Search

Bio-artificial livers (BAL) with microencapsulated hepatocytes have the typical limitations in maintaining hepatocyte functions, mechanical stability and uniform perfusion in packed or fluidized-bed bioreactors. We have previously developed microcapsules with enhanced hepatocyte functions. Here we have introduced a fibrin network inside microcapsules by (1) mixing collagen and fibrinogen with the encapsulated hepatocytes to support the cells; (2) submerging the microcapsules into a thrombin solution to induce the formation of an insoluble fibrin network inside the microcapsules. Fracture analysis on the microcapsules revealed significant improvement in mechanical stability. We have also introduced different amounts of gold nano-particles into microcapsules to achieve different densities for uniform bioreactor perfusion. These gold nano-particles also improved the mechanical stability of the microcapsules. Both the fibrin network and gold nano-particles exhibited the additional benefits of enhancing certain bio-functions of the encapsulated hepatocytes. The applications of these improved microcapsules in the development of bio-artificial livers are discussed.

Bank failure prediction is an important issue for the regulators of the banking industries. The collapse and failure of a bank could trigger an adverse financial repercussion and generate negative impacts such as a massive bail out cost for the failing bank and loss of confidence from the investors and depositors. Very often, bank failures are due to financial distress. Hence, it is desirable to have an early warning system (EWS) that identifies potential bank failure or high-risk banks through the traits of financial distress. Various traditional statistical models have been employed to study bank failures [J Finance 1 (1975) 21; J Banking Finance 1 (1977) 249; J Banking Finance 10 (1986) 511; J Banking Finance 19 (1995) 1073]. However, these models do not have the capability to identify the characteristics of financial distress and thus function as black boxes. This paper proposes the use of a new neural fuzzy system [Foundations of neuro-fuzzy systems, 1997], namely the Generic Self-organising Fuzzy Neural Network (GenSoFNN) [IEEE Trans Neural Networks 13 (2002c) 1075] based on the compositional rule of inference (CRI) [Commun ACM 37 (1975) 77], as an alternative to predict banking failure. The CRI based GenSoFNN neural fuzzy network, henceforth denoted as GenSoFNN-CRI(S), functions as an EWS and is able to identify the inherent traits of financial distress based on financial covariates (features) derived from publicly available financial statements. The interaction between the selected features is captured in the form of highly intuitive IF-THEN fuzzy rules. Such easily comprehensible rules provide insights into the possible characteristics of financial distress and form the knowledge base for a highly desired EWS that aids bank regulation. The performance of the GenSoFNN-CRI(S) network is subsequently benchmarked against that of the Cox's proportional hazards model [J Banking Finance 10 (1986) 511; J Banking Finance 19 (1995) 1073], the multi-layered perceptron (MLP) and the modified cerebellar model articulation controller (MCMAC) [IEEE Trans Syst Man Cybern: Part B 30 (2000) 491] in predicting bank failures based on a population of 3635 US banks observed over a 21 years period. Three sets of experiments are performed-bank failure classification based on the last available financial record and prediction using financial records one and two years prior to the last available financial statements. The performance of the GenSoFNN-CRI(S) network as a bank failure classification and EWS is encouraging.

Neural fuzzy networks proposed in the literature can be broadly classified into two groups. The first group is essentially fuzzy systems with self-tuning capabilities and requires an initial rule base to be specified prior to training. The second group of neural fuzzy networks, on the other hand, is able to automatically formulate the fuzzy rules from the numerical training data. Examples are the Falcon-ART, and the POPFNN family of networks. A cluster analysis is first performed on the training data and the fuzzy rules are subsequently derived through the proper connections of these computed clusters. This correspondence proposes two new networks: Falcon-FKP and Falcon-PFKP. They are extensions of the Falcon-ART network, and aimed to overcome the shortcomings faced by the Falcon-ART network itself, i.e., poor classification ability when the classes of input data are very similar to each other, termination of training cycle depends heavily on a preset error parameter, the fuzzy rule base of the Falcon-ART network may not be consistent Nauck, there is no control over the number of fuzzy rules generated, and learning efficiency may deteriorate by using complementarily coded training data. These deficiencies are essentially inherent to the fuzzy ART, clustering technique employed by the Falcon-ART network. Hence, two clustering techniques--Fuzzy Kohonen Partitioning (FKP) and its pseudo variant PFKP, are synthesized with the basic Falcon structure to compute the fuzzy sets and to automatically derive the fuzzy rules from the training data. The resultant neural fuzzy networks are Falcon-FKP and Falcon-PFKP, respectively. These two proposed networks have a lean and efficient training algorithm and consistent fuzzy rule bases. Extensive simulations are conducted using the two networks and their performances are encouraging when benchmarked against other neural and neural fuzzy systems.

Packed-bed or fluidized-bed bioreactor filled with microencapsulated hepatocytes has been proposed as one of the promising designs for bioartificial liver assist device (BLAD) because of potential advantages of high mass transport rate and optimal microenvironment for hepatocyte culture. Recently, we have developed a microcapsule system for the encapsulation of hepatocytes. The microcapsules consist of an inner core of modified collagen and an outer shell of terpolymer of methyl methacrylate, methacrylate and hydroxyethyl methacrylate. Cells encapsulated in these microcapsules exhibit enhanced cellular functions. Improving the mechanical stability of the microcapsules to withstand the shear stress induced by high perfusion rate would be crucial to the success of BLAD applications. In this study, we investigated the effects of terpolymer molecular weight (M(w)) on the mechanical property of these microcapsules and the differentiated functions of encapsulated hepatocytes. Six terpolymers with different M(w) were synthesized using radical polymerization in solution by adjusting the reaction temperature and the initiator concentration. All the terpolymers formed microcapsules with the methylated collagen. While the terpolymer M(w) had little effect on the capsule membrane thickness and permeability of serum albumin, the mechanical property of the microcapsules was significantly improved by the higher M(w) of the terpolymer. Differentiated functions of the hepatocytes cultured in the microcapsules, including urea synthesis, albumin synthesis and cytochrome P450 metabolic activity, were not significantly affected by the terpolymer M(w)., (Copyright 2003 Elsevier Science Ltd.)

A pseudo-outer product based fuzzy neural network using the compositional rule of inference and singleton fuzzifier [POPFNN-CRI(S)] is proposed in this paper. The correspondence of each layer in the proposed POPFNN-CRI(S) to the compositional rule of inference using standard T-norm and fuzzy relation gives it a strong theoretical foundation. The proposed POPFNN-CRI(S) training consists of two phases; namely: the fuzzy membership derivation phase using the novel fuzzy Kohonen partition (FKP) and pseudo Kohonen partition (PFKP) algorithms, and the rule identification phase using the novel one-pass POP learning algorithm. The proposed two-phase learning process effectively constructs the membership functions and identifies the fuzzy rules. Extensive experimental results based on the classification performance of the POPFNN-CRI(S) using the Anderson's Iris data are presented for discussion. Results show that the POPFNN-CRI(S) has taken only 15 training iterations and misclassify only three out of all the 150 patterns in the Anderson's Iris data.

A cogent understanding of the microstructure, and indeed nano-structure, of hydroxyapatite (HA) and the interface between Ti-6Al-4V and HA is crucial to its appropriateness as a biomaterials. This paper reports the analysis of plasma-sprayed HA/Ti-6Al-4V composites by transmission electron microscopy (TEM) and scanning transmission electron microscopy (STEM) to elucidate the intricate nature of the materials following plasma spray processing and in vitro evaluation. The novel Ti-6Al-4V/HA composite coating, with approximately 48 wt% HA, had demonstrated attractive tensile adhesion strength (approximately 28 MPa) and improved Young's modulus (approximately 55 GPa). Experimental results demonstrated that amorphous calcium phosphate and fine HA grains were formed during rapid splat solidification in the as-sprayed composite coatings. Small Ti-6Al-4V grains were observed adjacent to the amorphous calcium phosphate. The coatings were further heat treated at 600 degrees C for 6 h, and significant crystallisation of the amorphous calcium phosphate phase took place. However, complete crystallisation was not achieved at this temperature, as the coatings invariably contained a small amount of amorphous calcium phosphate phase in some local regions. After immersion in simulated body fluid for 2 weeks and 10 weeks, TEM and STEM confirmed that the interfaces inside the coating maintained good microstructural integrity., (Copyright 2002 Elsevier Science Ltd.)

Mechanical stability, complete encapsulation, selective permeability, and suitable extra-cellular microenvironment, are the major considerations in designing microcapsules for cell encapsulation. We have developed four types of multi-layered microcapsules that allow selective optimization of these parameters. Primary hepatocytes were used as model cells to test these different microcapsule configurations. Type-1 microcapsules with an average diameter of 400 microm were formed by complexing modified collagen with a ter-polymer shell of 2-hydroxyethyl methylacrylate (HEMA), methacrylic acid (MAA) and methyl methacrylate (MMA), resulting in a capsule thickness of 2-5 microm. Cells in these microcapsules exhibited improved cellular functions over those cultured on collagen monolayers. Type-II microcapsules were formed by encapsulating the Type-I microcapsules in another 2-5 microm ter-polymer shell and a approximately 5 microm collagen layer between the two ter-polymer shells to ensure complete cell encapsulation. Type-II microcapsules comprised of a macro-porous exoskeleton with materials such as alumina sol-gel coated on the Type-I microcapsules. Nano-indendation assay indicated an improved mechanical stability over the Type-I microcapsules. Type-IV microcapsules were created by encapsulating Type-III microcapsules in another 2-5 microm ter-polymer shell, with the aim of imparting a negatively charged smooth surface to minimize plasma protein absorption and ensure complete cell encapsulation. The permeability for nutrient exchange, cellular functions in terms of urea production and mechanical stability of the microcapsules were characterized. The advantages and limitations of these microcapsules for tissue engineering are discussed.

Existing neural fuzzy (neuro-fuzzy) networks proposed in the literature can be broadly classified into two groups. The first group is essentially fuzzy systems with self-tuning capabilities and requires an initial rule base to be specified prior to training. The second group of neural fuzzy networks, on the other hand, is able to automatically formulate the fuzzy rules from the numerical training data. No initial rule base needs to be specified prior to training. A cluster analysis is first performed on the training data and the fuzzy rules are subsequently derived through the proper connections of these computed clusters. However, most existing neural fuzzy systems (whether they belong to the first or second group) encountered one or more of the following major problems. They are (1) inconsistent rule-base; (2) heuristically defined node operations; (3) susceptibility to noisy training data and the stability-plasticity dilemma; and (4) needs for prior knowledge such as the number of clusters to be computed. Hence, a novel neural fuzzy system that is immune to the above-mentioned deficiencies is proposed in this paper. This new neural fuzzy system is named the generic self-organizing fuzzy neural network (GenSoFNN). The GenSoFNN network has strong noise tolerance capability by employing a new clustering technique known as discrete incremental clustering (DIC). The fuzzy rule base of the GenSoFNN network is consistent and compact as GenSoFNN has built-in mechanisms to identify and prune redundant and/or obsolete rules. Extensive simulations were conducted using the proposed GenSoFNN network and its performance is encouraging when benchmarked against other neural and neural fuzzy systems.

The results of a retrospective study of fall incidence during a 1-year period in a psychiatric hospital in Singapore are reported, involving 309 patients who fell one or more times during their stay. The profile of inpatients involved in falls was identified. Data were derived from standard incident forms completed whenever patients had a fall. A total of 348 falls were identified for the 1-year period. Results revealed higher fall rates in younger male epileptic patients; elderly female patients, age 70 and older with depression or dementia; individuals with concomitant medical conditions; and patients taking three types of medication. The majority of falls occurred when the activity level was high in the ward. Young epileptic patients and elderly depressed patients were prone to repeated falls. These results have the potential to assist in identifying patients at high risk and in designing and implementing strategies to prevent such incidents.

A novel fuzzy neural network, the pseudo outer-product-based fuzzy neural network using the singleton fuzzifier together with the approximate analogical reasoning schema, is proposed in this paper. The network is referred to as the singleton fuzzifier POPFNN-AARS, the singleton fuzzifier POPFNN-AARS employs the approximate analogical reasoning schema (AARS) instead of the commonly used truth value restriction (TVR) method. This makes the structure and learning algorithms of the singleton fuzzifier POPFNN-AARS simple and conceptually clearer than those of the POPFNN-TVR model. Different similarity measures (SM) and modification functions (FM) for AARS are investigated. The structures and learning algorithms of the proposed singleton fuzzifer POPFNN-AARS are presented. Several sets of real-life data are used to test the performance of the singleton fuzzifier POPFNN-AARS and their experimental results are presented for detailed discussion.

The aim of this project was to compare dual and tri-colour reagents for lymphocyte immunophenotyping. A total of 37 patient and normal specimens were immunophenotyped concurrently with the following mean values (% dual vs tri-colour): CD3 (69.4 vs 68.3) CD4 (24.0 vs 24.2) and CD19 (13.9 vs 12.6). A comparison of the results obtained using the paired t test showed that there were no significant differences for cells expressing CD3, CD4 and CD19. However, there was a significant difference in the NK (18.3 vs 16.3) cell component. A major advantage in using 3 colour immunophenotyping is the ability to analyse specimens that cannot be analysed using dual colour reagents due to debris or contamination of the gate with non-lymphocytic cells.

A novel fuzzy neural network, called the pseudo outer-product based fuzzy neural network (POPFNN), is proposed in this paper. The functions performed by each layer in the proposed POPFNN strictly correspond to the inference steps in the truth value restriction method in fuzzy logic [[Mantaras (1990)] Approximate reasoning models, Ellis Horwood]. This correspondence gives it a strong theoretical basis. Similar to most of the existing fuzzy neural networks, the proposed POPFNN uses a self-organizing algorithm ([Kohonen, 1988], Self-organization and associative memories, Springer) to learn and initialize the membership functions of the input and output variables from a set of training data. However, instead of employing the popularly used competitive learning [[Kosko (1990)] IEEE Trans. Neural Networks, 3(5), 801], this paper proposes a novel pseudo outer-product (POP) learning algorithm to identify the fuzzy rules that are supported by the training data. The proposed POP learning algorithm is fast, reliable, and highly intuitive. Extensive experimental results and comparisons are presented at the end of the paper for discussion. Copyright 1996 Elsevier Science Ltd.

A cross-sectional study on the expression of 6 lymphocyte markers was carried out on 481 patients with human immunodeficiency virus (HIV) and 79 normals after stratification based on absolute CD4 counts. The data were stratified according to the following groups: (I) 1201 to 1600, (II) 801 to 1200, (III) 401 to 800 and (IV) 0 to 400 (x 10(6) CD4 cells per mm3). The mean percentages of the subsets before stratification showed that HIV patients had increased percentages of CD3+ (75.7 against 66.9), CD3+CD8+ (52.2 against 32.3) and CD3+HLA-DR+ (36.1 against 14.4) cells and lower percentages of CD19 (10.3 against 13.3) and natural killer cells (13.7 against 20.4) when compared to controls in the same group. A definite trend, however, was only seen in CD3+CD8+ (47.4, 50.0, 54.0, 57.5 for groups I, II, III and IV respectively) and CD3+HLA-DR+ (29.1, 32.9, 38.4, 43.9 for groups I, II, III and IV respectively).

The aim of this study was to establish the lymphocyte subset reference ranges in a defined Malaysian population as well as to determine inter-racial differences for these values. Normal blood obtained from 152 subjects (55.9% Malay, 26.3% Chinese and 17.7% Indian) was immunophenotyped. Results obtained (expressed as mean +/- SD %), absolute count (x 10(6) cells/mm3) were as follows: CD3:66.5 +/- 8.6%, 2,066; CD4:33.2 +/- 8.5%, 1,028; CD831.6 +/- 8.9%, 982; CD19:12.0 +/- 0%, 5,374, and CD56+CD16:20.9 +/- 9%, 1,638. There were no significant differences between the percent lymphocyte subsets of the three racial groups. However, the absolute number of CD4 cells and CD19 cells in Chinese was significantly lower (p < 0.05) compared to the Indian and the Indian and Malay groups respectively. Comparison of our results with other reports showed that the percentage of Natural Killer cells in this population is higher than that reported for Caucasian population.

House dust mite allergens constitute one of the most important allergens in house dust. In this study, the levels of two common dust mite allergens, Der p I and Der f I, in a general hospital in Singapore were evaluated. Our results showed that these allergens were detected in 42/74 (or 57%) of the dust samples. Der p I was found to be the predominant allergen detected (p < 0.001). The allergen levels were, however, low with only 1/74 having a Der p I concentration above 2 micrograms g-1 dust. None of the samples had Der f I concentrations above this level. Of the various niches studied (mattresses, pillows, sofas, carpets, blinds and floors), the blinds and floors had the lowest concentration of allergen (p < 0.05). These low levels in the hospital compared to homes were attributed to the vigorous cleaning schedule in the hospital, the use of plastic to encased mattresses and pillows, vinyl covered sofas and vinyl lined floors. These practices may be adopted in the home as a means to reduce mite allergen exposure.

Cross-sectional data of 210 Chinese children aged three to 16 years, of whom 29 (13.8%) were obese, and their biological parents were statistically assessed whether the body mass index (BMI = weight/height2) of the parents were predictive of childhood obesity. Using the logistic regression and proportional hazards models to predict the probability of obesity and to control for potential confounding due to the age and sex of the children, families with one parent having a BMI score of 25 or higher were 3.7 times, and families with two such parents were five times more likely to have an obese child than families with no such parent. Childhood obesity was more markedly correlated with the mother's BMI than with the father's BMI. These findings were all stochastically significant (p < 0.05) inspite of the modest sample size. One in two parents with an obese child did not perceive their child to be overweight and 62% of the parents with an obese child did not want their child to lose weight.