study was partly supported by Japan Society for Promotion of Science KAKENHI Grant Nos. JP16H06524, JP16H06525 and 26220004. |
NIRSで早い時期から結合解析をされている保前先生の，動的機能的ネットワークについての発表でした．結果として，発話音提示に伴う機能的ネットワークの構成は年齢に依存するとのことで，非常に興味深かったです．今回，3ヶ月と6ヶ月の乳児を対象として実験を行っていましたが，たった3ヶ月で変化が起こることに驚きました．また，研究室でも用いられている動的解析をやられていたので動的解析のメリットも知っていきたいと思いました．
| 発表タイトル ： NIRSTORM, a Brainstorm plugin inspired by electrophysiology dedicated to fNIRS data analysis, advanced 3D reconstructions and optimal probe design 著者 ： T. Vincent, Z. Cai, F. Tadel, A. Spilkin, A. Machado, S. Baillet, L. Bherer, J.M. Lina, C. Grova セッション名 ： Data analysis & algorithms Abstract: Measurements of bioelectrical activity using EEG or hemodynamic processes using fNIRS enable wearable neuroimaging in realistic lifestyle and clinical conditions. Despite measuring different physiological signals, EEG and fNIRS are sharing several similarities: (i) they consist in scalp measurements, (ii) they offer an excellent temporal resolution, (iii) their spatial resolution is limited and 3D reconstruction of the generators of these scalp recordings requires solving an ill-posed inverse problem. Brainstorm software is internationally recognized for EEG/MEG processing [Tadel et al. Comp Intell Neurosci 2011], featuring advanced databasing, visualization, signal processing, source localization and statistical analysis methods. Benefiting from powerful ergonomic features available in Brainstorm, we are proposing NIRSTORM, a plugin dedicated to fNIRS data analysis. Methods: NIRSTORM provides classical channel-space fNIRS processing consisting in band pass filtering, Modified Beer-Lambert Law, motion correction and window averaging. To allow advanced fNIRS modeling, NIRSTORM is now featuring integration with MCXLab software [Fang and Boas Opt. Express 2009] to estimate light sensitivity profiles within head tissue, using head models derived either from a standard template MRI (Colin 27) or a subject-specific MRI. Segmentation of the template was carefully computed and accurate Monte Carlo simulations were launched to generate fluence data for 690 and 830 nm wavelengths. This fluence template consists in ~6000 head vertices with a spatial resolution of ~5mm (Fig.A). Using light sensitivity profiles estimated from any vertex of the scalp, we implemented our proposed personalized optimal montage design targeting a predefined brain region. This method consists in linear programming optimization maximizing light sensitivity to the target region, while ensuring spatial overlap between sensors to allow local 3D reconstruction [Machado et al JBO 2014, JNS-Meth. in rev, Pellegrino et al Front. Neurosc. 2016] (Fig.B: optimal montages targeting the hand knob). The optimal montage can be estimated from the MRI template or from an individual MRI, using any freely defined optode position available on the scalp or positions from the 10/20 system. Integration within Brainstorm allowed adapting to NIRS advanced 3D reconstruction methods we developed and validated for EEG/MEG source imaging [Pellegrino et al HBM 2018] within the Maximum Entropy on the Mean framework [Cai et al submitted] (Fig.C: 3D reconstruction for a visual task, left checkerboard reversal). Conclusion: NIRSTORM is an open-source initiative and welcomes any contribution. It is currently hosted on github(https://github.com/Nirstorm/nirstorm), where the wiki pages of our first training session organized in Montreal in May 2018 are available. We expect NIRSTORM to become an ideal platform for multimodal EEG/fNIRS integration. |
この発表は，NIRSTORMというNIRSやEEGのデータ処理を行うことができるソフトウェアについてでした．このソフトはBrainstormソフトウェアから開発されたものであり，結合解析も可能とのことでした．学会の直前にはじめて知り，まだほとんど知らなかったため興味を持ちました．便利なソフトだと思うので，どんな処理ができるのかを知るとともに触れてみたいと思いました．
| 発表タイトル ： The Role of Mirror Neuron System in Encoding Motor Complexity 著者 ： Xinge Li, Manon A. Krol, Sahar Jahani, David A. Boas, Helen Tager-Flusberg and Meryem A. Yücel セッション名 ： Cognitive & social neuroscience Background: The mirror neuron system (MNS), including pars opercularis of the inferior frontal gyrus (IFG) and adjacent ventral premotor cortex (PMv) and inferior parietal lobule (IPL), is activated when individuals execute an action and when they observe a similar action performed by another individual¹. Converging neuroimaging evidence has shown the functional role of the MNS in action understanding. Although most studies have focused on the effects of modulations in goals and kinematics of observed actions on MNS activity, little research has explored the effects of manipulations in motor complexity. Thus, the present study adopted fNIRS to examine MNS activity, aiming to better understand the functional role of this system in encoding motor complexity. We hypothesized that, compared with observed and executed simple actions, MNS will activate more during observed and executed complex actions in order to reinforce motor planning and precision. Methods: Twenty-one healthy adults executed as well as observed two hand actions that differed in motor complexity. The simple action was to place a card into an open box, while the complex action was to insert a card into a box with a narrow slot on the lid (Fig.1 c-f). MNS activity was recorded by an fNIRS system with 16 sources and 24 long-separation and 8 short-separation detectors. Hemodynamic response function was estimated by a general linear model with short separation regression. Results: Our results show that the observation complex and execution complex tasks both resulted in stronger brain response in the ipsilateral mirror neuron regions compared to contralateral. Motor complexity, represented as the contrast between the simple and complex task conditions, on the other hand, is represented in IFG, IPL and motor regions during execution and in IPL and motor regions during observation (Fig.2). Discussion: Our findings suggest that the MNS encodes motor representation of varying motor complexity in the observed and executed actions and facilitates action understanding in a direct, pre-cognitive and motor-based way. References 1. Rizzolatti, G., & Sinigaglia, C. (2016). The mirror mechanism: A basic principle of brain function. Nature Reviews Neuroscience, 17(12), 757–765. |
この発表では，ミラーニューロンシステムの役割について調べていました．そのために，複雑さが異なる2つの手の動きをタスクとして用いていましたが，手の動きをすることによって頭も動いてしまい，体動が乗っている可能性があると思い気になりました．プローブ配置の図など，発表する見せ方において参考にできる点もあったので活かしていきたいと思います．
参考文献
- fNIRS2018， http://fnirs2018.org/
学会参加報告書
| 報告者氏名 |
谷口 尚 |
| 発表論文タイトル | 協調課題時の脳間神経同期のfNIRSを用いたhyperscanningの研究 |
| 発表論文英タイトル | A fNIRS-based hyperscanning study of inter-brain neural synchronization during a cooperative task |
| 著者 | Megumi Mizuno, Sho Taniguchi, Satoru Hiwa, Tomoyuki Hiroyasu |
| 主催 | The Society for fNIRS |
| 講演会名 | fNIRS2018 |
| 会場 | 東京大学 本郷キャンパス |
| 開催日程 | 2018/10/05-2018/10/08 |
- 講演会の詳細
2018/10/05から2018/10/08にかけて，東京大学本郷キャンパスにて開催されましたfNIRS2018に参加いたしました．この学会は，The Society for fNIRSによって主催され，光学的方法を用いて生物組織，特に脳の機能特性を理解しようとする基礎科学者および臨床科学者の集まりで，アイデアの交換，学際的協力，教育を促進することを目的に開催されています．
私は2018/10/05から2018/10/08の全日参加いたしました．本研究室からは他に廣安先生，日和先生，池田，西澤，水野，山本が参加しました．
- 研究発表
- 発表概要
私は10/07のPoster Session「PosterⅡ」にて水野と共に発表いたしました．発表の形式はポスター発表で，午前，午後共に１時間ずつ自由に発表と質疑応答を行う時間となっておりました．
今回の発表は，「A fNIRS-based hyperscanning study of inter-brain neural synchronization during a cooperative task」です．以下に抄録を記載致します．
| Introduction: Social interaction is a dynamic behavior between individuals who modify their actions and reactions depending on the actions of their partner. In this study, we investigate the relationship between social interaction and brain functions. Cui et al. analyzed the neural synchronization between two subjects who played a cooperation game involving synchronizing each other’s response timing and revealed that the interpersonal brain coherence of the subjects increased during the cooperative task [1]. In this study, to easily facilitate the cooperative behavior of the participants, we improved the experimental design and examined the inter-brain neural synchronization during the cooperative task. Methods: Twenty-two healthy adult males (11 pairs, age: 22.7 ± 1.0 years old, right-handed) participated in this experiment. The experimental environment is shown in Fig 1. The brains of the two participants during their social interaction were simultaneously measured by hyperscanning using a single functional near-infrared spectroscopy (fNIRS) device (ETG-7100, Hitachi, Ltd.). A 3 × 10 probe consisting of 47 measurement channels was attached to the forehead of each subject. Participants were instructed: (1) to synchronize the response timing of their partner after the cue (synchronization task), (2) to respond faster than their partner (competition task), and (3) to respond quickly to the cue (single A/B task) [1]. In the synchronization task, the time difference between the responses of the two participants were fed back to each of them. Each participant predicted his partner’s behavior based on the time difference, and was asked to synchronize his responses at the next cue. Wavelet transform coherence was calculated from two sets of time-series data of cerebral blood flow changes. Then we performed a one-sample t-test (p < 0.05) of the coherence increase of each task and investigated the brain regions where the coherence increased. Results & Discussion: The difference in response time between the two participants in the synchronization task was significantly larger than that of the single A and B task (p < 0.05). This result indicates that the participants not only responded quickly to the cue, but also reacted by anticipating the behavior of their opponents. A t-score map of coherence increase is shown in Fig 2. The coherence within the left superior frontal gyrus (SFG) and left middle frontal gyrus (MFG) increased significantly in the synchronization task. The left SFG is involved in building a trust relationship [2]. Hence, it is assumed that this region affects cooperative behavior. Furthermore, the coherence in these two regions did not increase significantly in either the competitive or single task. Therefore, we suggest that the inter-brain synchronization in these brain regions can be utilized as a metric of cooperativeness. |
- 質疑応答
今回の講演発表では，以下のような質疑を受けました．
・質問内容1
質問者の氏名を控え損ねてしまいました．
Single課題の結果に対してどのように考えているかというものでした．同じ状態であると考えていたSingleA課題とSingleB課題のコヒーレンス増加量のt-mapが異なっており，これは各被験者を対象にしてとったEQSのアンケート結果から，個人の性質などの違いが影響を与えたと考えているとお答えしました．
・質問内容2
質問者の氏名を控え損ねてしまいました．
Synchronization課題の影響とみなしたものがその課題による影響だとは分からないのではないかというものでした．この質問に対しては同じ相手を必要とするCompetition課題と比較をし，Competition課題に現れていないものがSynchronization課題の影響であるとみなしたとお答えしました．
・質問内容3
質問者の氏名を控え損ねてしまいました．
なぜ前頭部を測定したのかというものでした．この質問に対しては，まずは参考にした先行研究で測定していた前頭部を中心により多くの領域を計測したと回答しました．現在は側頭頭頂接合部などを測定している社会性のhyperscanning研究も多々あるため，今後はさらに文献を調査して関心領域について決めていく必要があると考えています．
・質問内容4
質問者の氏名を控え損ねてしまいました．
Wavelet変換によるコヒーレンスと移動窓を用いたFourier変換によるコヒーレンスを求めたのはその違いを見たかったためかというものでした．その違いの検討にはまだ至っておらず，まずは両方のコヒーレンスから得られる結果に関して総合的に考察を行ったと回答しました．
・質問内容5
質問者の氏名を控え損ねてしまいました．
違う課題とのコヒーレンスを算出し比較することで結果の信頼性が増すのではないかとご意見をいただきました．現在は同じ課題の同じCHで評価を行っていますが，ご指摘いただいたように課題の影響がコヒーレンスに現われたことを裏付けるためにもそのようなことを今後検討していく必要があると思いました．
- 感想
今回は２度目の学会参加ではありましたが初の国際学会であり，ポスター発表でした．最初は英語ということもあり緊張してなかなか思うように自分の伝えたいことを伝えることができませんでした．しかし，慣れていくにつれ，自分から積極的に声をかけ議論を行うことができ，楽しめました．同じ分野の方と話す機会が研究室外ではあまりなかったため，発表や聴講を通して新たな気づきや学びがあるとともに，自分の研究をより客観的に見つめ直すことができました．しっかり振り返りを行い，今回の学会参加を通して得られたものを自分のみならず研究室全体に還元していきたいです．
- 聴講
今回の講演会では，下記の４件の発表を聴講しました．
| 発表タイトル ： Verifying wavelet coherence analysis to measure neural coupling using pseudo-random visual stimulation sequences and fNIRS 著者 ： Xian Zhang, J. Adam Noah, Swethasri Dravida, Joy Hirsch セッション名 ： PosterⅠ Abstruct ： Neural mechanisms that underlie dynamic interpersonal interactions are thought to include temporally synchronous signals that reflect coupled processes between two brains [1]. Wavelet coherence analysis of hemodynamic signals acquired simultaneously during hyper-scanning experiments has been proposed for analysis of these neural processes [2,3]. However, this computational approach has not been validated against known markers. Here we generate a set of known neural coherences using pseudo random sequences of reversing checkerboard patterns. It is expected that input sequences that were more highly correlated generated higher cross-brain output correlations of wavelets in the visual cortex across pairs of subjects. Each visual stimulus event was a 2-second full-field reversing checkerboard pattern. Three random sequences, called sequence A, B and C (inset in figure), of such visual stimuli were presented for 2 minutes and repeated twice. The sequences were generated to have varying levels of correlation (e g: A-B more correlated than A-C). The expected coherence (left panel) was obtained by coherence analysis on the modeled fNIRS waveform, which was generated by convolving the stimulus sequence and the hemodynamic response function (HRF).The overall expected coherence is summarized as the average coherence within 0-30 second wavelength range (horizontal lines and numbers in left panel). Ten subjects participated. Subject’s data were randomly paired with every other subject’s data, resulting in ninety possible pairs. The combinations included A-A, B-B, or C-C, where both subjects viewed identical sequences of the stimuli (green). The other pairings were A- B (red), A-C (blue) and B-C (cyan). The wavelet coherence results from measured fNIRS deOxyHB data (right panel) confirmed the expected coherence. Our results validate wavelet coherence as a technique for quanitfying the coherence of brain signals across participants during hyperscanning to study social interaction in ecologically valid paradigms. |
この発表ではhyperscanning研究でよく使われているwavelet coherence解析の検証を行っていました．空間フィルタリングされたデータの方が生データよりも本来の予測されたコヒーレンスの値をよりよく表しているとのことでした．今後の解析方法などに参考にしていきたいと思いました．
| 発表タイトル ：Transition of phase synchrony of fNIRS signals depending on sleep states in infants 著者 ： G. Taga, H. Watanabe, R. Saji, F. Homae セッション名 ： PosterⅢ Abstruct ： Sleep reflects spontaneous activity of the brain and includes sleep state transitions such as NREM sleep and REM sleep. The gold standard for classification of sleep state is to use electroencephalogram (EEG) and electrooculogram (EOG). On the other hand, fNIRS can measure brain tissue oxygenation dynamics related to spontaneous activity of the brain and is expected to provide complementary information to the electric method with respect to the sleep state. Especially, studying infants’ sleep is important for elucidating the developmental mechanism of the brain. fNIRS has been used to study spontaneous activity (Taga et al. 2000), stimulus induced activity (Taga et al. 2018), functional network (Homae et al. 2010), and hemoglobin phase of oxygenation and deoxygenation (hPod) (Watanabe et al. 2017) in sleeping infants. The present study aims to clarify the spatiotemporal dynamics of spontaneous activity of the brain according to sleep state of infants using fNIRS. We measured 2- and 3-month-old infants. 94 channel probe of the NIRS device (ETG 7000, Hitachi) and the electrodes of electroencephalogram (EEG) and electrooculogram (EOG) were attached to infants while they were naturally sleeping in the laboratory in the daytime. Measurement was terminated when the infant woke up or when 25 minutes passed while sleeping. Sleep states were classified into active sleep (AS), quiet sleep (QS), AS/QS and wake every 30 seconds using EEG, EOG and video recordings. Data on infants whose sleeping states showed transition from AS to QS, each of which lasted more than 10 minutes, were selected, and 16 2-month-old (2M) and 19 3-month-old (3M) infants were used for the further analysis. Time-series data of oxygenated hemoglobin were bandpass-filtered (0.05 to 0.1 Hz) and subjected to Hilbert transformation to obtain instantaneous phases. Furthermore, the vector sum of instantaneous phases of 94 channels were taken at each time, and then time series of phase synchronization index (PSI) was obtained (Taga et al. 2011). PSIs were time-averaged within the sleep state and compared between the states. In the 2M group, PSI = 0.70 for AS and PSI = 0.48 for QS. In the 3M group, PSI = 0.73 for AS and PSI = 0.46 for QS. Analysis of variance of age x sleep state showed a significant main effect on sleep state (p <0.001). This result implies that when the sleep changes from AS to QS, the synchrony decreases from spatially synchronous state to asynchronous state. There was no significant difference in age. It has been assumed that AS and QS in infants correspond to REM and NREM sleep in adults, respectively. REM and NREM sleep have been thought to reflect synchronous and asynchronous neural activity of the cerebral cortex, respectively. While the spatiotemporal scale of the signals is different between EEG and fNIRS, the present study showed that the sleep state dependence on synchronization property of fNIRS signals is opposite to the one of EEG. This study suggests that the functional connectivity among the global network is enhanced in AS and diminished in QS. In future, further investigations are needed to understand functional roles of cortical activity in different sleep states in infants. |
この発表ではfNIRSを用いて幼児の睡眠状態に応じた脳の自発活動の時空間ダイナミクスを明らかにすることを目的としていました．EEGを用いるものは知っていましたがfNIRSを用いた睡眠の研究はあまり聞いたことがなく新鮮でした．また解析に用いていた瞬時位相を自分の研究でも利用できないかと考えていたためとても参考になりました．
| 発表タイトル ：Inter-brain synchronization during a freestyle PC game: an fNIRS hyperscanning study 著者 ： M. Xu, S. Morimoto, E. Hoshino, セッション名 ： PosterⅢ Abstruct ： Motivation: Human social interaction is a highly dynamic process consisting of continuous live social signal exchanges. Thus, studying interaction between multiple humans is critical to understanding the brain basis of social behavior, which is an emerging field of interest. fNIRS hyperscanning has played a significant role in advancing this field by enabling real-world neuroimaging. However, previous studies mainly adopted tasks during which multiple participants had performed synchronized actions. It remains largely unclear about the underlying mechanisms and any influential factors of inter-brain couplings during real-world interactions when participants have a common goal but much flexibility in action. Method: Thirty-nine same-sex dyads played a turn-based PC game (Fig.1) cooperatively or independently while their hemodynamic responses in the right frontal and temporal regions were measured simultaneously. At the same time, their physiological responses (heart rate and skin conductance response) were recorded simultaneously as well. Any type of verbal/nonverbal interaction was allowed during the cooperative sessions (the dyads were prompted to design the interiors of a big room that satisfy both themselves and their partners) but not during the independent sessions (to design the interiors of two small rooms without considering their partners). Their behaviors during the tasks were videotaped. The inter-brain synchronization (IBS) was evaluated using wavelet transform coherence (WTC) (1). Results and discussion: The IBS was found to display different patterns during the cooperative sessions and the independent sessions, and the IBS was more prominent during cooperation. Moreover, the period during which the IBS occurred was shown to correlate with the dyads’ span of turn-takings. These findings indicate that cooperative behaviors facilitate between-brain neural couplings. The relationship between the IBS and other types of interactive signals (e.g., eye contact, verbal communication, etc.) and physiological synchronization will be further examined. |
部屋のデザインを行うゲームを利用したオリジナルな実験設計でした．脳だけでなく他の計測機器を用いて心電図や皮膚コンダクタンス反応などの生理学的データを取得していました．今後は，それらに加えて，撮影した動画などからアイコンタクトや言葉によるコミュニケーションなどの行動データと脳活動の関係もみていきたいとのことでした．他の生理学的データや行動データを豊富に取っており，脳活動との関連が分かれば面白いと思いました．
| 発表タイトル ： THE NEW NEUROSCIENCE OF TWO: HYPERSCANNING WITH fNIRS TO UNDERSTAND COMMUNICATING BRAINS 著者 ： Joy Hirsch セッション名 ： Cognitive & social neuroscience Abstruct ： Humans are a profoundly social species. However, little is known about the most fundamental brain functions that mediate live social interactions. This knowledge gap between single and social brain processes, is, in part, a consequence of conventional neural imaging methods that are generally restricted to single individuals, static tasks, and nonverbal responses. However, recent developments of fNIRS hyperscanning (imaging two individuals simultaneously) enables rigorous investigation of this unexplored neural domain of human social behavior. Two existing theoretical frameworks converge as a foundation for this new “neuroscience of two”. The first is the Interactive Brain Hypothesis (De Jaegher, et al, 2016) which proposes that social interactions are mediated by dedicated brain processes. This hypothesis provides a general framework for the investigation of neural mechanisms underlying social interaction. The second is a quantifiable model of dynamic neural coupling, i.e. the correlation between temporal patterns of neural signals across two interacting partners (Hasson and Frith, 2016). This model proposes that dynamic coupling is the mechanism by which information is shared between the sender and the receiver across brains. Hyperscanning with fNIRS, together with these emerging theoretical frameworks, establishes fNIRS as the neuroimaging method of choice for investigations of live human-to-human interactions. Paradigms for investigation of two individuals engaged in live and spontaneous communications include developments for imaging in natural environments as well as the computational methods necessary to quantify live cross-brain interactions. We have applied these tools to test specific cases of the interactive brain hypothesis using fNIRS hyperscanning and primary social cues such as real eye-to-eye contact compared to eye gaze at a picture-face (Hirsch, et al., 2017), a video-face (Noah, et al.,2018), and talking and listening (Hirsch, et al., 2018) with and without social interaction. Consistent with the Interactive Brain Hypothesis cross-brain coherence measured by Wavelet Coherence Analysis (MATLAB Wavelet Toolbox) between local brain areas is greater for interactive than non-interactive conditions. Specifically, dynamic neural coupling increases during interactions such as eye-to-eye contact and interactive speaking and listening. Further, the dynamic neural coupling is associated with neural activity in temporal-parietal regions of the brain consistent with specialized cross-brain neural mechanisms for live social interaction regardless of the modality or task. These, and other similar findings recently reported using fNIRS hyperscanning, contribute to an expanding experimental and theoretical foundation for a new neuroscience where the aim is to understand the dynamic signal exchanges and neural mechanisms within a dyad that underlie live episodes of social interaction in natural conditions. |
導入においてhyperscanningの研究について丁寧に説明をしており，hyperscanning研究の過去・現在・未来についてよく分かり，話の構成や仮説の立て方など大変勉強になりました．また，fNIRSを用いたhyperscanning研究が脳の社会性を理解するための主流になるだろうとおっしゃっており，自分の研究の意義を再確認することができました．
参考文献
- fNIRS2018, http://fnirs2018.org