On the Selection of Non-Invasive Methods Based on Speech Analysis Oriented to Automatic Alzheimer Disease Diagnosis

The work presented here is part of a larger study to identify novel technologies and biomarkers for early Alzheimer disease (AD) detection and it focuses on evaluating the suitability of a new approach for early AD diagnosis by non-invasive methods. The purpose is to examine in a pilot study the potential of applying intelligent algorithms to speech features obtained from suspected patients in order to contribute to the improvement of diagnosis of AD and its degree of severity. In this sense, Artificial Neural Networks (ANN) have been used for the automatic classification of the two classes (AD and control subjects). Two human issues have been analyzed for feature selection: Spontaneous Speech and Emotional Response. Not only linear features but also non-linear ones, such as Fractal Dimension, have been explored. The approach is non invasive, low cost and without any side effects. Obtained experimental results were very satisfactory and promising for early diagnosis and classification of AD patients.

Object oriented simulation of hybrid renewable energy systems focused on supervisor control

EFTA 2009

Hybrid Qm/Mm Simulation Of The Hydration Phenomena Of Dipalmitoylphosphatidylcholine Headgroup

The polar headgroup of dipalmitoylphosphatidylcholine (DPPC) molecule both in gas phase and aqueous Solution is investigated by the hybrid quantum mechanical/molecular mechanical (QM/MM) method, in which the polar head of DPPC molecule and the bound water molecules are treated with density functional theory (DFT), while the apolar hydrocarbon chain of DPPC molecule is treated with MM method. It is demonstrated that the hybrid QM/MM method is both accurate and efficient to describe the conformations of DPPC headgroup. Folded structures of headgroup are found in gas phase calculations. In this work, both monohydration and polyhydration phenomena are investigated. In monohydration, different water association sites are studied. Both the hydration energy and the quantum properties of DPPC and water molecules are calculated at the DFT level of theory after geometry optimization. The binding force of monohydration is estimated by using the scan method. In polyhydration, more extended conformations are found and hydration energies in different polyhydration styles are estimated. (C) 2008 Elsevier Inc. All rights reserved.

An Atomistic-Continuum Hybrid Simulation Of Fluid Flows Over Superhydrophobic Surfaces

Recent experiments have found that slip length could be as large as on the order of 1 mu m for fluid flows over superhydrophobic surfaces. Superhydrophobic surfaces can be achieved by patterning roughness on hydrophobic surfaces. In the present paper, an atomistic-continuum hybrid approach is developed to simulate the Couette flows over superhydrophobic surfaces, in which a molecular dynamics simulation is used in a small region near the superhydrophobic surface where the continuum assumption is not valid and the Navier-Stokes equations are used in a large region for bulk flows where the continuum assumption does hold. These two descriptions are coupled using the dynamic coupling model in the overlap region to ensure momentum continuity. The hybrid simulation predicts a superhydrophobic state with large slip lengths, which cannot be obtained by molecular dynamics simulation alone.

Turbulent collision of inertial particles: point-particle based, hybrid simulations and beyond

Point-particle based direct numerical simulation (PPDNS) has been a productive research tool for studying both single-particle and particle-pair statistics of inertial particles suspended in a turbulent carrier flow. Here we focus on its use in addressing particle-pair statistics relevant to the quantification of turbulent collision rate of inertial particles. PPDNS is particularly useful as the interaction of particles with small-scale (dissipative) turbulent motion of the carrier flow is mostly relevant. Furthermore, since the particle size may be much smaller than the Kolmogorov length of the background fluid turbulence, a large number of particles are needed to accumulate meaningful pair statistics. Starting from the relative simple Lagrangian tracking of so-called ghost particles, PPDNS has significantly advanced our theoretical understanding of the kinematic formulation of the turbulent geometric collision kernel by providing essential data on dynamic collision kernel, radial relative velocity, and radial distribution function. A recent extension of PPDNS is a hybrid direct numerical simulation (HDNS) approach in which the effect of local hydrodynamic interactions of particles is considered, allowing quantitative assessment of the enhancement of collision efficiency by fluid turbulence. Limitations and open issues in PPDNS and HDNS are discussed. Finally, on-going studies of turbulent collision of inertial particles using large-eddy simulations and particle- resolved simulations are briefly discussed.