Search

Objective. Advances in brain–machine interfaces (BMIs) are expected to support patients with movement disorders. Electrocorticogram (ECoG) measures electrophysiological activities over a large area using a low-invasive flexible sheet placed on the cortex. ECoG has been considered as a feasible signal source of the clinical BMI device. To capture neural activities more precisely, the feasibility of higher-density arrays has been investigated. However, currently, the number of electrodes is limited to approximately 300 due to wiring difficulties, device size, and system costs. Approach. We developed a high-density recording system with a large coverage (14 × 7 mm2) and using 1152 electrodes by directly integrating dedicated flexible arrays with the neural-recording application-specific integrated circuits and their interposers. Main results. Comparative experiments with a 128-channel array demonstrated that the proposed device could delineate the entire digit representation of a nonhuman primate. Subsampling analysis revealed that higher-amplitude signals can be measured using higher-density arrays. Significance. We expect that the proposed system that simultaneously establishes large-scale sampling, high temporal-precision of electrophysiology, and high spatial resolution comparable to optical imaging will be suitable for next-generation brain-sensing technology.

In the over-realizable learning scenario of multilayer perceptrons, in which the student network has a larger number of hidden units than the true or optimal network, some of the weight parameters are unidentifiable. In this case, the teacher network consists of a union of optimal subspaces included in the parameter space. The optimal subspaces, which lead to singularities, are known to affect the estimation performance of neural networks. Using statistical mechanics, we investigate the online learning dynamics of two-layer neural networks in the over-realizable scenario with unidentifiable parameters. We show that the convergence speed strongly depends on the initial parameter conditions. We also show that there is a quasi-plateau around the optimal subspace, which differs from the well-known plateaus caused by permutation symmetry. In addition, we discuss the property of the final learning state, relating this to the singular structures.

We study the process of growth of graphene by the SiC surface decomposition method to improve the technique of growing graphene. In this study, the initial stage of the graphitization process on a periodically faceted SiC surface is investigated by the Naval Research Laboratory tight-binding approach. The results suggest that the decomposition of the SiC surface started by the desorption of Si atoms at the step edge. It is also found that released C atoms are trapped at the step-edge sites and that neighboring Si atoms are subsequently desorbed. These results suggest that graphene nucleation on the SiC surface proceeds in a negative step-flow mode.

Statistical haplotype inference is an indispensable technique in the field of medical science. The method usually has two steps: inference of haplotype frequencies and inference of diplotype for each subject. The first step can be done by using the expectation-maximization (EM) algorithm, but it incurs an unreasonably large calculation cost when the number of single-nucleotide polymorphism (SNP) loci of concern is large. In this article, we describe an approximate probabilistic model of haplotype frequencies. The model is constructed by using several distributions of nearby local SNPs. This approximation seems good because SNPs are generally more strongly correlated when they are close to one another on a chromosome. To implement this approach, we use a log linear model, the Walsh-Hadamard transform, and a combinatorial optimization method. Artificial data suggested that the overall haplotype inference of our method is good if there are nine or more local consecutive SNPs. Some minor problems should be dealt with before this method can be applied to real data.

We investigated the thermal properties of a single-walled carbon nanotube with vacancy defects to determine its applicability to solid-state thermal rectification. Nonequilibrium molecular dynamics simulation of a nanotube with randomly located defects only along half the length revealed asymmetric heat conduction at room temperature. The direction of rectification is in good agreement with that obtained in a past experiment in which C9H16Pt was asymmetrically deposited on a nanotube, as far as the local deposition is supposed to cause defects in the nanotube lattice. The mechanism underlying the current thermal rectification effect is discussed considering the temperature dependence of the local thermal conductivity and the phonon filtering effect. The calculated phonon density of states shows larger overlapping when heat flows from the defective part to the pristine part and intermediate-frequency phonons are mainly responsible for rectification.

The method of statistical haplotype inference is an indispensable technique in the field of medical science. The authors previously reported Hardy-Weinberg equilibrium-based haplotype inference that could manage single nucleotide polymorphism (SNP) data. We recently extended the method to cover copy number variation (CNV) data. Haplotype inference from mixed data is important because SNPs and CNVs are occasionally in linkage disequilibrium. The idea underlying the proposed method is simple, but the algorithm for it needs to be quite elaborate to reduce the calculation cost. Consequently, we have focused on the details on the algorithm in this study. Although the main advantage of the method is accuracy, in that it does not use any approximation, its main disadvantage is still the calculation cost, which is sometimes intractable for large data sets with missing values.

Statistical mechanical informatics (SMI) is an approach that applies statistical-mechanics techniques developed for many-body problems in natural science to information sciences. SMI has been extensively applied to various research fields of information sciences, including in particular information and communication theory, probabilistic inference, and combinatorial optimization problems, and has brought about novel advances in these fields. From a formal point of view, the most significant advantage of the SMI approach would be that it facilitates our look at hierarchical organization of a complex system, on the basis of which reduction of many `microscopic' degrees of freedom is performed to obtain a `macroscopic' description of the system with a small number of parameters, which is in turn utilized to analyze the system microscopically. Hierarchy, reduction, and macroscopic description should also be essential in order to understand biological systems. These observations imply that the conceptual framework of SMI should be extensible to various research fields in bioinformatics, and new theoretical frameworks for bioinformatics will be established from the system science viewpoint. To advance research activities in these directions, however, interactions between SMI and bioinformatics are needed. The International Workshop on Statistical-Mechanical Informatics has been organized in order to explore such new directions for SMI and bioinformatics through interactions between various research disciplines, and to incubate joint research initiatives. This workshop, held at Mielparque-Kyoto, Kyoto, Japan, 13–16 September 2009, and sponsored by the Grant-in-Aid for Scientific Research on Priority Areas `Deepening and Expansion of Statistical Mechanical Informatics (DEX-SMI)' (Head investigator: Yoshiyuki Kabashima, Tokyo Institute of Technology) (Project http://dex-smi.sp.dis.titech.ac.jp/DEX-SMI), was intended to provide leading researchers in the fields of DNA microarray analysis, brain science, medical engineering and mathematical biology, to exchange advanced findings on and to detail achievements in bioinformatics and statistical mechanics. We would like to thank the contributors of the workshop as well as all the participants, who have enjoyed the workshop as well as their stay in Kyoto. This successful workshop will stimulate further development of the interdisciplinary research field of bioinformatics and SMI. Editors Masato Inoue Shin Ishii Yoshiyuki Kabashima Masato Okada The IW-SMI 2009 Organizing Committee Masato Okada, General Chair (University of Tokyo) Yoshiyuki Kabashima, Vice-General Chair (Tokyo Institute of Technology) Shin Ishii, Program Chair (Kyoto University) Masato Inoue, Publications Chair (Waseda University) Kazuyuki Tanaka (Tohoku University) Toshiyuki Tanaka (Kyoto University)