Blind signal separation: statistical principles

Blind signal separation (BSS) and independent component analysis (ICA) are emerging techniques of array processing and data analysis that aim to recover unobserved signals or “sources” from observed mixtures (typically, the output of an array of sensors), exploiting only the assumption of mutual independence between the signals. The weakness of the assumptions makes it a powerful approach, but it requires us to venture beyond familiar second order statistics, The objectives of this paper are to review some of the approaches that have been developed to address this problem, to illustrate how they stem from basic principles, and to show how they relate to each other     

Datapath BIST Insertion Using Pre-Characterized Area and Testability Data

There are several ways to insert Built-in Self-Test (BIST) circuitry on a circuit, each of them with particular consequences on area overhead, test application time and fault coverage. This paper presents a BIST insertion methodology applied to datapaths described at the RTL level that uses a database containing: (a) testability data of several types of test pattern generators (TPGs) and signature analyzers (SAs) when connected to several types of functional units and (b) area overhead due to the implementation by a datapath register of each type of those test resources. The availability of this database makes then possible to choose the best test resource types associated to each functional unit in a datapath, leading to good testability and area results.     

The immobilisation of enzymes on titanium-activated inorganic supports

Methods are described for the immobilisation of enzymes on inorganic supports activated by the surface deposition of an imperfectly crystallised film of TiO₂. The stability of the preparations is shown to be strongly dependent on the integrity of the support-TiO₂ link which is in turn dependent on the pretreatment of the solid support. The enzyme may be linked to the activated support using 5-aminosalicylic acid (with or without diazotisation), tannic acid or stannous chloride, or by direct contact. These preparations are shown to have activities and operational stabilities superior to those of a preparation made using the silane-glutaraldehyde method on the same glass support.     

Optimal cross-section and configuration design of elastic-plastic structures subject to dynamic cyclic loading

This paper shows an optimal design problem with continuum variational formulation, applied to nonlinear elasticplastic structures subject to dynamic loading. The total Lagrangian procedure is used to describe the response of the structure. The direct differentiation method is used to obtain the sensitivities of the structural response that are needed to solve the optimization problem. Since unloading and reloading of the structure are allowed, the structural response is path-dependent and an additional step is needed to integrate the constitutive equations. It can be shown, consequently, that design sensitivity analysis is also path-dependent. A finite element method with implicit time integration is used to discretize the state and sensitivity equations.A mathematical programming approach is used for the optimization process. Numerical applications are performed on a 3-D truss structure, where cross-sectional areas and nodal point coordinates are treated as design variables. Optimal designs have been obtained and compared by using two different strategies: a twolevel strategy where the levels are defined according to the type of design variables, cross sectional areas or node coordinates, and optimizing simultaneously with respect to both types of design variables. Comparisons have also been made between optimal designs obtained by considering or not considering the inertial term of the structural equilibrium.     

Diastolic dysfunction in patients with chronic kidney failure on a hemodialysis program

OBJECTIVE: The aim of this study was to analyse different ultrasound parameters for the assessment of isolated left ventricular diastolic dysfunction (LVDD) in patients with chronic renal failure (CRF) on periodic hemodialysis (HD), comparing pulsed wave Doppler with pulsed tissue Doppler. MATERIALS AND METHODS: Forty-seven patients with CRF on HD (61% were male; mean age was 51.0 +/- 16.5 years, mean HD time--3.7 +/- 3.8 years, 38% had hypertension, 17% had diabetes) were studied by echocardiography (bidimensional, M-Mode, flow pulsed Doppler and tissue Doppler imaging). All patients had symptoms of left heart failure-class II NYHA, were in sinus rhythm and had no symptoms of ischemic heart disease. The presence of abnormal LV regional contractility was the exclusion criteria. According to their mitral inflow profile Doppler characteristics, patients were included in two groups: Group A (E/A > 1; n = 21) and B (E/A < 1; n = 26). We compared: LV dimensions and function, left atrial (LA) dimension. Gaasch index, LV mass index. E and A wave velocities (in flow pulsatile Doppler and tissue Doppler). E/N ratio in tissue Doppler, isovolumetric relaxation time (IVRT) and deceleration time (DT). RESULTS: There were no significant differences in the prevalence of age > or = 65 years male sex, hypertension or diabetes between group A and B patients, and almost all patients were on hemodialytic treatment for more than one year (81% vs 85%: NS). LV hypertrophy was present in almost all group A and B patients (A--95% vs B--85.5%; NS). Group A, compared with group B, had a difference in the Gaasch index (2.45 +/- 0.3 vs 2.08 +/- 0.4; p < 0.05), E wave velocity in flow pulsatile Doppler and tissue Doppler (cm/sec) (110 +/- 27 vs 62 +/- 20; p < 0.001 and 41 +/- 15 vs 28.5 +/- 16; p < 0.05), E/A ratio in tissue Doppler (1.3 +/- 0.4 vs 0.8 +/- 0.3; p < 0.001). IVRT (msec) (80.7 +/- 15.2 vs 113.5 +/- 28.3; p < 0.001) and DT (msec) (189.7 +/- 24 vs 278.2 +/- 17.9; p < 0.001). According to the E'/A' ratio in tissue Doppler, group A patients were divided in another two groups: E'/A' > 1 (13/21--62%) and < 1 (8/21--38%) and a significantly longer IVRT (75.8 +/- 9.3 vs 100.9 +/- 3.2; p < 0.001) and DT (178 +/- 15 vs 240 +/- 20; p < 0.001) and a greater LA dimension (37.6 +/- 6.9 vs 44.6 +/- 6.9; p < 0.05) were found. CONCLUSIONS: Pulsed wave Doppler is the most useful non invasive method for assessment of global diastolic dysfunction. In our study, 17% of the patients had E/A < 1 only in the tissue Doppler study. These patients probably had a pseudonormal mitral pattern.     

Maximal functional capacity in patients with endomyocardial fibrosis

We evaluated objectively the cardiocirculatory dysfunction of patients with EMF, and related the functional class (FC-NYHA) to the data obtained by analysis of maximal functional capacity. We studied 55 subjects, divided into three groups: group 1 (G1) 21 (38.1%) normal persons, group 2 (G2) seven (12.7%) patients with EMF who were in FC I or II, and group 3 (G3) 27 (49%) patients in FC III or IV. Maximal oxygen consumption (VO2max) and O2 pulse index (PO2max) were obtained using a Beckman computerized gas analyser. The mean values were significantly different among the three groups. We conclude that the greater the clinical impairment, the greater the alterations that occur in the indices of cardiac function.     

Comparison between the use of unifoil double balloon and bifoil balloon in successful mitral valvuloplasty by balloon catheter

PURPOSE: To compare immediate and long term results balloon mitral valvuloplasty (BMV) using double balloon or bifoil balloon. METHODS: One hundred and thirteen consecutive cases of BMV used aleatory double balloon (group DB--55 cases) or bifoil balloon (group BF--16 cases). Patients were similar regarding to age, sex, valvopaty etiology, functional class and echocardiographic score. Seventy one (63%) patients achieved 12 months follow-up. RESULTS: In group DB there were 2 (4%) insuccess, 2 (4%) cardiac tamponade and 2 (4%) deaths, 91% patients had immediate criteria of success. Mitral valve area (MVA) increased from 0.8 to 1.69cm2 and mitral gradient (G) by echodopplercardiographic (ECHO) decreased from 17.9 to 4.8mmHg. Three (5%) patients developed severe mitral regurgitation (MR) and needed surgical intervention. At follow-up 2 (4%) developed mitral restenosis. MVA estimated by ECHO study after one year follow-up was inferior to 1.15cm2 in 15 (32%) cases, between 1.5 and 2.0cm2 in 17 (37%) and superior to 2.0cm2 in 14 (31%). In group BF there were 2 (12.5%) insuccess, 4 (25%) developed severe MR occurring 1 death immediate post-operative mitral valve replacement. Among 14 (87%) success cases, MVA increased from 0.8 to 1.89cm2 and G decreased from 18 to 6.4mmHg. Lately 2 (12.5%) needed surgical intervention because significative MR. At 12th month follow-up the ECHO study showed that in one (10%) case MVA was < 1.5cm2, and in 3 (27%) cases was between 1.5 and 2.0cm2. CONCLUSION: There were similar good results in both groups, however group DB had more restenosis, cardiac tamponade and vascular complications and group BF had more severe MR.     

Equivariant adaptive source separation

Source separation consists of recovering a set of independent signals when only mixtures with unknown coefficients are observed. This paper introduces a class of adaptive algorithms for source separation that implements an adaptive version of equivariant estimation and is henceforth called equivariant adaptive separation via independence (EASI). The EASI algorithms are based on the idea of serial updating. This specific form of matrix updates systematically yields algorithms with a simple structure for both real and complex mixtures. Most importantly, the performance of an EASI algorithm does not depend on the mixing matrix. In particular, convergence rates, stability conditions, and interference rejection levels depend only on the (normalized) distributions of the source signals. Closed-form expressions of these quantities are given via an asymptotic performance analysis. The theme of equivariance is stressed throughout the paper. The source separation problem has an underlying multiplicative structure. The parameter space forms a (matrix) multiplicative group. We explore the (favorable) consequences of this fact on implementation, performance, and optimization of EASI algorithms     

Enhanced in vivo adenovirus-mediated gene transfer to rat hepatocarcinomas by selective administration into the hepatic artery

Adenovirus-mediated gene therapy of experimental hepatocarcinoma is hindered by low transduction efficacy in vivo. We evaluated the extent of gene expression following various routes of administration of recombinant adenovirus AdCMVlacZ in diethylnitrosamine-induced rat hepatocarcinoma. We first characterized the vascularization of diethylnitrosamine-induced hepatocarcinomas using a computerized tomography scanner approach. The efficacy of gene transfer was then evaluated by three routes of administration: intraportal, selective injection through the hepatic artery and direct injection into the tumor. Diethylnitrosamine-induced hepatocarcinomas had predominantly an arterial blood supply, 67% of the total liver blood supply. Compared with intraportal administration, arterial injection improved gene transfer into tumors whereas that to the non-tumor areas was diminished. In addition, this route of injection allowed the efficient transduction of dysplastic nodules. Diethylnitrosamine-induced hepatocarcinoma in rats is a relevant model for the study of human hepatocarcinoma due to its vascularization. Arterial infusion improved the ratio of transduced tumorous to nontumorous cells and allowed targeting of gene transfer to dysplastic nodules. This will be useful in the design of gene therapy for hepatocarcinoma.     

Application of a model using the phenomenological approach for prediction of growth and xanthan gum production with sugar cane broth in a batch process

Production of xanthan gum was studied and modelled using unstructured kinetic models composed of three differential equations, which considered the microbial biomass, carbon source, and xanthan concentration. The fermentation process, using Xanthomonas campestris pv. campestris NRRL B-1459, was conducted under controlled conditions with diluted sugar cane broth at different initial sucrose concentrations (15.0, 25.0, and 35.0 g L⁻¹). Unstructured kinetic models proposed in the literature for this system were reviewed and applied. These models were tested against the experimental results, calculating the parameters by nonlinear regression. The kinetic models used in this study provided estimations of microbial growth, substrate consumption, and product formation, and, therefore, these parameters were quantified in the fermentation experiments. Higher yield of xanthan per amount of sucrose (0.58 g g⁻¹) and productivity (0.63 g L⁻¹ h⁻¹) were obtained using initial sucrose concentrations of 25.0 and 35.0 g L⁻¹, respectively. The models were used to predict the kinetic parameters for a medium containing an intermediate and a larger initial sucrose concentration (27.0 and 40.0 g L⁻¹). When tested experimentally, the measured fermentation parameters were in close agreement with the values predicted by the model that presented the best adjustment, demonstrating its validity.