Inter-patient arrhythmia ECG signal classification based on SVM+XGBoost ensemble classifier / 国际生物医学工程杂志

Objective:To study a classifier used to classify arrhythmia electrocardiogram (ECG) signals under the inter-patient paradigm to improve the accuracy of automatic classification and solve the limitations of manual diagnosis of arrhythmia.Methods:A SVM+XGBoost ensemble classifier with four modules including preprocessing, feature extraction, support vector machine (SVM) training and ensemble classification was constructed. ECG signal was preprocessed, and R-R interval, high-order statistics, local binary patterns and wavelet components were used as features to train independent SVM classifiers. Then, XGboost algorithm was used to integrate independent SVM classifiers and output arrhythmia classification results. The integrated classifiers were trained and tested on MIT-BIH database.Results:The overall classification accuracy of the ensemble classifier for arrhythmia was 0.867 and the average sensitivity was 0.782.Conclusions:The proposed ensemble classifier can realize automatic and accurate classification of arrhythmia ECG signals under the inter-patient paradigm, and can be used for clinical auxiliary diagnosis.

Finite element analysis of temperature field of retina by electrical stimulation with microelectrode array / 生物医学工程学杂志

We studied the influence of electrode array parameters on temperature distribution to the retina during the use of retinal prosthesis in order to avoid thermal damage to retina caused by long-term electrical stimulation. Based on real epiretinal prosthesis, a three-dimensional model of electrical stimulation for retina with 4 X 4 microelectrode array had been established using the finite element software (COMSOL Multiphysics). The steady-state temperature field of electrical stimulation of the retina was calculated, and the effects of the electrode parameters such as the distance between the electrode contacts, the materials and area of the electrode contact on temperature field were considered. The maximum increase in the retina steady temperature was about 0. 004 degrees C with practical stimulation current. When the distance between the electrode contacts was changed from 130 microm to 520 microm, the temperature was reduced by about 0.006 microC. When the contact radius was doubled from 130 microm to 260 microm, the temperature decrease was about 0.005 degrees C. It was shown that there were little temperature changes in the retina with a 4 x 4 epiretinal microelectrode array, reflecting the safety of electrical stimulation. It was also shown that the maximum temperature in the retina decreased with increasing the distance between the electrode contacts, as well as increasing the area of electrode contact. However, the change of the maximum temperature was very small when the distance became larger than the diameter of electrode contact. There was no significant difference in the effects of temperature increase among the different electrode materials. Rational selection of the distance between the electrode contacts and their area in electrode design can reduce the temperature rise induced by electrical stimulation.

A new method for extracting visual saliency map for visual prosthesis / 国际生物医学工程杂志

Objective Salient regions may express the main content (salient objects) in an image or scene.Since the number of implanted electrodes for visual prosthesis is limited,only the low-resolution image is used.Extraction of the salient regions may help capture salient objects in the scenes of visual prosthesis.Methods Salient regions detected by Itti model,one of salient detection models,have some differences with human visual perceptions and its salient objects boundary are also unclear.The Itti model was optimized by removing the orientation and the color characteristics,and converting the RGB image in an image or scene into the HSI color space with three new feature components of hue,saturation and intensity.In saliency map,the ratio of the salient point area in salient objects regions and the total salient point area was defined as salient accuracy and computed,which was used to compare the effects of the two methods of extracting salient map.Results Compared with the Itti method for extracting saliency map,the salient accuracy of the improved method was increased by about 20％,and the time for detecting the same salient regions by our method was shorter by nearly 50％.Conclusions The proposed algorithm can be an alternative method to extract salient map from a complex image or scene for retinal prosthesis with higher salient accuracy and less running time.

Simulation analysis of extracellular selective stimulation of optic nerve with new helix electrode / 国际生物医学工程杂志

ObjectiveElectric stimulation of the central nervous system has been served as a treatment method for variety of neurological,psychiatric and sensory disorders.Despite considerable success in some applications,current stimulation techniques offer little control over which neuronal targets are activated by stimulation.This study aimed to present a new shape self-adaptive helix electrode for selective activation of optic nerve.MethodsThe geometric model of optic nerve and new helix electrode was elaborated with COMSOL Multiphysics.The new helix electrodes consist of silicone helix frame,which acted as support and insulation,and platinum contacts embedded within the frame.The activating function (AF) was introduced to characterize the stimulation effects,and the selectivity of activating optic nerve fascicle with new helix electrode was simulated by COMSOL Multiphysics,taking into account the variations of electrode contact locations.ResultsAssuming normalized AF threshold was 0.1 V/m2,the ratio difference of AF over threshold between new helix electrode and traditional electrode was only 1.2410％.With contacts in two ends of helix electrode closer to the middle contact,the small nerve fascicle was first activated and then the large one.ConclusionThe results show that the new helix electrodes have the same stimulation effects as that of traditional cuff electrodes.The new helix electrodes can selectively activate optic nerve fascicle with variations of electrode contact locations.

Simulation of axonal excitability using LABVIEW / 国际生物医学工程杂志

ObjectiveIn this paper a new application of LABVIEW in neural electrophysiological simulation was applied to simulate axonal excitability.MethodsA variety of axonal excitability of the squid giant axon was modeled with LABVIEW,including action potential threshold,refractoriness,temporal summation,anode break excitation,repetitive firing and conduction velocity.ResultsThe LABVIEW program accurately described the permeability changes of Na+ and K+ and their effects on the action potential,and the membrane excitability had been well predicted.ConclusionThe graphical programming language of LABVIEW allows the user to complete the programming by dragging and dropping icons and connecting wires,the users can quickly and easily construct their own biological research systems.

Finite element analysis of electric field of extracellular stimulation of optic nerve with a spiral cuff electrode / 生物医学工程学杂志

In order to study the underlying electrode-nerve functional mechanism, optimize the electrode design and guide the prosthesis application, we applied finite element method to analyze the spatial distribution of electric field generated by optic nerve electrical stimulation with spiral cuff electrode. A macroscopic cylindrical model of optic nerve was elaborated, taking into account of electrode contact configurations and possible variations of the thickness of cerebrospinal fluid (CSF). By building an appropriate mesh on this model and under some boundary conditions, the finite element method was applied to compute the 3D electric field generated by the electrode with finite element software COMSOL Multiphysics. The stimulation results indicated that, under the same conditions of stimulation, the longitudinal tripolar electrode structure could generate larger current density than that of biopolar electrode structure (located in the opposite of nerve trunk). However biopolar electrode structure requirs less leads, and is more easily implanted. By means of parametric sweep, the results suggest that, with the increase of the CSF thickness and a higher conductivity of CSF than those of other tissues, the distribution of electric field generated by electrodes is extended but scattered, and the diffuse current distribution makes nerve stimulation less effective.

Simulation analysis of optic nerve stimulation with penetrating electrode / 国际生物医学工程杂志

ObjectiveOptic nerve stimulation with penetrating electrode.is a new method for developing visual prosthesis.A simulation system was developed with computer software MATLAB to investigate its mechanism.MethodsVolume conductor model of optic nerve and optic nerve fiber model were developed.With the stimulation of monopolar penetrating electrode,the excitation thresholds of fibers at different depth were calculated.The activating function and activation region were employed to characterize the external stimulation effect.The impact of different parameters on stimulation effects were explored by changing the fiber diameter and the stimulation pulse width.Results Excitation thresholds increased as well as the percentage of activated fibers with the increase of electrode-fiber distance.Excitation threshold of the same depth fiber decreased as the fiber diameter increased,and the longer the electrode-fiber distance was,the more significant the drop was.Excitation threshold of the same depth fiber decreased as the pulse width of monophasic rectangle wave increased.The excitation threshold of the fiber at the same depth hardly changed for pulse durations greater than 0.5 ms.ConclusionThe simulation results indicates that optical nerve stimulation can be realized with penetratingelectrode,and different stimulation parameters can produce different activation effects.The present results providetheoretical guidance for the design of experiments.

Effects of potassium concentration and ion channel conductance on spontaneous firing of Hippocampal CA1 neuron / 国际生物医学工程杂志

Objective To analyse the variability of spontaneous firing frequencies and firing patterns of a Hippocampal CA1 neuron with the changes of extracellular potassium concentration[K+]0 and ion channel conductance.Methods A 16-compartment cable model of a Hippocampal CA1 neuron was developed based on the Warman model with computer software MATLAB.The dendrites contained no active channels,while five active channels (INa,INap,IDR,IA,IM)were contained in soma.In the model,the calcium currents and potassium currents(ICm,,ICT and IAHP) regulated by calcium concentration were not included.ResultsSimulation results showed that the neuron could generate periodicspontaneous firing activity.The spiking frequency increased with the increasing of[K+]0 and sodium conductance and decreasing of potassium conductance.Spontaneous single neuron activity appeared in singlet or in grouped bursts of two or three spikes.Conclusion The variability of spontaneous firing frequencies and firing patterns of single neuron are relevant to[K+]0 and ion channel conductance.

Modeling and simulating study on disorder of associative memory function of hippocampal CA3 region caused by synaptic deletion based on Hopfield-like spiking neural network model / 国际生物医学工程杂志

Objective A spiking Hopfield-like neural network was proposed and used to simulate the disorder of hippocampal associative memory disorder caused by synaptic deletion.Methods A three-layered Hopfield-like spiking neural network model with auto-associative memory function and hetero-associative memory function was proposed according to anatomical structure of hippocampal CA3,and both associative memories of the models were simulated under Matlab platform.Disorder of hippocampal associative memory was also simulated according to Ruppin's 'synaptic deletion' theory.Results With the increasing of synaptic deletion level,both associative memory functions impaired gradually.Conclusion Synaptic deletion of hippocampal CA3 region can lead to the disorder of autoassociative memory as well as heteroassociative memory.

Simulation of extracellular electrical stimulation of rat optic nerve fiber based on experimental data / 中国组织工程研究

BACKGROUND:Currently,there are three types of visual prosthesis:retinal prosthesis,optic nerve prosthesis and visual cortical prosthesis.Normal optic nerve fiber is the basis to apply visual prosthesis.Extracellular stimulation of myelinated optic nerve fiber is developed from intracellular stimulation,with feasible selection of stimulation locations and low requirements for operation.OBJECTIVE: Using computer to simulate the response of myelinated rat optic nerve fiber to extracellular point current source and to analyze the effect of the position of the point source on the response of the optic nerve fiber.METHODS:The experiment data of ion channels of rat optic nerve fiber and the characteristic of common ion channels were summarized,and the existing Hodgkin-Huxley like optic nerve fiber models were analyzed.The response of the optic nerve fiber model was simulated by computer,and the differences of action potentials of Ranvier node were compared by changing the position of point source.RESULTS AND CONCLUSION:Simulated models showed that the response of optic nerve is affected by the position of extracellular exciting electrode.The myelinated rat optic nerve fiber is easier to activate when the distance between point source and the Ranvier node center of the optic nerve fiber is short.

Neural network modeling of hippocampal CA3 associative memory functions / 中国组织工程研究

BACKGROUND: Hippocampus is one of an important brain areas related with memory,and plays a critical role in associative memory function.Hippocampal CA3 is one of the most important regions to form associative memory.CA3 is functionally divided into autoassociative and heteroassociative memories,and memory formation and retrieval require the development of detailed models of hippocampal function.OBJECTIVE: To establish a detailed model of hippocampal CA3 function according to CA3 structure.METHODS: The model was a three-layered Hopfield-like neural network and was constituted by 280 Izhikevich artificial neurons,and is modulated by Hebbian rules.The model was simulated using MATLAB under the condition of adding the Gaussian white noise to its input.In the simulation,memories were represented by synchronous firing sequences.RESULTS AND CONCLUSION: The simulating results show that the third layer of model had heteroassociative memory function;the first and the second layer of the model could implement autoassociative memory.The model implemented well the memory functions of three subregions of hippecampal CA3.But it is impossible to understand the functions and dynamics of a real biological neural network by constructing a simple model.The model proposed has 280 neurons,which are far less than the real number of neurons.It suggests that there is a big gap between the properties of the model and a real biological neural network of CA3.

Development and Study of Associative Memory Neural Network / 医疗卫生装备

The concept and important role of associative memory neural network are introduced.The development of associative memory neural network is also summarized along with other neural networks such as the Hopfield neural network,the neural network composed of FitzHugh-Nagumo neuron model,the neural network composed of Hodgkin-Huxley neuron model.

Development of PET scanner for small animals / 医疗卫生装备

In comparison with the clinical PET scanner,PET scanner for small animals puts forward a higher demand in spacial resolution and sensitivity. This article describes the development of the PET scanner for small animals and compares the characteristics of several kinds of contemporary typical PET scanners for small animals in different aspects based on the development of detectors. Meanwhile,it also introduces the development of the detectors designed for measuring depth of interaction (DTC) to reduce parallax error.