Chenglin Peng

A multilayer perceptron-based medical decision support system for heart disease diagnosis

The medical diagnosis by nature is a complex and fuzzy cognitive process, and soft computing methods, such as neural networks, have shown great potential to be applied in the development of medical decision support systems (MDSS). In this paper, a multiplayer perceptron-based decision support system is developed to support the diagnosis of heart diseases. The input layer of the system includes 40 input variables, categorized into four groups and then encoded using the proposed coding schemes. The number of nodes in the hidden layer is determined through a cascade learning process. Each of the 5 nodes in the output layer corresponds to one heart disease of interest. In the system, the missing data of a patient are handled using the substituting mean method. Furthermore, an improved back propagation algorithm is used to train the system. A total of 352 medical records collected from the patients suffering from five heart diseases have been used to train and test the system. In particular, three assessment methods, cross validation, holdout and bootstrapping, are applied to assess the generalization of the system. The results show that the proposed MLP-based decision support system can achieve very high diagnosis accuracy (>90%) and comparably small intervals (<5%), proving its usefulness in support of clinic decision process of heart diseases.

Selecting critical clinical features for heart diseases diagnosis with a real-coded genetic algorithm

In clinic, normally a lot of diagnostic features are recorded from a patient for a certain disease. It will be beneficial for the prompt and correct diagnosis of the disease by selecting the important and relevant features and discarding those irrelevant and redundant ones. In this paper, a real-coded genetic algorithm (GA)-based system is proposed to select the critical clinical features essential to the heart diseases diagnosis. The heart disease database used in this study includes 352 cases, and 40 diagnostic features were recorded for each case. Using the proposed genetic algorithm, 24 critical features have been identified, and their corresponding diagnosis weights for each heart disease of interest have been determined. The critical diagnostic features and their clinic meanings are in sound agreement with those used by the physicians in making their clinic decisions.

Development of a decision support system for heart disease diagnosis using multilayer perceptron

In this paper, a computational model based on a multilayer perceptron (MLP) neural network with three layers is employed to develop a decision support system for the diagnosis of five major heart diseases. The input layer of the system includes 38 input variables, extracted from a large number of patient cases. The number of nodes in the hidden layer is determined through a cascade learning process. Each of the 5 nodes in the output layer corresponds to one heart disease of interest. The proposed decision support system is trained using a back propagation algorithm augmented with the momentum term, the adaptive learning rate and the forgetting mechanics. In addition, the missing data are handled using the substituting mean method. The experimental results have shown that the adopted MLP-based decision model can achieve high accuracy level (63.6-82.9%) on the classification of heart diseases, qualifying it as a good decision support system deployable in clinics.

The internet-based knowledge acquisition and management method to construct large-scale distributed medical expert systems

The Internet offers an unprecedented opportunity to construct powerful large-scale medical expert systems (MES). In these systems, a cost-effective medical knowledge acquisition (KA) and management scheme is highly desirable to handle the large quantities of, often conflicting, medical information collected from medical experts in different medical fields and from different geographical regions. In this paper, we demonstrate that a medical KA/management system can be built upon a three-tier distributed client/server architecture. The knowledge in the system is stored/managed in three knowledge bases. The maturity of the medical know-how controls the knowledge flow through these knowledge bases. In addition, to facilitate the knowledge representation and application in these knowledge bases as well as information retrieval across the Internet, an 8-digit numeric coding scheme with a weight value system is proposed. At present, a medical KA and management system based on the proposed method is being tested in clinics. Current results have showed that the method is a viable solution to construct, modify, and expand a distributed MES through the Internet.

Handgrip Force Estimation Based on a Method Using Surface Electromyography (sEMG) of Extensor Carpi Radialis Longus

Both flexor and extensor muscle activated together when hand-grip task conducted, but there is little work that attempts to specifically investigate the relationship of the hand-grip force level and the EMG activity of extensor muscles. The present study was designed to investigate the correlation between hand-grip force level and sEMG of ECRL (extensor carpi radialis longus, ECRL). A pseudo-randomized sequence of hand-grip tasks with some specific force ranges has been defined for calibration. Eight subjects (university students, five males and three females) were recruited to conduct both calibration trials and voluntary trials. EMG signals have been preprocessed with RMS (root-mean-square) method, after which EMG signals are normalized with amplitude value of MVC-related EMG. With data regression of calibration trials, a linear model has been developed to correlate the handgrip force output with sEMG activities of ECRL and this linear model then is employed to estimate the hand-grip force production of voluntary trials. The root-mean-square-error (RMSE) of the estimated force output for all the voluntary trials are statistically compared in different force ranges. The results indicate that the linear model is useful to estimate the handgrip force based on the EMG activities of forearm extensor muscle, and the accuracy of this model is dependent on the force levels. That is the linear model can provide best estimation in moderate force range (30%-50% MVC), while the force prediction error tends to be large for weak force (20%-30% MVC) or strong force (50%-80% MVC).

Experimental Study of Magnetic-based Localization Model for Miniature Medical Device Placed Indwelling Human Body

More and more miniature medical devices (MMD), such as alimentary canal miniature drug delivering device, wireless endoscope, feeding tubes and treatment probes, have been researched and developed for clinical diagnosis and therapy. When a MMD was put into the human body for diagnosis and therapy, it is necessary to provide a non-invasive locating method that can locate the MMD in long-term. Here, a novel magnetic-based position method has been explored. According to the principle of magnet field distribution of magnetic dipole, a special multi-sensor magnetic detecting system has been designed and a modified magnetic position model was proposed based on curve-fitting process with detected data. Models with different coefficient have been compared, and a better one has been constructed. The results demonstrate that the proposed magnetic-based position method is promising and effective

Force measurement on fracture site with external fixation.

A force measurement device has been designed to monitor the mechanical properties of fracture site with external fixation. Forces are measured through electric resistance strain gauges mounted on fixation framework and the measurement results are displayed on an LCD screen. The device features a force range of 0–10 kg with linearity and repeatability less than 1% and accuracy less than 0.1 kg.

A STUDY OF MODELS FOR HANDGRIP FORCE PREDICTION FROM SURFACE ELECTROMYOGRAPHY OF EXTENSOR MUSCLE

Force production involves the coordination of multiple muscles, and the produced force levels can be attributed to the electrophysiology activities of those related muscles. This study is designed to explore the activity modes of extensor carpi radialis longus (ECRL) using surface electromyography (sEMG) at the presence of different handgrip force levels. We attempt to compare the performance of both the linear and nonlinear models for estimating handgrip forces. To achieve this goal, a pseudo-random sequence of handgrip tasks with well controlled force ranges is defined for calibration. Eight subjects (all university students, five males, and three females) have been recruited to conduct both calibration and voluntary trials. In each trial, sEMG signals have been acquired and preprocessed with Root–Mean–Square (RMS) method. The preprocessed signals are then normalized with amplitude value of Maximum Voluntary Contraction (MVC)-related sEMG. With the sEMG data from calibration trials, three models, Linear, Power, and Logarithmic, are developed to correlate the handgrip force output with the sEMG activities of ECRL. These three models are subsequently employed to estimate the handgrip force production of voluntary trials. For different models, the Root–Mean–Square–Errors (RMSEs) of the estimated force output for all the voluntary trials are statistically compared in different force ranges. The results show that the three models have different performance in different force ranges. Linear model is suitable for moderate force level (30%–50% MVC), whereas a nonlinear model is more accurate in the weak force level (Power model, 10%–30% MVC) or the strong force level (Logarithmic model, 50%–80% MVC).

Characterization of finger isometric force production with maximum power of surface electromyography

Fingers action has been controlled by both intrinsic and extrinsic hand muscles. Characterizing the finger action with the activations of hand muscles could be useful for evaluating the neuromuscular control strategy of fingers motor functions. This study is designed to explore the correlation of isometric fingertip force production and frequency-domain features of surface electromyography (sEMG) recorded on extrinsic hand muscles. To this end, 13 subjects (five male and eight female university students) have been recruited to conduct a target force-tracking task. Each subject is required to produce a certain level of force with either the index or middle fingertip to match the pseudo-random ordered target force level (4N, 6N, or 8N) as accurate as possible. During the finger force production process, the sEMG signals are recorded on two extrinsic hand muscles: flex digitorum superficials (FDS) and extensor digitorum (ED). For each sEMG trail, the power spectrum is estimated with the autoregressive (AR) model and from which the maximum power is obtained. Our experimental results reveal three findings: (1) the maximum power increases with the force level regardless of the force producing finger (i.e. index or middle) and the extrinsic hand muscle (i.e. FDS or ED). (2) The sEMG maximum power of index finger is significantly lower than that of the middle finger under the same force level and extrinsic hand muscle. (3) No significant difference can be found between the maximum powers of FDS and ED. The results indicate that the activations of the extrinsic muscles are affected by both the force level and the force producing finger. Based on our findings, the sEMG maximum power of the extrinsic hand muscles could be used as a key parameter to describe the fingers actions.

Design of Additional Power Supply for Subretinal Implants device

In the paper, a methodology of highly efficient wireless power transmission in retinal prosthesis was presented. It is based on inductive coupled coil pair which is the most common method in medical implants. In the design, a class-E driver was used in the primary side. In order to reduce the dissipation in the retina, an improved half wave rectifier circuit was adopted to get effective voltage in the secondary side. Then the design is optimized by orcad/pspice9.2. The third part is the simulation result; the load voltage of design can reach 5 volts. The last part is conclusion.