Artificial neural network analysis of common femoral artery Doppler shift signals: classification of proximal disease

Narrowing of the middle cerebral artery: artificial intelligence methods and comparison of transcranial color coded duplex sonography with conventional TCD

The goal of the study was to compare performances of transcranial color-coded duplex sonography (TCCS) and transcranial Doppler sonography (TCD) in the diagnosis of the middle cerebral artery (MCA) narrowing in the same population of patients using statistical and nonstatistical intelligent models for data analysis. We prospectively collected data from 179 consecutive routine digital subtraction angiography (DSA) procedures performed in 111 patients (mean age 54.17+/-14.4 years; 59 women, 52 men) who underwent TCD and TCCS examinations simultaneously. Each patient was examined independently using both ultrasound techniques, 267 M1 segments of MCA were assessed and narrowings were classified as < or =50% and >50% lumen reduction. Diagnostic performance was estimated by two statistical and two artificial neural networks (ANN) classification methods. Separate models were constructed for the TCD and TCCS sonographic data, as well as for detection of "any narrowing" and "severe narrowing" of the MCA. Input for each classifier consisted of the peak-systolic, mean and end-diastolic velocities measured with each sonographic method; the output was MCA narrowing. Arterial narrowings less or equal 50% of lumen reduction were found in 55 and >50% narrowings in 26 out of 267 arteries, as indicated by DSA. In the category of "any narrowing" the rate of correct assignment by all models was 82% to 83% for TCCS and 79% to 81% for TCD. In the diagnosis of >50% narrowing the overall classification accuracy remained in the range of 89% to 90% for TCCS data and 90% to 91% for TCD data. For the diagnosis of any narrowing, the sensitivity of the TCCS was significantly higher than that of the TCD, while for diagnosis of >50% MCA narrowing, sensitivity of the TCCS was similar to sensitivity of the TCD. Our study showed that TCCS outperforms conventional TCD in detection of < or =50% MCA narrowing, whereas no significant difference in accuracy between both methods was found in the diagnosis of >50% MCA narrowing. (E-mail: jaroslaw.krejza@uphs.upenn.edu).

Prediction of aortic diameter values in healthy Turkish infants, children, and adolescents by using artificial neural network

The aorta is the largest vessel in the systemic circuit. Its diameter is very important to guess for child before adult age, due to growing up body. Aortic diameter, one of the cardiac values, changes in time. Evaluation of the cardiac structures and generating a valid regional curve requires a large study group experience for accurate data on normal values. In this study, our aim is to estimate aortic diameter values without curve of charts. Using real sample of the all groups has been predicted using a hybrid system based on combination of Line Based Normalization Method (LBNM) and Artificial Neural Network (ANN) with Levenberg-Marquardt (LM) algorithm. In this study, aortic diameter values dataset divided into two groups as 50% training-50% testing of whole dataset. In order to show the performance of the proposed method, two fold cross validation and prevalent performance measuring methods, Mean Square Error (MSE), Absolute Deviation (AD), Root Mean Square Error (RMSE), statistical relation factor T and R2, have been used. The obtained MSE results from combination of Min-Max normalization and ANN, combination of Decimal Scaling and ANN, combination of Z-score and ANN, and combination of LBNM and ANN (the proposed method) are 0.00517, 0.001299, 0.006196, and 0.000145, respectively. For the suggested method, error's results have been given discretely for every age up to adult age. The results are compared to real aortic diameter values by expert with nine year experiences in medical area. These results have shown that the proposed method can be confidently used in the prediction of aortic diameter values in healthy Turkish infants, children and adolescents.

Comparison of short-time Fourier transform and Eigenvector MUSIC methods using discrete wavelet transform for diagnosis of atherosclerosis

In this paper, a more effective use of Doppler techniques is presented for the purpose of diagnosing atherosclerosis in its early stages using the carotid artery Doppler signals. The power spectral density (PSD) graphics are obtained by applying the short-time Fourier transform (STFT)-Welch and the Eigenvector MUSIC methods to the discrete wavelet transform (DWT) of Doppler signals. The PSDs for the fourth approximation component (A4) of both methods estimated that the patients with atherosclerosis in its early phase had lower maximum frequency components. On the other hand, the healthy subjects had higher maximum frequency components. The area under the curve (AUC), which belongs to the receiver operating characteristic (ROC) curve for the frequency level of the maximum PSDs of the A4 approximation obtained from the STFT modeling, is computed as 0.97. The AUC for the MUSIC modeling is computed as 0.996. The AUC belonging to the ROC curve for the higher maximum frequency component is computed as 0.87. The AUC belonging to the ROC curve for the test parameter of the frequency level of the maximum PSDs derived from the MUSIC modeling is determined to be 0.882. The results of this study clearly demonstrate that it is possible to distinguish between the healthy people and the patients with atherosclerosis by using the frequency level of the maximum PSDs for the A4 approximation. Furthermore, it is concluded that the power of Eigenvector-MUSIC method in terms of the resolution of the high frequencies is better than that of the STFT methods.

Utilization of Discretization method on the diagnosis of optic nerve disease

The optic nerve disease is an important disease that appears commonly in public. In this paper, we propose a hybrid diagnostic system based on discretization (quantization) method and classification algorithms including C4.5 decision tree classifier, artificial neural network (ANN), and least square support vector machine (LSSVM) to diagnose the optic nerve disease from Visual Evoked Potential (VEP) signals with discrete values. The aim of this paper is to investigate the effect of Discretization method on the classification of optic nerve disease. Since the VEP signals are non-linearly-separable, low classification accuracy can be obtained by classifier algorithms. In order to overcome this problem, we have used the Discretization method as data pre-processing. The proposed method consists of two phases: (i) quantization of VEP signals using Discretization method, and (ii) diagnosis of discretized VEP signals using classification algorithms including C4.5 decision tree classifier, ANN, and LSSVM. The classification accuracies obtained by these hybrid methods (combination of C4.5 decision tree classifier-quantization method, combination of ANN-quantization method, and combination of LSSVM-quantization method) with and without quantization strategy are 84.6-96.92%, 94.20-96.76%, and 73.44-100%, respectively. As can be seen from these results, the best model used to classify the optic nerve disease from VEP signals is obtained for the combination of LSSVM classifier and quantization strategy. The obtained results denote that the proposed method can make an effective interpretation and point out the ability of design of a new intelligent assistance diagnosis system.

Comparison of MLP neural network and neuro-fuzzy system in transcranial Doppler signals recorded from the cerebral vessels

Transcranial Doppler signals recorded from cerebral vessels of 110 patients were transferred to a personal computer by using a 16 bit sound card. Spectral analyses of Transcranial Doppler signals were performed for determining the Multi Layer Perceptron (MLP) neural network and neuro Ankara-fuzzy system inputs. In order to do a good interpretation and rapid diagnosis, FFT parameters of Transcranial Doppler signals classified using MLP neural network and neuro-fuzzy system. Our findings demonstrated that 92% correct classification rate was obtained from MLP neural network, and 86% correct classification rate was obtained from neuro-fuzzy system.

Usage of eigenvector methods to improve reliable classifier for Doppler ultrasound signals

A new approach based on the implementation of the automated diagnostic systems for Doppler ultrasound signals classification with the features extracted by eigenvector methods is presented. In practical applications of pattern recognition, there are often diverse features extracted from raw data which needs recognizing. Because of the importance of making the right decision, the present work is carried out for searching better classification procedures for the Doppler ultrasound signals. Decision making was performed in two stages: feature extraction by the eigenvector methods and classification using the classifiers trained on the extracted features. The aim of the study is classification of the Doppler ultrasound signals by the combination of eigenvector methods and the classifiers. The present research demonstrated that the power levels of the power spectral density (PSD) estimates obtained by the eigenvector methods are the features which well represent the Doppler ultrasound signals and the probabilistic neural networks (PNNs), recurrent neural networks (RNNs) trained on these features achieved high classification accuracies.

Neural network-based diagnosing for optic nerve disease from visual-evoked potential

In this paper, we purpose a diagnostic procedure to identify the optic nerve disease from visual evoked potential (VEP) signals using an Artificial Neural Network (ANN). Multilayer feed forward ANN trained with a Levenberg Marquart backpropagation algorithm was implemented. The correct classification rate was 96.87% for subjects having optic nerve disease and 96.66% for healthy subjects. The end results are classified as healthy and diseased. Testing results were found to be compliant with the expected results that are derived from the physician's direct diagnosis, angiography, VEP and pattern electroretinography. The stated results show that the proposed method could point out the ability of design of a new intelligent assistance diagnosis system.

Probabilistic neural networks employing Lyapunov exponents for analysis of Doppler ultrasound signals

The implementation of probabilistic neural networks (PNNs) with the Lyapunov exponents for Doppler ultrasound signals classification is presented. This study is directly based on the consideration that Doppler ultrasound signals are chaotic signals. This consideration was tested successfully using the nonlinear dynamics tools, like the computation of Lyapunov exponents. Decision making was performed in two stages: computation of Lyapunov exponents as representative features of the Doppler ultrasound signals and classification using the PNNs trained on the extracted features. The present research demonstrated that the Lyapunov exponents are the features which well represent the Doppler ultrasound signals and the PNNs trained on these features achieved high classification accuracies.

Combining eigenvector methods and support vector machines for detecting variability of Doppler ultrasound signals

In this paper, the multiclass support vector machines (SVMs) with the error correcting output codes (ECOC) were presented for detecting variabilities of the multiclass Doppler ultrasound signals. The ophthalmic arterial (OA) Doppler signals were recorded from healthy subjects, subjects suffering from OA stenosis, subjects suffering from ocular Behcet disease. The internal carotid arterial (ICA) Doppler signals were recorded from healthy subjects, subjects suffering from ICA stenosis, subjects suffering from ICA occlusion. Methods of combining multiple classifiers with diverse features are viewed as a general problem in various application areas of pattern recognition. Because of the importance of making the right decision, better classification procedures for Doppler ultrasound signals are searched. Decision making was performed in two stages: feature extraction by eigenvector methods and classification using the SVMs trained on the extracted features. The research demonstrated that the multiclass SVMs trained on extracted features achieved high accuracy rates.

Training a learning vector quantization network using the pattern electroretinography signals

In this study, the pattern electroretinography (PERG) signals derived from evoked potential across retinal cells of subjects after visual stimulation were analyzed using artificial neural network (ANN) with 172 healthy and 148 diseased subjects. ANN was employed to PERG signals to distinguish between healthy eye and diseased eye. Supervised network examined was a competitive learning vector quantization network. The designed classification structure has about 94% sensitivity, 90.32% specifity, 5.94% false negative, 9.67% false positive and correct classification is calculated to be 92%. Testing results were found to be compliant with the expected results that are derived from the physician's direct diagnosis. The end benefit would be to assist the physician to make the final decision without hesitation.

Automated diagnostic systems with diverse and composite features for Doppler ultrasound signals

In this paper, we present the automated diagnostic systems for Doppler ultrasound signals classification with diverse and composite features and determine their accuracies. We compared the classification accuracies of six different classifiers, namely multilayer perceptron neural network (MLP), combined neural network (CNN), mixture of experts (ME), modified mixture of experts (MME), probabilistic neural network (PNN), and support vector machine (SVM), which were trained on diverse or composite features. The present study was conducted with the purpose of answering the question of whether the automated diagnostic systems improve the capability of classification of ophthalmic arterial (OA) and internal carotid arterial (ICA) Doppler signals. Our research demonstrated that the SVM trained on composite feature and the MME trained on diverse features achieved accuracy rates which were higher than that of the other automated diagnostic systems.

Utilization of artificial neural networks and autoregressive modeling in diagnosing mitral valve stenosis

This research is concentrated on the diagnosis of mitral heart valve stenosis through the analysis of Doppler Signals' AR power spectral density graphic with the help of ANN. Multilayer feedforward ANN trained with a Levenberg Marquart backpropagation algorithm was implemented in the MATLAB environment. Correct classification of 94% was achieved, whereas 4 false classifications have been observed for the test group of 68 subjects in total. The designed classification structure has about 97.3% sensitivity, 90.3% specifity and positive prediction is calculated to be 92.3%. The stated results show that the proposed method can make an effective interpretation.

Utilization of artificial neural networks in the diagnosis of optic nerve diseases

This research is concentrated on the diagnosis of optic nerve disease through the analysis of pattern electroretinography (PERG) signals with the help of artificial neural network (ANN). Multilayer feed forward ANN trained with a Levenberg Marquart (LM) backpropagation algorithm was implemented. The designed classification structure has about 96.4% sensitivity, 90.4% specifity and positive prediction is calculated to be 94.2%. The end results are classified as healthy and diseased. Testing results were found to be compliant with the expected results that are derived from the physician's direct diagnosis. The end benefit would be to assist the physician to make the final decision without hesitation.

Examination of Electric Field Effects on Tissues by Using Back Propagation Neural Network

The aim of this study is to determine lipid peroxidation and antioxidant enzyme levels in spleen and testis tissues of guinea pigs which were exposed to different intensities and periods of DC (direct current) and AC (alternating current) electric fields. The experimental results are applied to neural networks as learning data and the training of the feed forward neural network is realized. At the end of this training; without applying electric field to the tissues, the determination of the effects of the electric field on tissues by using computer is predicted by the neural network. After the experiments, the prediction of the neural network is averagely 99%.

Adaptive neuro-fuzzy inference systems for analysis of internal carotid arterial Doppler signals

In this study, a new approach based on adaptive neuro-fuzzy inference system (ANFIS) was presented for detection of internal carotid artery stenosis and occlusion. The internal carotid arterial Doppler signals were recorded from 130 subjects that 45 of them suffered from internal carotid artery stenosis, 44 of them suffered from internal carotid artery occlusion and the rest of them were healthy subjects. The three ANFIS classifiers were used to detect internal carotid artery conditions (normal, stenosis and occlusion) when two features, resistivity and pulsatility indices, defining changes of internal carotid arterial Doppler waveforms were used as inputs. To improve diagnostic accuracy, the fourth ANFIS classifier (combining ANFIS) was trained using the outputs of the three ANFIS classifiers as input data. The proposed ANFIS model combined the neural network adaptive capabilities and the fuzzy logic qualitative approach. Some conclusions concerning the impacts of features on the detection of internal carotid artery stenosis and occlusion were obtained through analysis of the ANFIS. The performance of the ANFIS model was evaluated in terms of classification accuracies and the results confirmed that the proposed ANFIS classifiers have some potential in detecting the internal carotid artery stenosis and occlusion. The ANFIS model achieved accuracy rates which were higher than that of the stand-alone neural network model.

Wavelet-based neural network analysis of ophthalmic artery Doppler signals

In this study, ophthalmic artery Doppler signals were recorded from 115 subjects, 52 of whom had ophthalmic artery stenosis while the rest were healthy controls. Results were classified using a wavelet-based neural network. The wavelet-based neural network model, employing the multilayer perceptron, was used for analysis of ophthalmic artery Doppler signals. A multilayer perceptron neural network (MLPNN) trained with the Levenberg-Marquardt algorithm was used to detect stenosis in ophthalmic arteries. In order to determine the MLPNN inputs, spectral analysis of ophthalmic artery Doppler signals was performed using wavelet transform. The MLPNN was trained, cross validated, and tested with training, cross validation, and testing sets, respectively. All data sets were obtained from ophthalmic arteries of healthy subjects and subjects suffering from ophthalmic artery stenosis. The correct classification rate was 97.22% for healthy subjects, and 96.77% for subjects having ophthalmic artery stenosis. The classification results showed that the MLPNN trained with the Levenberg-Marquardt algorithm was effective to detect ophthalmic artery stenosis.

Wavelet-packet-based texture analysis for differentiation between benign and malignant liver tumours in ultrasound images

The purpose of this study was to apply a novel method of multiscale echo texture analysis for distinguishing benign (hemangiomas) from malignant (hepatocellular carcinomas (HCCs) and metastases) focal liver lesions in B-mode ultrasound images. In this method, regions of interest (ROIs) extracted from within the lesions were decomposed into subimages by wavelet packets. Multiscale texture features that quantify homogeneity of the echogenicity were calculated from these subimages and were combined by an artificial neural network (ANN). A subset of the multiscale features was selected that yielded the highest performance in the classification of lesions measured by the area under the receiver operating characteristic curve (Az). In an analysis of 193 ROIs consisting of 50 hemangiomas, 87 hepatocellular carcinomas and 56 metastases, the multiscale features yielded a high A: value of 0.92 in distinguishing benign from malignant lesions, 0.93 in distinguishing hemangiomas from HCCs and 0.94 in distinguishing hemangiomas from metastases. Our new multiscale texture analysis method can effectively differentiate malignant from benign lesions, and thus has the potential to increase the accuracy of diagnosis of focal liver lesions in ultrasound images.

Fourier transformation of arterial Doppler waveforms of the lower extremity

PURPOSE

Although it is well known that the normal, triphasic pulsatile arterial Doppler waveform changes in shape as flow is impaired, interpretation of the waveform has largely been subjective. We aimed to describe the Doppler waveforms of the lower extremity objectively using Fourier transformation.

METHODS

Sixty-eight zero-crossing detector arterial recordings from 25 lower extremities were grouped as follows: group 1, no ischemic symptoms with an ankle-brachial index (ABI) > 0.9 (n = 17, 8 limbs); group 2, no ischemic symptoms with ABI < 0.9 (n = 18, 5 limbs); group 3, symptoms of claudication (n = 19, 7 limbs); group 4, rest pain or tissue loss (n = 14, 5 limbs). The waveforms were Fourier transformed and their amplitudes and phases were compared up to the third harmonic (H3).

RESULTS

Amplitudes of both the fundamental (H1) and second harmonic (H2) were predominant in group 1. In contrast, amplitudes of the H2 and H3 decreased with altered flow (p < 0.0001 for group 1 versus others). The phases of the H1 and H2 were delayed with altered flow (p < 0.05 for group 1 versus others). Phases of the H1 were different between group 2 and 4 (p < 0.05). The difference of phase between the H3 and H1 was shortened with altered flow (p < 0.05 for group 1 or 2 versus group 4). Multivariate analysis revealed that the relative amplitudes of the H2 and H3, the phases of the H1 and H2, and the relative phase of the H3 were significant discriminators among the groups.

CONCLUSION

Abnormal waveforms could be characterized by the predominant amplitude of the H1, phase delay of the H1 and H2, and shortening of the relative phase of the H3. These parameters may be useful in the evaluation of Doppler waveforms in patients with peripheral arterial disease.

Detection of ophthalmic artery stenosis by least-mean squares backpropagation neural network

Doppler ultrasound is a noninvasive technique that allows the examination of the direction, velocity, and volume of blood flow. In this study, ophthalmic artery Doppler signals were obtained from 105 subjects, 48 of whom had suffered from ophthalmic artery stenosis. A least-mean squares backpropagation neural network was used to detect the presence or absence of ophthalmic artery stenosis. Spectral analysis of ophthalmic artery Doppler signals was done by the Welch method for determining the neural network inputs. The network was trained, cross validated and tested with subject records from the database. Performance indicators and statistical measures were used for evaluating the neural network. Ophthalmic artery Doppler signals were classified with the accuracy varying from 88.9% to 90.6%.

Classification of transcranial Doppler signals using artificial neural network

Transcranial Doppler signals, recorded from the temporal region of brain on 110 patients were transferred to a personal computer by using a 16-bit sound card. The fast Fourier transform (FFT) method was applied to the recorded signal from each patient. Since FFT method inherently can not offer a good spectral resolution at jet blood flows, it sometimes causes wrong interpretation of transcranial Doppler signals. To do a correct and rapid diagnosis, transcranial Doppler blood flow signals were statistically arranged so that they were classified in artificial neural network. Back propagation neural network and self-organization map algorithms of artificial neural network were used for training, whereas momentum and delta-bar-delta algorithms were used for learning. The results of these algorithms were compared in the case of classification and learning.

Evaluation of inherent performance of intelligent medical decision support systems: utilising neural networks as an example

Researchers who design intelligent systems for medical decision support, are aware of the need for response to real clinical issues, in particular the need to address the specific ethical problems that the medical domain has in using black boxes. This means such intelligent systems have to be thoroughly evaluated, for acceptability. Attempts at compliance, however, are hampered by lack of guidelines. This paper addresses the issue of inherent performance evaluation, which researchers have addressed in part, but a Medline search, using neural networks as an example of intelligent systems, indicated that only about 12.5% evaluated inherent performance adequately. This paper aims to address this issue by concentrating on the possible evaluation methodology, giving a framework and specific suggestions for each type of classification problem. This should allow the developers of intelligent systems to produce evidence of a sufficiency of output performance evaluation.

Implementation of intelligent decision support systems in health care

The full implementation of any intelligent system in health care, which is designed for decision support, has several stages, from initial problem identification through development and, finally, cost-benefit analysis. Central to this is formal objectivist evaluation with its core component of inherent performance of the outputs from these systems. A Medline survey of one type of intelligent system is presented, which demonstrates that this issue is not being addressed adequately. Lack of criteria for dealing with the outputs from these "black box" systems to prescribe adequate levels of inherent performance may be preventing their being accepted by those in the health-care domain and, thus, their being applied widely in the field.

A neural network approach for insulin regime and dose adjustment in type 1 diabetes

BACKGROUND

A decision support system based on a neural network approach is proposed to advise on insulin regime and dose adjustment for type 1 diabetes patients.

METHOD

The system consists of two feed-forward neural networks, trained with the back-propagation algorithm with momentum and adaptive learning rate. The input to the system consists of patient's glucose levels, insulin intake, and observed hypoglycemia symptoms during a short time period. The output of the first neural network provides the insulin regime, which is applied as input to the second neural network to estimate the appropriate insulin doses for a short time period.

RESULTS

The system's ability in order to recommend on insulin regime is excellent, while its performance in adjusting the insulin dosages for a specific patient is highly dependent on the data set used during the training procedure.

CONCLUSIONS

Despite the limitations of computer-based approaches, this study shows that artificial neural networks can assist diabetes patients in insulin adjustment.