Three-dimensional medical image analysis of the heart by the revised GMDH-type neural network self-selecting optimum neural network architecture

In this study, a revised group method of data handling (GMDH)-type neural network algorithm which self-selects the optimum neural network architecture is applied to 3-dimensional medical image analysis of the heart. The GMDH-type neural network can automatically organize the neural network architecture by using the heuristic self-organization method, which is the basic theory of the GMDH algorism. The heuristic self-organization method is a kind of evolutionary computation method. In this revised GMDH-type neural network algorithm, the optimum neural network architecture was automatically organized using the polynomial and sigmoid function neurons. Furthermore, the structural parameters, such as the number of layers, the number of neurons in the hidden layers, and the useful input variables, are selected automatically in order to minimize the prediction error criterion, defined as the prediction sum of squares (PSS).     

A study on the effect of morphological filters on computer-aided medical image diagnosis

We have developed several morphological image filters that can be useful for computer-aided medical image diagnosis. Several computer-aided diagnosis (CAD) systems for lung cancer and breast cancer have been developed to assist the radiologist’s diagnostic work. The CAD systems for lung cancer can automatically detect pathological changes (pulmonary nodules) with a high true-positive rate (TP) even under low false-positive rate (FP) conditions. On the other hand, the conventional CAD systems for breast cancer can automatically detect some pathological changes (calcifications and masses), but the TP for other changes, such as architectural distortion, is still very low. Motivated by the radiologist’s cognitive processes to increase TP for breast cancer, we propose new methods to extract novel morphological features from X-ray mammography. Simulation results demonstrate the effectiveness of the morphological methods for detecting tumor shadows.     

Dynamic biventricular response to alterations in preload in patients undergoing left ventricular device implantation

Ventricular interdependence is important for the successful use of a left ventricular assist device (LVAD) because the filling of the device depends on right ventricular (RV) function as well as the interactions between the ventricles. The pulmonary arterial (PAP) and systemic arterial (AP) response to inferior vena caval (IVC) occlusion before and after insertion of an LVAD in 15 patients was used to "dissect out" the determinants of these interactions. PAP and AP were recorded during each IVC occlusion and peak systolic values calculated for each beat. Linear regression analysis was used to calculate the slope (k) between peak systolic AP values and peak systolic PAP values. k, a measure of preload responsiveness of the heart, is predominantly linear. k is relatively "flat" in selected LV failure patients pre-LVAD but increases significantly (P < 0.001) after LVAD (0.67 +/- 0.55 vs. 2.71 +/- 1.39). The increase in this parameter after LVAD suggests that the loss of RV-to-LV ventricular interdependence in patients with congestive heart failure appears to recover somewhat once an LVAD is inserted.     

Zinc oxide thin-film chemical sensors in conjunction with neural network pattern recognition for trimethylamine and dimethylamine gases

Transient response curves for exposure to several gases are observed using zinc oxide (ZnO) thin-film gas sensors. It is found that an aluminium-doped ZnO (ZnO:Al) sensor exhibits a high sensitivity and an excellent selectivity for amine gases. In order to discriminate between gas species such as trimethylamine (TMA), dimethylamine (DMA) and other gases pattern recognition analysis using a neural network is carried out using parameters which characterize the transient responses of the sensor for exposure to gases. The recognition probability of the neural network is 90% for TMA and DMA with constant concentration and is 100% for TMA and DMA with different concentrations, except for a concentration of 1 p.p.m.     

Medical image diagnosis of lung cancer by multi-layered GMDH-type neural network self-selecting functions

In this study, a revised Group Method of Data Handling (GMDH)-type neural network self-selecting functions is applied to the computer aided image diagnosis (CAD) of lung cancer. The GMDH-type neural network algorithm has an ability of self-selecting optimum neural network architecture from three neural network architectures, such as sigmoid function neural network, radial basis function neural network and polynomial neural network. The GMDH-type neural network also has abilities of self-selecting the number of layers, the number of neurons in hidden layers and useful input variables. This algorithm is applied to CAD of lung cancers, and it is shown that this algorithm is useful for the CAD, and is very easy to apply to practical complex problems because optimum neural network architecture is automatically organized.     

Synthesis of Lead Barium Niobate Powders and Thin Films by the Sol-Gel Method

Crack-free, dense, and transparent Pb_0.6Ba_0.4Nb₂O₆ (PBN60) thin films have been prepared by a sol-gel method with metal alkoxides and metal acetate. A homogeneous and stable precursor solution was obtained from Ba metal, Pb(CH₃COO)₂, and Nb(OEt)₅ in 2-methoxyethanol. PBN60 powder crystallized to the hexagonal phase at 600°C and then completely transformed to the orthorhombic phase of the tungsten bronze structure at 1250°C. The hexagonal phase was formed on SiO₂ glass, MgO(lOO), and sapphire(R) substrate at 600°C, while the orthorhombic phase was only on a sapphire(C) substrate. Orthorhombic PBN60 films with c -axis preferred orientation were successfully synthesized on sapphire(C) substrates at 600°C.     

Analysis of factors controlling catalytic activity by neural network

An artificial neural network was applied to the analysis of factors controlling catalytic activity by taking, as examples, experimentally established correlations of catalytic activities with primary factors including both monotonous and volcano-type correlations. Three equations were proposed and applied to the estimation of relative importance of each given factor from the weightings of connecting links in the trained artificial neural network of an error back-propagation model. In all the examples, the primary factors that had been proposed in experimental studies were successfully identified by using an equation based on our previous proposal. Further, the possibility of identifying secondary factor was also discussed.     

Highly controlled carbodiimide reaction for the modification of lysozyme. Modification of Leu129 or Asp119

In the cross-linking reaction of lysozyme between Leu¹²⁹ (-COO^–)and Lys¹³ (-NH₃⁺ using imidazole and 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimidehydrochloride (EDC), a side reaction of the peptide bond inversionfrom to ß between A and Gly¹⁰² was greatly reducedby addition of ß-(1,4)-linked trimer of N-acetyl-D-glucosamine[(NAG)₃] When methylamine or 2-hydroxyethylamine was furtheradded, the extent of the cross-link formation was decreasedand the derivative where the -carboxyl group of Leu¹²⁹ was modifiedwith the amine was newly obtained. On the other hand, when ammoniawas added, the ß-carboxyl group of Asp¹¹⁹ insteadof the -carboxyl group was mainly amidated. From these results,the presence of a salt bridge between Asp¹¹⁹ and Arg¹²⁵ besidesthat between Lys¹³ and Leu¹²⁹ is proposed. Enzymatic activitiesof the derivatives prepared here indicated that the modificationof the -carboxyl group reduced the activity to {small tilde}90% of that of native lysozyme. Des-Leu¹²⁹ lysozyme, which lacksLeu¹²⁹ also showed {small tilde} 90% of the activity of nativelysozyme. Therefore, the salt bridge between Lys¹³ and Leu¹²⁹may play some role in maintaining the active conformation oflysozyine.     

Chitosan-hybridized acrylic resins prepared in emulsion polymerizations and their application as interior finishing coatings

Chitosan was used to introduce formaldehyde adsorption abilities into an emulsion binder for interior finishing coatings. Chitosan-hybridized acrylic emulisions were prepared by two methods. To produce stable chitosan-hybridized acrylic emulsions, the pre-emulsion dropping method is superior to the monomer dropping method. The adsorption performances for fonnaldehyde in chitosan-hybridized acrylic resin films increased with increasing chitosan contents; the films also had adsorption abilities for hydrogen sulfide and ammonia. Tensile strengths and elongation at breaking points decreased with increasing chitosan contents. Interior finishing coatings made from chitosan-hybridized acrylic resin emulsions have the qualities necessary for an interior finishing coating and showed excellent adsorption abilities for formaldehyde.