Human brain disturbance by methylmercury poisoning, focusing on the long-term effect on brain weight

The long-term effects of the initial methylmercury exposure in the Minamata area were examined, with focus on brain weight. Comparisons were made between the brain weights of methylmercury-contaminated persons in the Minamata area (cases) and those of Japanese who suffered sudden or accidental death without brain disease (controls). Since over thirty years have passed since the outbreak of Minamata disease, subjects aged 30 or over at death were included in the analysis. The total number of cases was 417 (273 males, 144 females) and that of controls was 2,934 (2,174 males, 760 females). The data suggest that the brain weights from Minamata cases were reduced by 80 to 200g as compared to the controls. These results may indicate that methylmercury exposure led to a decrease of nerve cells (single cell necrosis), so-called "thinning out", in the brain cortex.     

Bacteriological, pharmacokinetic and clinical evaluation of azithromycin in the pediatric field]

Azithromycin (AZM), a new macrolide antibiotic, in fine granules and in capsules was studied for pharmacokinetic and clinical evaluation in the pediatric patients. Antibacterial activity of AZM against 43 clinical isolates: AZM exhibited slightly lower activity against Gram-positive bacteria and 2-8-fold higher activity against Gram-negative bacteria than erythromycin or clarithromycin. Plasma or urine samples were collected from eight patients receiving the drug in fine granular form, and two patients receiving it in capsules for the determination of drug levels. The elimination half-lives of AZM after administration at dose of 10 mg/kg/day for 3 days were 50.0 and 51.2 hours for fine granules, and 41.5 hours for capsules. AUC0-infinity was 11.7 and 24.3 micrograms.hr/ml for fine granules, and 8.3 micrograms.hr/ml for capsules. The cumulative excretion rates up to 120 hours after the start of treatment were 8.24 and 13.84% for fine granules, and 3.83% for capsules. AZM was administered to 123 patients once daily at 3.7-20.0 mg/kg body weight over 3 to 5 days with reference to the standard dose of 10 mg/kg. The drug was used to treat patients with pharyngitis, tonsillitis, scarlet fever, pneumonia, mycoplasmal pneumonia, chlamydial pneumonia, otitis media, pertussis, intestinal infection, and SSTI. The effectiveness of AZM was evaluated in 109 cases. The drug was rated "excellent" in 65.1% of the patients and "good" in 29.4%, resulting in an efficacy rate of 94.5%. Furthermore, AZM eradicated 43 of 46 (93.5%) bacteria that had been identified before the treatment. Three patients complained of side effects of urticaria (1 case) and diarrhea (2 cases). Abnormal laboratory changes were reported as follows: decreased leukocyte (3 cases), increased eosinophil (5), increased platelet (2), increased eosinophil and platelet, elevated GPT (1), and elevated GOT and GPT (1). The abnormalities, however, were mild enough to raise no clinically significant problems. In conclusion, AZM in once daily regimen was effective and safe in treatment of pediatric infections.     

Efficient Leave-<Emphasis Type="Italic">m</Emphasis>-out Cross-Validation of Support Vector Regression by Generalizing Decremental Algorithm

We propose a computationally efficient method for cross-validation of the Support Vector Regression (SVR) by generalizing the decremental algorithm of SVR. Incremental and decremental algorithm of Support Vector Machines (SVM) ^{2, 8, 9)} efficiently update the trained SVM model when a single data point is added to or removed from the training set. The computational cost of leave-one-out cross-validation can be reduced using the decremental algorithm. However, when we perform leave-m-out cross-validation (m > 1), we have to repeatedly apply the decremental algorithm for each data point. In this paper, we extend the decremental algorithm of SVR^{8, 9)} in such a way that several data points can be removed more efficiently. Experimental results indicate that the proposed approach can reduce the computational cost. In particular, we observed that the number of breakpoints, which is the main computational cost of the involved path-following, were reduced from \({\mathcal O}(m)\) to \({\mathcal O}(\sqrt{m})\).     

A three-dimensional shape memory alloy microelectrode with clippingstructure for insect neural recording

A microelectrode, with clipping structure for neural recording from a free-moving insect, was designed and fabricated using a shape memory alloy (SMA) thin film. The SMA thin films (titanium nickel; Ti-48 at.%Ni) are deposited by RF magnetron sputtering and patterned by HF-HNO ₃ wet etching. The transformation temperatures of the SMA thin films were measured at 54°C (A*) and 50°C (M*). The SMA microelectrode consists of a “hook” structure (720 μm×480 μm) and two “C”-shape probes (600 μm×70 μm). The electrode impedance is about 5 kΩ at 1 kHz. The desired three-dimensional (3D) shape is given to the electrode by a bonded wire. The clinging force of the electrode to the nerve is enhanced by the 3-D structures. The SMA microelectrode can clip a nerve cord tightly. The damages of the nerve by thermal actuation of the clip are not observed by physiological analysis. The neural activity from a living insect was successfully recorded with this SMA microelectrode     

A knowledge assimilation method for logic databases

In this paper we consider a deductive question-answering system for relational databases as a logic database system, and propose a knowledge assimilation method suitable for such a system. The concept of knowledge assimilation for deductive logic is constructed in an implementable form based on the notion of amalgamating object language and metalanguage. This concept calls for checks to be conducted on four subconcepts, provability, contradiction, redundancy, independency, and their corresponding internal database updates. We have implemented this logic database knowledge assimilation program in PROLOG, a logic programming language, and have found PROLOG suitable for knowledge assimilation implementation.     

Object oriented programming in Concurrent Prolog

It is shown that the basic operations of object-oriented programming languages — creating an, object, sending and receiving messages, modifying an object’s state, and forming class-superclass hierarchies — can be implemented naturally in Concurrent Prolog. In addition, a new object-oriented programming paradigm, called incomplete messages, is presented. This paradigm subsumes stream communication, and greatly simplifies the complexity of programs defining communication networks and protocols for managing shared resources. Several interesting programs are presented, including a multiple-window manager. All programs have been developed and tested using the Concurrent Prolog interpreter described in.¹⁾     

Flood Management in Japan—From Rivers to Basins

Abstract

The River Commission of the Ministry of Construction in Japan recommended a basin response to floods in its December 2000 Interim Report to Minister of Construction. This is practically an official declaration of the policy accepting inundation in habited areas. By this policy shift, it is anticipated that more basin response measures and non-structural measures against floods will be promoted to complement structural measures such as continuous levees and dams. Such a policy is not new in the world but is a significant governmental move in Japan, a densely populated, highly developed, and highly flood hazardous country. Although living with floods is a universal coping strategy, the implementation of measures should be different from region to region reflecting region's nature and socio-economic conditions. This paper reviews the flood fighting history of Japan and her statutory evolution against floods and shows the process by which the ever-increasing flood damage potential and the recent increase of flood damages necessitated the official move in flood control management from rivers to basins.     

First principles total energy studies of the adsorption of germane on Ge(001)-c(2 × 4)

We perform first principles total energy calculations to study the energetics, and the atomic structure of the adsorption of germane (GeH₄) molecules on the Ge(001)-c(2 × 4) surface. The adsorption of a GeH₄ unit occurs after its dissociation into a germanium trihydride (GeH₃) and a hydrogen atom and a subsequent decomposition into a germanium dihydride (GeH₂) subunit and H atoms. Consequently, we first consider the adsorption of GeH₂ in two different configurations; the on-dimer and the intra-row geometries. Similar to the adsorption of SiH₂ and GeH₂ on Si(001), it is found that the on-dimer site is more stable than the intra-row geometry by 0.13 eV. However, in the adsorption of a GeH₂ fragment together with two H atoms we find that the intra-row geometry is energetically more favorable, again, similar to the adsorption of SiH₂ and GeH₂ (plus two H atoms) on the Si(001) surface.     

Efficient distribution-free population learning of simple concepts

We consider a variant of the ‘population learning model’ proposed by Kearns and Seung [8], in which the learner is required to be ‘distribution-free’ as well as computationally efficient. A population learner receives as input hypotheses from a large population of agents and produces as output its final hypothesis. Each agent is assumed to independently obtain labeled sample for the target concept and output a hypothesis. A polynomial time population learner is said to PAC-learn a concept class, if its hypothesis is probably approximately correct whenever the population size exceeds a certain bound which is polynomial, even if the sample size for each agent is fixed at some constant. We exhibit some general population learning strategies, and some simple concept classes that can be learned by them. These strategies include the ‘supremum hypothesis finder’, the ‘minimum superset finder’ (a special case of the ‘supremum hypothesis finder’), and various voting schemes. When coupled with appropriate agent algorithms, these strategies can learn a variety of simple concept classes, such as the ‘high–low game’, conjunctions, axis-parallel rectangles and others. We give upper bounds on the required population size for each of these cases, and show that these systems can be used to obtain a speed up from the ordinary PAC-learning model [11], with appropriate choices of sample and population sizes. With the population learner restricted to be a voting scheme, what we have is effectively a model of ‘population prediction’, in which the learner is to predict the value of the target concept at an arbitrarily drawn point, as a threshold function of the predictions made by its agents on the same point. We show that the population learning model is strictly more powerful than the population prediction model. Finally, we consider a variant of this model with classification noise, and exhibit a population learner for the class of conjunctions in this model. This revised version was published online in June 2006 with corrections to the Cover Date.