Estimating waveforms common across trials from MEG data in various stimulus–response tasks

s / Neuroscience Research 71S (2011) e108–e415 e209 P2-v07 Facilitation of OKR adaptation by GABAb receptor activation in mice: An analysis using a machine visionengined measurement system Toshihide Tabata , Yoshihiro Shirai, Shu Uchiyama, Toshifumi Sasajima, Masahiro Takegoshi, Kimihiro Tachikawa, Kenkichi Fukurotani Lab. for Neural Info. Tech, Grad. Sch. of Scis. & Eng., Univ. of Toyama, Toyama, Japan Optokinetic reflex (OKR) is an automatic eye movement to follow moving objects. The spatiotemporal fidelity of OKR to the stimulus movement can be measured without ad-hoc training and powerfully quantifies the functional status of various brain regions including the cerebellovestibular and oculomotor systems. We developed an easy-to-install/use yet reliable system for measuring OKR in mice and studied how activation of cerebellar B-type -amino butyric acid receptor (GABAbR) influences OKR adaptation. The eye tracker of our system consisted of industrial-use video equipment governed by home-made software created with a machine vision application development platform. Non-engineer/computer expert researchers may readily program and operate the eye tracker through the graphical interface of the platform. The eye tracker located the pupil center on line at a high success rate and based on its outputs, the pupil azimuth was estimated accurately with a high linearity to the real value. We also devised a stereotaxic apparatus which held the mouse at the multiple points and thereby minimized artifacts due to voluntary body movements. We injected saline with or without baclofen, a GABAbR-selective agonist into the cerebellar flocculi of a mouse and then exposed the mouse to horizontally oscillating black and white stripes for 1 h/day. On the first day, OKR gain increased to similar extents under both conditions. On the second day, OKR gain was maintained in the baclofen-injected mice but not in the vehicle-injected mice. Such a facilitatory effect was observed with a wide range of the dose of baclofen (50 pM–50 M, 500 nl/side). Activation of cerebellar GABAbR may facilitate consolidation of the long-term memory of OKR adaptation. Research fund: KAKENHI 21026011, 20500284, 20022025, and 19045019; grants from Presidential Office, Cooperative Res. Ctr., and Collabofesta 2010, Univ. of Toyama. doi:10.1016/j.neures.2011.07.905 P2-v08 Three-dimensional proliferation culture of differentiated retinal horizontal interneurons in a porous sponge Itsuki Ajioka 1 , Shizuko Ichinose 2 1 Center for Brain Integration Research, Tokyo Medical Dental U, Tokyo, Japan 2 Research Center for Medical and Dental Sciences, Tokyo Medical Dental U, Tokyo, Japan As the neurons in the central nervous system (CNS) form a neuronal network in a three-dimensional (3D) manner, a 3D proliferation culture system for differentiated neurons has been desired. Although differentiated neurons were previously thought to never proliferate, differentiated horizontal interneurons of Rb−/−; p107+/−; p130−/− (p107-single) retina clonally proliferated without dedifferentiation in vivo. In this study, we developed a basement membrane-like matrix sponge (BM-sponge) for the 3D proliferation culture of differentiated horizontal interneurons. p107-single horizontal interneurons, but not other types of retinal neurons, proliferated in the BM-sponge in a 3D manner. These interneurons expressed presynaptic marker and developed synaptic vesicles. These data demonstrated that p107-single horizontal interneurons cultured in the BM-sponge proliferate while maintaining their differentiated features. This is the first report of the 3D proliferation culture system for differentiated neurons. Research fund: Industrial Technological Research Grant Program in 2009 from NEDO. doi:10.1016/j.neures.2011.07.906 P2-v09 Isolating cortical activities from artifacts in simulated EEG data during smooth pursuit eye movements Ken-ichi Morishige 1,2 , Takatsugu Aihara 2, Mitsuo Kawato 2, Rieko Osu 2, Masa-aki Sato 3 1 Dept. Intelligent Systems Design Eng., Toyama Pref. Univ., Toyama 2 CNS, ATR, Kyoto 3 NIA, ATR, Kyoto The measurement of EEG signals is contaminated by large electrical artifacts, such as eye movements. These artifacts can be orders of magnitude larger than the signal from the brain, thus making cortical current estimation extremely difficult. To remove the artifacts, here we improve the hybrid hierarchical variational Bayesian method (hyVBED) proposed by Fujiwara et al. (NeuroImage 2009). hyVBED simultaneously estimates the cortical and artifact currents while placing dipoles not only on cortical surfaces but also on artifactual sources. It requires the EOG data for constructing an EOG forward model describing the relationship between eye dipoles and electric potentials. In contrast, our improved approach does not require EOG, and in more general terms less knowledge about the current variance of artifacts is needed a priori. We varied the value of prior information for cortical surface and artifactual source dipoles, and their optimal values for estimation were explored systematically. Two criteria were taken for this search. First, the prior and posterior current variances should be the same. Second, the spatial patterns of fMRI activities and estimated current power should be highly correlated. In our previous studies, we applied this method to MEG data. Here, it is applied to EEG data. EEG is more difficult than MEG because the scalp electrical potential reflects the shells with different electric conductivities. The three-shell boundary element method (BEM) model derived from the MRI data set was employed. By introducing proper prior information, our approach can reasonably remove the effect of artifacts from contaminating EEG signals. Additionally, we compared the performance of our approach to the conventional statistical methods such as PCA and ICA. When the brain and artifact signals are statistically similar, our method outperforms these conventional statistical methods. Research fund: A part of this work was supported by NICT, and SRBPS from MEXT. doi:10.1016/j.neures.2011.07.907 P2-v10 Estimating waveforms common across trials from MEG data in various stimulus–response tasks Yusuke Takeda 1 , Kentaro Yamanaka 2, Noriko Yamagishi 3,4,5, Masaaki Sato 1 1 ATR NIA, Kyoto, Japan 2 Showa Women’s Univ., Tokyo, Japan 3 ATR CMC, Kyoto, Japan 4 JST PRESTO, Saitama, Japan 5 NICT Brain ICT Lab, Kyoto, Japan A generalized method to estimate waveforms common across trials from EEG/MEG was recently proposed. From single/multi-channel EEG/MEG epochs, the method estimates the number of waveforms common across trials, their delays in individual trials, and all of the pure waveforms that are not contaminated with each other. In this study, we apply the method to MEG data in various stimulus–response tasks, such as a Go/NoGo task, and compare the estimated waveforms across the tasks. We obtain two main findings. First, as the tasks become demanding, the numbers of the estimated waveforms become larger. This suggests that demanding tasks involve a larger number of brain processes than simple tasks. Second, the properties of waveforms time-locked to neither cue nor motor responses are similar across the tasks. This suggests that these waveforms reflect the same brain process in these tasks, such as recognizing the shape of stimuli. We confirm that it is the new method for neuroscience to reveal hypothesis-free data driven findings. Research fund: National Institute of Information and Communications Technology. doi:10.1016/j.neures.2011.07.908 P2-v11 Simultaneous NIRS and fMRI measurement of cortical hemodynamic fluctuations during the resting state Shuntaro Sasai 1 , Fumitaka Homae 2, Hama Watanabe 1, Akihiro Sasaki 3,4,5, Hiroki Tanabe 3,4, Norihiro Sadato 3,4,6, Gentaro Taga 1 1 Graduate School of Education, University of Tokyo, Tokyo, Japan 2 Department of Language Sciences, Tokyo Metropolitan University, Hachioji, Japan 3 National Institute for Physiological Sciences, Okazaki, Japan 4 The Graduate University for Advanced Studies (SOKENDAI), Okazaki, Japan 5 Japan Society for the Promotion of Science, Tokyo, Japan 6 Japan Science and Technology Agency, Kawaguchi, Japan The study of temporal correlation between different brain regions during the resting state, which is important to understand the intrinsic functional architecture of the brain, has been performed using fMRI (Fox and Raichle, 2007) and, more recently, by NIRS (Homae et al., 2010; Lu et al., 2010; Masquita et al., 2010; Sasai et al., in press; White et al., 2009; Zhang et al., 2010). The local relationship between simultaneous NIRS and fMRI signals recorded during tasks (Cui et al., in press; Steinbrink et al., 2006; Toronov et al., 2001; Toyoda et al., 2008) and the relationship between NIRS signals measured from limited regions and the blood-oxygen-level-dependent (BOLD) signals from the entire brain during the resting state (Tong and Frederick, 2010) have been examined. In this study, in order to clarify whether these different modalities are consistent with respect to spontaneous hemodynamic