Assessment of respiratory system mechanics by artificial neural networks: an exploratory study.

We evaluated 1) the performance of an artificial neural network (ANN)-based technology in assessing the respiratory system resistance (Rrs) and compliance (Crs) in a porcine model of acute lung injury and 2) the possibility of using, for ANN training, signals coming from an electrical analog (EA) of the lung. Two differently experienced ANNs were compared. One ANN (ANN(BIO)) was trained on tracings recorded at different time points after the administration of oleic acid in 10 anesthetized and paralyzed pigs during constant-flow mechanical ventilation. A second ANN (ANN(MOD)) was trained on EA simulations. Both ANNs were evaluated prospectively on data coming from four different pigs. Linear regression between ANN output and manually computed mechanics showed a regression coefficient (R) of 0.98 for both ANNs in assessing Crs. On Rrs, ANN(BIO) showed a performance expressed by R = 0.40 and ANN(MOD) by R = 0.61. These results suggest that ANNs can learn to assess the respiratory system mechanics during mechanical ventilation but that the assessment of resistance and compliance by ANNs may require different approaches.     

Sequential modelling of a full-scale wastewater treatment plant using an artificial neural network

This work proposes a sequential modelling approach using an artificial neural network (ANN) to develop four independent multivariate models that are able to predict the dynamics of biochemical oxygen demand (BOD), chemical oxygen demand (COD), suspended solid (SS), and total nitrogen (TN) removal in a wastewater treatment plant (WWTP). Suitable structures of ANN models were automatically and conveniently optimized by a genetic algorithm rather than the conventional trial and error method. The sequential modelling approach, which is composed of two parts, a process disturbance estimator and a process behaviour predictor, was also presented to develop multivariate dynamic models. In particular, the process disturbance estimator was first employed to estimate the influent quality. The process behaviour predictor then sequentially predicted the effluent quality based on the estimated influent quality from the process disturbance estimator with other process variables. The efficiencies of the developed ANN models with a sequential modelling approach were demonstrated with a practical application using a data set collected from a full-scale WWTP during 2 years. The results show that the ANN with the sequential modelling approach successfully developed multivariate dynamic models of BOD, COD, SS, and TN removal with satisfactory estimation and prediction capability. Thus, the proposed method could be used as a powerful tool for the prediction of complex and nonlinear WWTP performance.     

Trunk acceleration for neuroprosthetic control of standing: a pilot study

This pilot study investigated the potential of using trunk acceleration feedback control of center of pressure (COP) against postural disturbances with a standing neuroprosthesis following paralysis. Artificial neural networks (ANNs) were trained to use three-dimensional trunk acceleration as input to predict changes in COP for able-bodied subjects undergoing perturbations during bipedal stance. Correlation coefficients between ANN predictions and actual COP ranged from 0.67 to 0.77. An ANN trained across all subject-normalized data was used to drive feedback control of ankle muscle excitation levels for a computer model representing a standing neuroprosthesis user. Feedback control reduced average upper-body loading during perturbation onset and recovery by 42% and peak loading by 29% compared with optimal, constant excitation.     

Prediction of chiral separations using a combination of experimental design and artificial neural networks

In this work the advantages of using artificial neural networks (ANNs) combined with experimental design (ED) to optimize the separation of amino acids enantiomers, with α‐cyclodextrin as chiral selector, were demonstrated. The results obtained with the ED‐ANN approach were compared with those of either the partial least‐squares (PLS) method or the response surface methodology where experimental design and the regression equation were used. The ANN approach is quite general, no explicit model is needed, and the amount of experimental work can be decreased considerably. Chirality 11:616–621, 1999. © 1999 Wiley‐Liss, Inc.     

Kinetic models of odor transduction implemented as artificial neural networks

We present a formal model of olfactory transduction corresponding to the biochemical reaction cascade found in chemosensory neurons. It assumes that odorants bind to receptor proteins which, in turn, activate transducer mechanisms corresponding to second messenger-mediated processes. The model is reformulated as a mathematically equivalent artificial neural network (ANN). To enable comparison of the computational power of our model, previously suggested models of chemosensory transduction are also presented in ANN versions. In ANNs, certain biological parameters, such as rate constants and affinities, are transformed into weights that can be fitted by training with a given experimental data set. After training, these weights do not necessarily equal the real biological parameters, but represent a set of values that is sufficient to simulate an experimental set of data. We used ANNs to simulate data recorded from bee subplacodes and compare the capacity of our model with ANN versions of other models. Receptor neurons of the nonpheromonal, general odor-processing subsystem of the honeybee are broadly tuned, have overlapping response spectra, and show highly nonlinear concentration dependencies and mixture interactions, i.e., synergistic and inhibitory effects. Our full model alone has the necessary complexity to simulate these complex response characteristics. To account for the complex response characteristics of honeybee receptor neurons, we suggest that several different receptor protein types and at least two second messenger systems are necessary that may interact at various levels of the transduction cascade and may eventually have opposing effects on receptor neuron excitability.     

Predictive modelling of Fe(III) precipitation in iron removal process for bioleaching circuits

In this study, the applicability of three modelling approaches was determined in an effort to describe complex relationships between process parameters and to predict the performance of an integrated process, which consisted of a fluidized bed bioreactor for Fe³⁺ regeneration and a gravity settler for precipitative iron removal. Self-organizing maps were used to visually evaluate the associations between variables prior to the comparison of two different modelling methods, the multiple regression modelling and artificial neural network (ANN) modelling, for predicting Fe(III) precipitation. With the ANN model, an excellent match between the predicted and measured data was obtained (R ² = 0.97). The best-fitting regression model also gave a good fit (R ² = 0.87). This study demonstrates that ANNs and regression models are robust tools for predicting iron precipitation in the integrated process and can thus be used in the management of such systems.     

Prediction of Streptococcus salivarius subsp. thermophilus and Lactobacillus delbrueckii subsp. bulgaricus populations in yoghurt by Curie point pyrolysis-mass spectrometry.

We evaluated the potential of pyrolysis-mass spectrometry (PyMS) for quantifying the binary mixed population of Streptococcus salivarius subsp. thermophilus and Lactobacillus delbrueckii subsp. bulgaricus in yoghurt. For this purpose, a new analytical approach was developed. The yoghurt was transparised and its total bacterial population was recovered by centrifugation and estimated by turbidimetric measurement. The quantity of each population (L. bulgaricus, S. thermophilus) was then estimated in the pellet by PyMS, and the data were analysed by artificial neural networks (ANNs). In parallel, streptococci and lactobacilli were numerated on SYL agar and these data were used as reference values to predict the bacterial counts of each population by PyMS. A close correlation was established between the streptococci and the lactobacilli counts on SYL agar and PyMS measurements (r(2)=0.98 for S. thermophilus and r(2)=0.96 for L. bulgaricus). Combined turbidimetric measurement and PyMS/ANNs seemed to be a powerful method for obtaining rapid counts of binary mixtures of bacteria in yoghurt.     

Artificial neural networks for decision-making in urologic oncology

Artificial neural networks (ANNs) are computational methodologies that perform multifactorial analyses, inspired by networks of biological neurons. Like neural networks, ANNs contain layers of simple points (nodes) of data that interract through carefully weighted connection lines. ANNs are "trained" and balanced by having been previously fed data, which the ANN uses as the means for adjusting its interconnections. Studies have shown that novel and highly accurate ANNs significantly enhance the ability to detect prostate cancer early (high sensitivity) while avoiding a greater number of unnecessary tissue samplings (high specificity). The use of ANNs in prostate cancer is ideal because of 1) multiple predicting factors that influence outcome; 2) the desire to offer individual consulting based on various tests; 3) the fact that prior logistic regression analysis results have had serious limitations in application; and 4) the need for an up-to-date tool that can apply easily to everyone. An ANN should be seen as an important tool that is complementary to the physician's personal knowledge and judgment in making decisions.     

String matching artificial neural networks.

Three artificial neural networks (ANNs) are proposed for solving a variety of on- and off-line string matching problems. The ANN structure employed as the building block of these ANNs is derived from the harmony theory (HT) ANN, whereby the resulting string matching ANNs are characterized by fast match-mismatch decisions, low computational complexity, and activation values of the ANN output nodes that can be used as indicators of substitution, insertion (addition) and deletion spelling errors.     

Feasibility of estimating isokinetic knee torque using a neural network model

Many studies have investigated the relationships between electromyography (EMG) and torque production. A few investigators have used adjusted learning algorithms and feed-forward artificial neural networks (ANNs) to estimate joint torque in the elbow. This study sought to estimate net isokinetic knee torque using ANN models. Isokinetic knee extensor and flexor torque data were measured simultaneously with agonist and antagonist EMG during concentric and eccentric contractions at joint velocities of 30 degrees /s and 60 degrees /s. Age, gender, height, body mass, agonist EMG, antagonist EMG, joint position and joint velocity were entered as predictive variables of net torque. A three-layer ANN model was developed and trained using an adjusted back-propagation algorithm. Accuracy results were compared against those of forward stepwise regression models. Stepwise regression models included body mass, body height and joint position as the most influential predictors, followed by agonist EMG for concentric and eccentric contractions. Estimation of eccentric torque included antagonist EMG following the agonist activation. ANN models resulted in more accurate torque estimation (R=0.96), compared to the stepwise regression models (R=0.71). ANN model accuracy increased greatly when the number of hidden units increased from 5 to 10, continuing to increase gradually with additional hidden units. The average number of training epochs necessary for solution convergence and the relative accuracy of the model indicate a strong ability for the ANN model to generalize these estimations to a broader sample. The ANN model appears to be a feasible technique for estimating joint torque in the knee.     

Improved Glomerular Filtration Rate Estimation by an Artificial Neural Network

Background

Accurate evaluation of glomerular filtration rates (GFRs) is of critical importance in clinical practice. A previous study showed that models based on artificial neural networks (ANNs) could achieve a better performance than traditional equations. However, large-sample cross-sectional surveys have not resolved questions about ANN performance.

Methods

A total of 1,180 patients that had chronic kidney disease (CKD) were enrolled in the development data set, the internal validation data set and the external validation data set. Additional 222 patients that were admitted to two independent institutions were externally validated. Several ANNs were constructed and finally a Back Propagation network optimized by a genetic algorithm (GABP network) was chosen as a superior model, which included six input variables; i.e., serum creatinine, serum urea nitrogen, age, height, weight and gender, and estimated GFR as the one output variable. Performance was then compared with the Cockcroft-Gault equation, the MDRD equations and the CKD-EPI equation.

Results

In the external validation data set, Bland-Altman analysis demonstrated that the precision of the six-variable GABP network was the highest among all of the estimation models; i.e., 46.7 ml/min/1.73 m² vs. a range from 71.3 to 101.7 ml/min/1.73 m², allowing improvement in accuracy (15% accuracy, 49.0%; 30% accuracy, 75.1%; 50% accuracy, 90.5% [P<0.001 for all]) and CKD stage classification (misclassification rate of CKD stage, 32.4% vs. a range from 47.3% to 53.3% [P<0.001 for all]). Furthermore, in the additional external validation data set, precision and accuracy were improved by the six-variable GABP network.

Conclusions

A new ANN model (the six-variable GABP network) for CKD patients was developed that could provide a simple, more accurate and reliable means for the estimation of GFR and stage of CKD than traditional equations. Further validations are needed to assess the ability of the ANN model in diverse populations.     

Artificial neural network approaches to one-step weekly prediction of Dinophysis acuminata blooms in Huelva (Western Andalucía, Spain)

On the Atlantic coasts of Andalucía, chronic spring–summer (March–June) diarrhetic shellfish poisoning (DSP) outbreaks are associated with blooms of Dinophysis acuminata, Claparède and Lachmann. Artificial neural networks (ANNs) have been successfully used to model primary production and have recently been tested for the prediction of harmful algae blooms. In this study, we evaluated the performance of feed forward ANN models trained to predict D. acuminata blooms. ANN models were trained and tested using weekly data (5 previous weeks) of D. acuminata cell counts from eight stations of the Andalucía HAB monitoring programme in the coasts of Huelva between 1998 and 2004. Principal component analysis (PCA) were previously carried out to find out possible similarities within time series from each zone with the aim of reducing the number of areas to model. Our results show that ANN models with a low number of input variables are able to reproduce trends in D. acuminata population dynamics.     

Neural network modeling of the light profile in a novel photobioreactor

An artificial neural network (ANN) was implemented to model the light profile pattern inside a photobioreactor (PBR) that uses a toroidal light arrangement. The PBR uses Tequila vinasses as culture medium and purple non-sulfur bacteria Rhodopseudomonas palustris as biocatalyzer. The performance of the ANN was tested for a number of conditions and compared to those obtained by using deterministic models. Both ANN and deterministic models were validated experimentally. In all cases, at low biomass concentration, model predictions yielded determination coefficients greater than 0.9. Nevertheless, ANN yielded the more accurate predictions of the light pattern, at both low and high biomass concentration, when the bioreactor radius, the depth, the rotational speed of the stirrer and the biomass concentration were incorporated in the ANN structure. In comparison, most of the deterministic models failed to correlate the empirical data at high biomass concentration. These results show the usefulness of ANNs in the modeling of the light profile pattern in photobioreactors.     

Dynamic muscle force predictions from EMG: an artificial neural network approach

EMG signals of dynamically contracting muscle have never been used to predict experimentally known muscle forces across subjects. Here, we use an artificial neural network (ANN) approach to first derive an EMG–force relationship from a subset of experimentally determined EMGs and muscle forces; second, we use this relationship to predict individual muscle forces for different contractile conditions and in subjects whose EMG and force data were not used in the derivation of the EMG–force relationship; and third, we validate the predicted muscle forces against the known forces recorded in vivo. EMG and muscle forces were recorded from the cat soleus for a variety of locomotor conditions giving a data base from three subjects, four locomotor conditions, and 8–16 steps per subject and condition. Considering the conceptual differences in the tasks investigated (e.g. slow walking vs. trotting), the intra-subject results obtained here are superior to those published previously, even though the approach did not require a muscle model or the instantaneous contractile conditions as input for the force predictions. The inter-subject results are the first of this kind to be presented in the literature and they typically gave cross-correlation coefficients between actual and predicted forces of >0.90 and root mean square errors of <15%, thus they were considered excellent.

From the results of this study, it was concluded that ANNs represent a powerful tool to capture the essential features of EMG–force relationships of dynamically contracting muscle, and that ANNs might be used widely to predict muscle forces based on EMG signals.     

Prediction of inhibitor binding free energies by quantum neural networks. Nucleoside analogues binding to trypanosomal nucleoside hydrolase

A computational method has been developed to predict inhibitor binding energy for untested inhibitor molecules. A neural network is trained from the electrostatic potential surfaces of known inhibitors and their binding energies. The algorithm is then able to predict, with high accuracy, the binding energy of unknown inhibitors. IU-nucleoside hydrolase from Crithidia fasciculata and the inhibitor molecules described previously [Miles, R. W. Tyler, P. C. Evans, G. Furneaux R. H., Parkin, D. W., and Schramm, V. L. (1999) Biochemistry 38, xxxx-xxxx] are used as the test system. Discrete points on the molecular electrostatic potential surface of inhibitor molecules are input to neural networks to identify the quantum mechanical features that contribute to binding. Feed-forward neural networks with back-propagation of error are trained to recognize the quantum mechanical electrostatic potential and geometry at the entire van der Waals surface of a group of training molecules and to predict the strength of interactions between the enzyme and novel inhibitors. The binding energies of unknown inhibitors were predicted, followed by experimental determination of K(i)() values. Predictions of K(i)() values using this theory are compared to other methods and are more robust in estimating inhibitory strength. The average deviation in estimating K(i)() values for 18 unknown inhibitor molecules, with 21 training molecules, is a factor of 5 x K(i)() over a range of 660 000 in K(i)() values for all molecules. The a posteriori accuracy of the predictions suggests the method will be effective as a guide for experimental inhibitor design.     

Real-Time Decision Fusion for Multimodal Neural Prosthetic Devices

Background

The field of neural prosthetics aims to develop prosthetic limbs with a brain-computer interface (BCI) through which neural activity is decoded into movements. A natural extension of current research is the incorporation of neural activity from multiple modalities to more accurately estimate the user''s intent. The challenge remains how to appropriately combine this information in real-time for a neural prosthetic device.

Methodology/Principal Findings

Here we propose a framework based on decision fusion, i.e., fusing predictions from several single-modality decoders to produce a more accurate device state estimate. We examine two algorithms for continuous variable decision fusion: the Kalman filter and artificial neural networks (ANNs). Using simulated cortical neural spike signals, we implemented several successful individual neural decoding algorithms, and tested the capabilities of each fusion method in the context of decoding 2-dimensional endpoint trajectories of a neural prosthetic arm. Extensively testing these methods on random trajectories, we find that on average both the Kalman filter and ANNs successfully fuse the individual decoder estimates to produce more accurate predictions.

Conclusions

Our results reveal that a fusion-based approach has the potential to improve prediction accuracy over individual decoders of varying quality, and we hope that this work will encourage multimodal neural prosthetics experiments in the future.     

Classification des stades de sommeil par des réseaux de neurones artificiels hiérarchiques

The goal of our work is to provide an automatic analysis and decision tool for sleep stages classification based on an artificial neural networks (ANN). The first difficulty lies in choosing the physiological signals representation and in particular the electroencephalogram (EEG). Once the representation adopted, the next step is to design the optimal neural network determined by a learning and validation process of data from a set of sleep records. We studied several configurations of conventional ANN giving results varying from 62 to 71 %, then we proposed a new hierarchical configuration, which gives a rate of 74 % correct classification for six stages. These results lead us to further explore this issue at the representation and design of ANNs to improve the performance of our tool.     

Evaluation of soil aggregate stability in Algerian northwestern soils using pedotransfer functions and artificial neural networks

Since, aggregate stability is the main physical property regulating erodibility; its observations can act as a useful indicator for monitoring and managing soil degradation. In this context, this study carried out in the alluvial plain of Cheliff, a semi-arid area aimed to predict aggregate stability through Mean Weight Diameter (MWD), using pedotransfer functions (PTFs) with different stratifications (textural, salinity and organic-textural) and artificial neural networks (ANNs). Results showed that the best MWD predictions were those related to organic-textural PTFs, in this stratification the silty-clay moderately rich OM class showed the highest significant determination coefficient R² (0.65) and the lowest mean square error (0.03), whereas, the textural and salinity PTFs were a very weak predictors with a very low R². It was also found that the performances of ANNs in predicting MWD were better than those of PTFs, regarding ANNs input variables the best predictions were those obtained with a large number of input variables, furthermore, by using a large number of hidden neurons, the performances of Radial Basis Function (RBF) were better than those of Multilayer Perceptron (MLP). It was also noted that the best RBF results were always related to the Gaussian hidden activation, whereas, MLP was not related to a specific hidden activation.