Frederic D. McKenzie

EOG artifact removal using a wavelet neural network

In this paper, we developed a wavelet neural network (WNN) algorithm for electroencephalogram (EEG) artifact. The algorithm combines the universal approximation characteristics of neural networks and the time/frequency property of wavelet transform, where the neural network was trained on a simulated dataset with known ground truths. The contribution of this paper is two-fold. First, many EEG artifact removal algorithms, including regression based methods, require reference EOG signals, which are not always available. The WNN algorithm tries to learn the characteristics of EOG from training data and once trained, the algorithm does not need EOG recordings for artifact removal. Second, the proposed method is computationally efficient, making it a reliable real time algorithm. We compared the proposed algorithm to the independent component analysis (ICA) technique and an adaptive wavelet thresholding method on both simulated and real EEG datasets. Experimental results show that the WNN algorithm can remove EEG artifacts effectively without diminishing useful EEG information even for very noisy datasets.

Training in peacekeeping operations using virtual environments

Proper training of military personnel, at all levels, has never been more crucial. In this article, we describe the application of virtual environment technology to a novel and complex task-the military checkpoint. In the process, we also address differences between an immersive virtual environment training system and a destop version.

Model-based, real-time control of electrical power systems

Automated control of a large electrical power distribution network through a single controller can provide advantages in efficiency and reliability as well as reduction in maintenance costs. For control to be most effective, it is necessary that a global view of the entire network be had by the controller, so that it can reason as to the cause of the readings of the various sensing devices located throughout the network, Traditional approaches to power system control have involved a set of local devices (i.e., protective relays) that base their decision on the instantaneous reading of a single sensor. These single-parameter decisions can sometimes be incorrect due to sensor failures. Furthermore, a special type of fault called a soft fault, where a fault impedance limits the current to a value below the relay operating point, are nearly impossible to detect with decisions based on a single local parameter. By reasoning over an entire suit of sensing devices spread throughout the entire network, protection decisions based on a global view can become more reliable as well as comprehensive. Some previous approaches have implemented global control with varying degrees of success through the use of rule-based knowledge-based systems. This paper describes an alternative knowledge-based approach that makes use of so-called models of structure-and-behavior, to which model-based diagnosis is applied. The objective of this approach is to develop a system that can reliably diagnose faults in power distribution networks (especially soft faults), identify sensor failures, and carry out appropriate corrective action automatically. An intelligent power controller (IPC) which has these capabilities is described. This IPC was rigorously tested in an DC electrical power distribution system testbed and found to successfully carry out the required functions. This paper also describes in detail the tests and the conclusions drawn from their results.

Blood glucose individualized prediction for type 2 diabetes using iPhone application

Type 2 diabetes is now the most rapidly growing form of diabetes and has become increasingly common among children. This paper presents our work of implementing an individualized real time predictive system for blood glucose in type 2 diabetes in an iPhone application. The developed application, called HealthiManage, provides relevant feedback to patients at each glucose input reading comparing the measured and predicted readings, facilitating improved self-management of the disease. The application incorporates activity recognition via a built-in accelerometer on the iPhone, which monitors any physical activity and adjusts predictions accordingly. Also, a reward component interface was incorporated that is intended to enhance patient compliance and encourage mainly teenagers to take control and improve their blood glucose regulation. The individualized prediction algorithm was tested and verified with real patient data. Different physical activities were also examined and classified for an accurate activity recognition component. The designed application with its predictive model, activity recognition, and other elements provide what we believe to be helpful feedback to monitor and manage type 2 diabetes and improve patient compliance

High-dimensional MRI data analysis using a large-scale manifold learning approach

A novel manifold learning approach is presented to efficiently identify low-dimensional structures embedded in high-dimensional MRI data sets. These low-dimensional structures, known as manifolds, are used in this study for predicting brain tumor progression. The data sets consist of a series of high-dimensional MRI scans for four patients with tumor and progressed regions identified. We attempt to classify tumor, progressed and normal tissues in low-dimensional space. We also attempt to verify if a progression manifold exists—the bridge between tumor and normal manifolds. By identifying and mapping the bridge manifold back to MRI image space, this method has the potential to predict tumor progression. This could be greatly beneficial for patient management. Preliminary results have supported our hypothesis: normal and tumor manifolds are well separated in a low-dimensional space. Also, the progressed manifold is found to lie roughly between the normal and tumor manifolds.

Usefulness of Software Architecture Description Languages for Modeling and Analysis of Federates and Federation Architectures

Software architecture is high-level software design dealing with the structure and organization of large software systems. Architecture description languages (ADLs) are languages designed to represent software designs at the architecture level. ADLs are not widely used in the development of simulation systems. This research investigates the utility and effectiveness of ADLs for architecture-level design and analysis of simulation systems. Experimental applications of two ADLs to the specification and analysis of simulation architectures were conducted. Rapide was used to model the EnviroFed federation architecture and analyze data volume with and without interest management. Acme was used to model the ModSAF federate architecture and to analyze execution time at the component and federate levels in ModSAF. The experiments showed that ADLs could be used to discover important features of simulation system architectures.

Automatic diagnosis for prostate cancer using run-length matrix method

Prostate cancer is the most common type of cancer and the second leading cause of cancer death among men in US1. Quantitative assessment of prostate histology provides potential automatic classification of prostate lesions and prediction of response to therapy. Traditionally, prostate cancer diagnosis is made by the analysis of prostate-specific antigen (PSA) levels and histopathological images of biopsy samples under microscopes. In this application, we utilize a texture analysis method based on the run-length matrix for identifying tissue abnormalities in prostate histology. A tissue sample was collected from a radical prostatectomy, H&E fixed, and assessed by a pathologist as normal tissue or prostatic carcinoma (PCa). The sample was then subsequently digitized at 50X magnification. We divided the digitized image into sub-regions of 20 X 20 pixels and classified each sub-region as normal or PCa by a texture analysis method. In the texture analysis, we computed texture features for each of the sub-regions based on the Gray-level Run-length Matrix(GL-RLM). Those features include LGRE, HGRE and RPC from the run-length matrix, mean and standard deviation of the pixel intensity. We utilized a feature selection algorithm to select a set of effective features and used a multi-layer perceptron (MLP) classifier to distinguish normal from PCa. In total, the whole histological image was divided into 42 PCa and 6280 normal regions. Three-fold cross validation results show that the proposed method achieves an average classification accuracy of 89.5% with a sensitivity and specificity of 90.48% and 89.49%, respectively.

Augmented standardized patients now virtually a reality

Standardized patients (SPs), individuals who realistically portray patients, are widely used in medical education to teach and assess communication skills, eliciting a history, performing a physical exam, and other important clinical skills. One limitation is that each SP can only portray a limited set of physical symptoms. Finding SPs with the abnormalities students need to encounter is typically not feasible. This project augments the SP by permitting the learner to hear abnormal heart and lung sounds in a normal SP.

An integrated model-based approach for real-time on-line diagnosis of complex systems

Abstract Model-based diagnostic programs have been shown to be useful in isolating unpredictable faults in various types of systems. Due to the complex nature of many of these systems, models used by these programs to represent monitored systems have traditionally imposed restrictions on domain representations. These restrictions can make it difficult (and often impossible) to model a domain whose behavior is global in nature. By global, is meant behavior that affects system variables in parts of the system not directly related to the component in question. Analog electrical circuits and hydraulic circuits are only a few examples of such global systems. Accurate modelling of the behavior of these global systems is very often essential for obtaining a correct diagnosis. In complex systems such as those typically found in the electrical power-distribution domain, global behavior can be observed when voltages and currents throughout an entire system are affected by local load fluctuations, transient disturbances, faults, or circuit re-configurations, even when these are in remote parts of the circuit. Traditional models used in diagnosis have not been able to easily reflect these global interactions, and as a result, monitoring and diagnostic capabilities of model-based systems dependent upon such models are significantly degraded. This paper presents an implementation that can correctly simulate power systems and other such complex systems by overcoming the problem of representing global behavior while preserving the diagnostic abilities of structure–function models in model-based reasoning methodologies. This paper describes the integration of robust models , within the conventional device-centered models. These robust models are mathematically accurate system models, normally used in quantitative simulation for the purpose of system analysis. If used within the conventional device-centered models, they can provide the functionality needed in a structure–function model-based diagnostic paradigm, and therefore eliminate the problem of representing global behaviors in diagnosis. This paper further describes a conflict-oriented diagnostic technique used in conjunction with robust models to obtain real-time on-line FDIR (Fault Diagnosis, Isolation, and Recovery).

Development of an Average Chest Shape for Objective Evaluation of the Aesthetic Outcome in the Nuss Procedure Planning Process

The Nuss procedure is a minimally invasive surgery for correcting pectus excavatum. Pectus excavatum (PE), also called sunken or funnel chest, is a congenital chest wall deformity which is characterized by a deep depression of the sternum. This condition affects primarily children and young adults and is responsible for about 90% of congenital chest wall abnormalities.