Double frequency synchronization for working memory operation in the brain - a computational study for bridging human scalp EEG data and neural dynamics

Abstract

Event Abstract Back to Event Double frequency synchronization for working memory operation in the brain - a computational study for bridging human scalp EEG data and neural dynamics Yoko Yamaguchi1*, David Chik1, Masahiro Kawasaki2 and Hideki Oka3 1 RIKEN BSI, Lab for Dynamics of Emergent Intelligence, Japan 2 RIKEN BTCC, Rhythm-based Brain Computation Unit, Japan 3 Osaka University, The Center for Adv Med Eng and Informatics, Japan Current studies on brain dynamics in various levels more and more reveal the dynamical brain coordinated by synchronization of oscillations. In the decade, the synchronization hypothesis for dynamical linking was extensively studied in animals and humans. Most of neural systems are known to show various oscillation frequencies in cellular levels and also during cognitive tasks in human scalp EEG. Among them, complexity of synchfonization dynamics provides fascinating problems for understanding neural mechanisms of intellectual properties. According to the dynamical system theory (e.g. the slaving principle by H Haken), slow oscillations of human EEG, such as the delta, theta and alpha can concern with higher cognitive functions to regulate local processes with fast oscillations. We hypothesize that synchronization network in each frequency band has each role in cognitive functions and that their dynamical linking among modules within the same frequency band or through cross-frequency coupling could execute the real-time computation by implementing multiple constraints for cognitive control. Recently we have reported that EEG theta rhythm related network (4~8 Hz) distributing over frontal midline regions and posterior regions is related with central executive functions (Mizuhara et al. 2007) and furthermore that the theta rhythm network emerges restrictively in manipulation period of working memory(Kawasaki et al. 2010). In the working memory task, alpha rhythm (~10 Hz) around relevant sensory regions is observed in both of storage period and operation period. interestingly, theta and alpha rhythms exhibit one to two phase-locking in the manipulation period. The network mechanism of this double frequency synchronization and its possible functional mechanism are unsolved. In this paper, we propose a alpha-theta cross frequency coupling network model for the working memory task. In the model, alpha oscillation networks are described by flip-flop oscillation networks (Colliaux et al,. 2009). A limit cycle attractor in the up state of the associative memory network represents the working memory storage. Theta network is assumed to consist of frontal midline module and sensory modality specific modules with mutual coupling. By introducing a coupling between theta and alpha rhythm units, we show theta network can regulate alpha rhythm synchronization so that dynamically linking among alpha oscillations can retrieve or manipulate memory. This is in agreement with the results suggested by Kawasaki et al. (2010). We conclude that double frequency phase locking between theta and alpha oscillations can work for cognitive control through hierarchical dynamical linking. Then, for the sake of direct comparison with our oscillation network model and human EEG scalp data, we developed a simulation system of scalp EEG by assuming electronic charge in a given location of the brain. Our programming Using InsilicoML and InsilicoDE, showed usefulness of this system on basic simulation and also suggested possible computational neuroscience approach from complex neural dynamics for understanding cognitive functions. This study was partially supported by Grant-in-Aid for Scientific Research on Innovative Areas No. 4103. The simulation study is supported by physiome.jp and dynamic brain platform (DBPF).