Search

Perimetry is essential for identifying visual field defects due to disorders of the eye and brain. However, young children are often unable to reliably perform the preferred method of visual field assessment known as automated static perimetry (ASP). This paper introduces a novel method of ASP specifically developed for children called Saccadic Vector Optokinetic Perimetry (SVOP). SVOP uses eye tracking to detect the natural saccadic eye response of gaze orientation towards visual field stimuli if they are seen. In this paper, the direction and magnitude of a sample of subject gaze responses to visual field stimuli is used to construct a software decision algorithm for use in SVOP. SVOP was clinically evaluated in a group of 24 subjects, comprising children and adults, with and without visual field defects, by comparison with an equivalent test on the Humphrey Field Analyser (HFA). SVOP provides promising visual field test results when compared with the reference HFA test, and has proven extremely useful in detecting visual field defects in children unable to perform traditional ASP.

Performance of robotics based hand-held surgical devices in real-time is mainly dependent on accurate filtering of physiological tremor. The presence of phase delay in sensors (hardware) and filtering (software) processes affects the cancellation accuracy. This paper focuses on developing an estimation algorithm to improve the estimation accuracy in the presence of phase delay for real-time implementations. Moving window based online training approach for least squares-support vector machines (LSSVM) is employed in this paper for tremor estimation. A study is conducted with tremor data recorded from the subjects to analyze the suitability of proposed approach for both single-step and multi-step prediction.

Anthropometric measurements, such as the circumferences of the hip, arm, leg and waist, waist-to-hip ratio, and body mass index, are of high significance in obesity and fitness evaluation. In this paper, we present a home based imaging system capable of conducting automatic anthropometric measurements. Body images are acquired at different angles using a home camera and a simple rotating disk. Advanced image processing algorithms are utilized for 3D body surface reconstruction. A coarse body shape model is first established from segmented body silhouettes. Then, this model is refined through an inter-image consistency maximization process based on an energy function. Our experimental results using both a mannequin surrogate and a real human body validate the feasibility of proposed system.

Dual-frequency steady-state visual evoked potential (SSVEP) was suggested to generate more stimuli using a few flickering frequencies for brain-computer interface. Dual--frequency SSVEP peaks at more than two frequencies-both main and harmonic frequencies. However multi-frequency recognition strategy has not been investigated for dual-frequency SSVEP. In this paper, three modified power spectral density analysis (PSDA) methods and two modified canonical correlation analysis (CCA) methods were tested for dual-frequency SSVEP classification. Three methods among the five methods used conventional features or classification techniques, and the other two methods used modified features for harmonic frequencies. As a result, CCA with novel features showed the best BCI performance. Also the use of harmonic frequencies improved BCI performance of dual-frequency SSVEP.

Electromyography (EMG) pattern-recognition based control strategies for multifunctional myoelectric prosthesis systems have been studied commonly in a controlled laboratory setting. Most previous efforts concentrated on evaluating the performance of EMG pattern-recognition algorithms in identifying one signal movement at a time. Therefore, the current motion classification methods would be limited with the difficulties in identifying the combined upper-limb motion classes that are commonly required in performing activities daily. In this paper, four improved classifier training schemes were proposed and investigated to address the difficulties mentioned above. Our preliminary results showed that three of the four proposed training schemes could improve the classification performance. The average classification accuracies of the three methods were 75.10% ± 9.71%, 76.95% ± 8.02%, and 77.56% ± 6.55% for the able-bodied subjects, and 63.38% ± 7.51%, 62.55% ± 9.06%, and 62.50% ± 9.36% for the transradial amputees, respectively. These results suggested that the proposed methods could provide better classification performance in identifying the combined motions than the current methods.

Accurate measurement of lumen diameter is essential for correct estimation of arterial compliance. We have been developing a new non-invasive arterial compliance measurement tool using a single element ultrasound transceiver. In this paper we propose a new method for measurement of lumen diameter from single line of Radio-Frequency Signal (RF) obtained from the common carotid artery (CCA). The method is free from fixed thresholds and uses shape fitting to get objective measurement. The accuracy of the algorithm was found to be better than 5 % for software simulated and phantom arteries and better than 10 % in case of data obtained from CCA of human volunteers.

This paper presents cardiac arrhythmia detection using the combination of a heart rate variability (HRV) analysis and a "potential of unbalanced complex kinetics" (PUCK) analysis. Detection performance was improved by adding features extracted from the PUCK analysis. Initially, R-R interval data were extracted from the original electrocardiogram (ECG) recordings and were cut into small segments and marked as either normal or arrhythmia. HRV analyses then were conducted using the segmented R-R interval data, including a time-domain analysis, frequency-domain analysis, and nonlinear analysis. In addition to the HRV analysis, PUCK analysis, which has been implemented successfully in a foreign exchange market series to characterize change, was employed. A decision-tree algorithm was applied to all of the obtained features for classification. The proposed method was tested using the MIT-BIH arrhythmia database and had an overall classification accuracy of 91.73%. After combining features obtained from the PUCK analysis, the overall accuracy increased to 92.91%. Therefore, we suggest that the use of a PUCK analysis in conjunction with HRV analysis might improve performance accuracy for the detection of cardiac arrhythmia.

Compression is increasingly used in medical applications to enable efficient and universally accessible electronic health records. However, lossy compression introduces artifacts that can alter diagnostic accuracy, interfere with image processing algorithms and cause liability issues in cases of diagnostic errors. Compression guidelines were introduced to mitigate these issues and foster the use of modern compression algorithms with diagnostic imaging. However, these guidelines are usually defined as maximum compression ratios for each imaging protocol and do not take compressibility variations due to image content into account. In this paper we have evaluated the compressibility of thousands of computed tomography slices of an anthropomorphic thoracic phantom acquired with different parameters. We have shown that exposure, slice thickness and reconstruction filters have a significant impact on compressibility suggesting that guidelines based solely on compression ratios may be inadequate.

In this paper, we introduce a Concept Graph Engine (CG-Engine) that generates patient specific personalized disease ranking based on the laboratory test data. CG-Engine uses the Unified Medical Language System database as medical knowledge base. The CG-Engine consists of two concepts namely, a concept graph and its attributes. The concept graph is a two level tree that starts at a laboratory test root node and ends at a disease node. The attributes of concept graph are: Relation types, Semantic types, Number of Sources and Symmetric Information between nodes. These attributes are used to compute the weight between laboratory tests and diseases. The personalized disease ranking is created by aggregating the weights of all the paths connecting between a particular disease and contributing abnormal laboratory tests. The clinical application of CG-Engine improves physician's throughput as it provides the snapshot view of abnormal laboratory tests as well as a personalized disease ranking.

Detection of spines in medical images are important tasks in medical applications. These tasks are relatively easy for CT/MR images because the bones are easily distinguishable from other tissues. However, they are difficult for x-ray images due to bone and soft tissue overlapping. This paper illustrates a method for detecting the medial axis of spine in x-ray images. Given an initial point on the spine in the x-ray image manually or automatically, the method iteratively searches for good feature points on the spine to locate the medial axis. As a result, the effort of determining the relevant medical information, such as Cobb's angle, can be minimized. The proposed method is fast and efficient. In average it took less than 1 second for localizing the spine on a 3000×1000 gray scale x-ray image.

Skeletal muscle is an important tissue of human body, and its contractions control and regulate body motions. Muscle contraction results in morphological changes of the related muscles. Ultrasound imaging is an effective tool for studying muscle architectures and monitoring the morphological changes of muscles. The latter process can be realized with a motion estimation algorithm. However, ultrasound images are usually corrupted by speckle noises and performance of motion estimation methods can be significantly affected by the noises. To get a better performance in motion analysis, in this paper, as a pre-processing step, an adaptive filter named adaptive guided image filtering (AGF) is suggested to reduce speckle noises. We first transformed the multiplicative noise model into an additive one by taking the logarithm of the original speckled data, then performed AGF to obtain the filtered image, and finally took the tackled image back into exponent. Experimental results showed that AGF had a better performance in terms of noise attenuation and edge preservation compared with other standard filters. In quantitative results, the filtered images also had the highest Peak-Signal-to-Noise Ratio (PSNR) using AGF. It's believed that AGF is a good choice for the pre-processing stage of muscle motion analysis.

This paper develops a novel method to reduce the influence of stimulus-induced artifacts in functional spinal cord imaging. The developed method employes a two-step procedure. The first step acquires artifact data, which contain artifacts but do not contain spinal cord evoked magnetic field (SCEF). The second step applies a method called common-mode subspace projection (CSP). The effectiveness of the developed method is validated using SCEF data measured from a healthy volunteer.

Nowadays numerous technologies are employed for tracking patients and assets in hospitals or nursing homes. Each of them has advantages and drawbacks. For example, WiFi localization has relatively good accuracy but cannot be used in case of power outage or in the areas with poor WiFi coverage. Magnetometer positioning or cellular network does not have such problems but they are not as accurate as localization with WiFi. This paper describes technique that simultaneously employs different localization technologies for enhancing stability and average accuracy of localization. The proposed algorithm is based on fingerprinting method paired with data fusion and prediction algorithms for estimating the object location. The core idea of the algorithm is technology fusion using error estimation methods. For testing accuracy and performance of the algorithm testing simulation environment has been implemented. Significant accuracy improvement was showed in practical scenarios.

In order to evaluate people’s lifestyle for health maintenance, this paper presents a segmentation method based on multi-sensor data recorded by a wearable computer called eButton. This device is capable of recording more than ten hours of data continuously each day in multimedia forms. Automatic processing of the recorded data is a significant task. We have developed a two-step summarization method to segment large datasets automatically. At the first step, motion sensor signals are utilized to obtain candidate boundaries between different daily activities in the data. Then, visual features are extracted from images to determine final activity boundaries. It was found that some simple signal measures such as the combination of a standard deviation measure of the gyroscope sensor data at the first step and an image HSV histogram feature at the second step produces satisfactory results in automatic daily life event segmentation. This finding was verified by our experimental results.

Accurate multi-tissue segmentation of magnetic resonance (MR) images is an essential first step in the construction of a realistic finite element head conductivity model (FEHCM) for electroencephalography (EEG) source localization. All of the segmentation approaches proposed to date for this purpose require manual intervention or correction and are thus laborious, time-consuming, and subjective. In this paper we propose and evaluate a fully automatic method based on a hierarchical segmentation approach (HSA) incorporating Bayesian-based adaptive mean-shift segmentation (BAMS). An evaluation of HSA-BAMS, as well as two reference methods, in terms of both segmentation accuracy and the source localization accuracy of the resulting FEHCM is also presented. The evaluation was performed using (i) synthetic 2D multi-modal MRI head data and synthetic EEG (generated for a prescribed source), and (ii) real 3D T1-weighted MRI head data and real EEG data (with expert determined source localization). Expert manual segmentation served as segmentation ground truth. The results show that HSA-BAMS outperforms the two reference methods and that it can be used as a surrogate for manual segmentation for the construction of a realistic FEHCM for EEG source localization.

Developing patient specific model for the simulation of chemotherapy drug injection is important in medical application. This paper proposed a two-phase fluidic method to simulate chemotherapy drug injection and an improved lumped element method to simulate deformation of vessel at real-time by using GPU for general computing. Firstly, a three-dimensional (3-D) model of hepatic vessels is reconstructed from clinical CT-images using multi-layer method. A 3-D thinning algorithm based on Valence Driven Spatial Median (VDSM) is applied to generate unit-width skeleton of the vessel tree. The two-phase flow simulation of drug injection is based on Hagen-Poiseuille model by introducing a friction factor using bubbly flow Reynolds number. The improved lumped element method achieves good simulation realism at high computational speed to simulate deformable object. Real-time rendering and interaction of vessel deformation, self collision, and surface tearing has been realized and demonstrated in a virtual experiment.

This paper presents a parallel implementation of an algorithm for automatic segmentation of white matter fibers from tractography data. We execute the algorithm in parallel using a high-end video card with a Graphics Processing Unit (GPU) as a computation accelerator, using the CUDA language. By exploiting the parallelism and the properties of the memory hierarchy available on the GPU, we obtain a speedup in execution time of 33.6 with respect to an optimized sequential version of the algorithm written in C, and of 240 with respect to the original Python/C++ implementation. The execution time is reduced from more than two hours to only 35 seconds for a subject dataset of 800,000 fibers, thus enabling applications that use interactive segmentation and visualization of small to medium-sized tractography datasets.

There are multiple diseases that cause intermittent claudication, including lumber spinal canal stenosis (LSS) and peripheral arterial disease (PAD). LSS is categorized on the basis of the diseased part: L4 and L5. The medical treatment for these groups is totally different and the differentiation is important. With this in mind, we examined walking-motion data for patients and derived several features for the differentiation in previous studies. However, these features were not specialized for classification, and there is no guarantee that the features are effective for real differentiation. The present study investigates the possibility of differentiation by gait analysis, via use of an L1-regularized support vector machine (SVM). An L1-regularized SVM can execute both classification and feature selections simultaneously. On the basis of this method, our paper presents the methodology for classifying the underlying disease of the intermittent claudication with an accuracy of 79.7%. In addition, new effective features for the differentiation are extracted. © 2013 IEEE.

This paper describes an experimental setup designed for sensing the luminescent light coming from an organic plastic scintillator stimulated with ionizing radiation. This device is intended to be a part of a complete dosimeter system for characterization of small radiation fields which is the project of the doctoral thesis of the medical physicist at the Radiation Oncology facility of Hospital San Vicente Fundación in conjunction with the Universidad de Antioquia of Medellín Colombia. Some preliminary results predict a good performance of the unit, but further studies must be conducted in order to have a completed evaluation of the system. This is the first step in the development of an accuracy tool for measurement of non-standard fields in the Radiotherapy or Radiosurgery processes.

The fetal electrocardiogram (fECG) obtained from the abdominal signals, to monitor the wellbeing of the fetus, is a weak signal, recorded by placing electrodes on the maternal abdomen surface. When recording the abdominal fECG, the main problem is to separate the fECG from the background noise, including the maternal electrocardiogram (mECG) and/or the power line interference (PLI), this leading to an improved fECG signal to noise ratio (SNR). This paper proposes and evaluates three types of recording configurations, having different reference location, and analyzes the performance of each recording setup, based on the corresponding SNRs, quantitatively evaluated. The fECG extraction is carried out in order to evaluate the performance of each proposed configuration.

Time-frequency decompositions (TFDs) are well known techniques that permit to extract useful information or features from EEG signals, being necessary to distinguish between irrelevant information and the features effectively representing the subjacent physiological phenomena, according to some evaluation measure. This work introduces a new method to obtain relevant features extracted from time-frequency plane for epileptic EEG signals. Particularly, EEG features are extracted by common spectral methods such as short time Fourier transform (STFT), wavelets transform and Empirical Mode Decomposition (EMD). Then, each method is evaluated by Stochastic Relevance Analysis (SRA) that is further used for EEG classification and channel selection. The classification measures are carried out based on the performance of the k-NN classifier, while the channels selected are validated by visual inspection and topographic scalp map. The study uses real and multi-channel EEG data and all the experiments have been supervised by an expert neurologist. Results obtained in this paper show that SRA is a good alternative for automatic seizure detection and also opens the possibility of formulating new criteria to select, classify or analyze abnormal EEG channels.

Common Spatial Pattern (CSP) is a state-of-the-art method for feature extraction in Brain-Computer Interface (BCI) systems. However it is designed for 2-class BCI classification problems. Current extensions of this method to multiple classes based on subspace union and covariance matrix similarity do not provide a high performance. This paper presents a new approach to solving multi-class BCI classification problems by forming a subspace resembled from original subspaces and the proposed method for this approach is called Approximation-based Common Principal Component (ACPC). We perform experiments on Dataset 2a used in BCI Competition IV to evaluate the proposed method. This dataset was designed for motor imagery classification with 4 classes. Preliminary experiments show that the proposed ACPC feature extraction method when combining with Support Vector Machines outperforms CSP-based feature extraction methods on the experimental dataset.

This paper explores how noise can improve prediction accuracy of the Event-Related Potential (ERP) based on P300 signals. We propose an array of ICA-Based P300 processing systems with additive white Gaussian noise. The array system attains maximum accuracy when noise intensity is not zero and thus the system shows the stochastic resonance effect. The prediction accuracy increases as the number of stages of the array increases. Experimental results show that increasing the array size with the proper amount of noise can improve the accuracy of the original P300 signal detection using ICA-based subspace projection technique.

Real-time spiral phase contrast magnetic resonance imaging (MRI) is capable of non-invasively measuring the stroke volume associated with each individual heartbeat. The quality of these measurements depends on how good the segmentation of the interface between aortic wall and lumen is. Such process is hampered by the low-resolution and low-contrast nature of real-time images. Image segmentation using traditional techniques has proven not robust. This paper presents a novel model-based approach, which is capable of very accurately segmenting aortic flow. Instead of attempting to achieve a millimetrically-accurate segmentation of the wall-lumen interface, the proposed algorithm focuses on separating the aortic flow from neighboring flows. This provides robustness, even when this interface is not visually distinguishable. The proposed segmentation takes real-time MRI one step further towards becoming the non-invasive gold standard for assessment of stroke volume variability.

Age-related macular degeneration (AMD) is the leading cause of blindness in developed countries. Subretinal fluid (SRF) and sub-retinal pigment epithelium (sub-RPE) fluid are signs of AMD and can be detected in optical coherence tomography images. However, manual detection and segmentation of SRFs and sub-RPE fluids are laborious and time consuming. In this paper, a novel pipeline is proposed for automatic detection of SRFs and sub-RPE fluids. First, top and bottom layers of retina are segmented using a graph cut method. Then, a Split Bregman-based segmentation method is used to segment dark regions between layers. These segmented regions are considered as potential fluid candidates, on which a set of features are generated. After that, a random forest classifier is trained to distinguish between the true fluid regions from the falsely detected fluid regions. This method shows reasonable performance in a leave-one-out evaluation using a dataset from 21 patients.

This paper outlines a probabilistic neural network developed on the basis of time-series discriminant component analysis (TSDCA) that can be used to classify high-dimensional time-series patterns. TSDCA involves the compression of high-dimensional time series into a lower-dimensional space using a set of orthogonal transformations and the calculation of posterior probabilities based on a continuous-density hidden Markov model that incorporates a Gaussian mixture model expressed in the reduced-dimensional space. The analysis can be incorporated into a neural network so that parameters can be obtained appropriately as network coefficients according to backpropagation-through-time-based training algorithm. The network is considered to enable high-accuracy classification of high-dimensional time-series patterns and to reduce the computation time taken for network training. In the experiments conducted during the study, the validity of the proposed network was demonstrated for EEG signals.

The purpose of this paper is to determine whether electroencephalograpy (EEG) can be used as a tool for hearing impairment tests such as hearing screening. For this study, we recorded EEG responses to two syllables, /a/ and /u/, in Korean from three subjects at Gwangju Institute of Science and Technology. The ultimate goal of this study is to classify speech sound data regardless of their size using EEG; however, as an initial stage of the study, we classified only two different speech syllables using Gaussian hidden markov model. The result of this study shows a possibility that EEG could be used for hearing screening and other diagnostic tools related to speech perception.

Traditional 2-class Motor Imagery (MI) Electroencephalography (EEG) classification approaches like Common Spatial Pattern (CSP) and Support Vector Machine (SVM) usually underperform when processing stroke patients' rehabilitation EEG which are flooded with unknown irregular patterns. In this paper, the classical CSP-SVM schema is improved and a feature learning method based on Gaussian Mixture Model (GMM) is utilized for depicting patients' imagery EEG distribution features. We apply the proposed modeling program in two different modules of our online BCI-FES rehabilitation platform and achieve a relatively higher discrimination accuracy. Sufficient observations and test cases on patients' MI data sets have been implemented for validating the GMM model. The results also reveal some working mechanisms and recovery appearances of impaired cortex during the rehabilitation training period.

This paper presents the first attempt to quantify the individual performance of the subject and of the computer agent on a closed loop Reinforcement Learning Brain Machine Interface (RLBMI). The distinctive feature of the RLBMI architecture is the co-adaptation of two systems (a BMI decoder in agent and a BMI user in environment). In this work, an agent implemented using Q-learning via kernel temporal difference (KTD)(λ) decodes the neural states of a monkey and transforms them into action directions of a robotic arm. We analyze how each participant influences the overall performance both in successful and missed trials by visualizing states, corresponding action value Q, and resulting actions in two-dimensional space. With the proposed methodology, we can observe how the decoder effectively learns a good state to action mapping, and how neural states affect the prediction performance.

Bio-robots based on brain computer interface (BCI) suffer from the lack of considering the characteristic of the animals in navigation. This paper proposed a new method for bio-robots' automatic navigation combining the reward generating algorithm base on Reinforcement Learning (RL) with the learning intelligence of animals together. Given the graded electrical reward, the animal e.g. the rat, intends to seek the maximum reward while exploring an unknown environment. Since the rat has excellent spatial recognition, the rat-robot and the RL algorithm can convergent to an optimal route by co-learning. This work has significant inspiration for the practical development of bio-robots' navigation with hybrid intelligence.

Biomedical data analytic system has played an important role in doing the clinical diagnosis for several decades. Today, it is an emerging research area of analyzing these big data to make decision support for physicians. This paper presents a parallelized web-based tool with cloud computing service architecture to analyze the epilepsy. There are many modern analytic functions which are wavelet transform, genetic algorithm (GA), and support vector machine (SVM) cascaded in the system. To demonstrate the effectiveness of the system, it has been verified by two kinds of electroencephalography (EEG) data, which are short term EEG and long term EEG. The results reveal that our approach achieves the total classification accuracy higher than 90%. In addition, the entire training time accelerate about 4.66 times and prediction time is also meet requirements in real time.

In extracellular neural recording experiments, spike detection is an important step for information decoding of neuronal activities. An ASIC implementation of detection algorithms can provide substantial data-rate reduction and facilitate wireless operations. In this paper, we present a 16-channel neural spike detection ASIC. The chip takes raw data as inputs, and outputs three data streams simultaneously: field potentials down sampled at 1.25 KHz, band-pass filtered neural data, and spiking probability maps sampled at 40 KHz. The functionality and the performance of the chip have been verified in both in-vivo and benchtop experiments. Fabricated in a 0.13 µm CMOS process, the chip has a peak power dissipation of 85 µW per channel and achieves a data-rate reduction of 98.44%.

Manually induced artefacts, like caliper marks and anatomical labels, render an ultrasound (US) image incapable of being subjected to further processes of Computer Aided Diagnosis (CAD). In this paper, we propose a technique to remove these artefacts and restore the image as accurately as possible. The technique finds application as a pre-processing step when developing unsupervised segmentation algorithms for US images that deal with automatic estimation of the number of segments and clustering. The novelty of the algorithm lies in the image processing pipeline chosen to automatically identify the artefacts and is developed based on the histogram properties of the artefacts. The algorithm was able to successfully restore the images to a high quality when it was executed on a dataset of 18 US images of the thyroid gland on which the artefacts were induced manually by a doctor. Further experiments on an additional dataset of 10 unmarked US images of the thyroid gland on which the artefacts were simulated using Matlab showed that the restored images were again of high quality with a PSNR > 38dB and free of any manually induced artefacts.

Seizure detection from electroencephalogram (EEG) plays an important role for epilepsy therapy. Due to the diversity of seizure EEG patterns between different individuals, multiple features are necessary for high accuracy since a single feature could hardly encode all types of epileptiform discharges. However, a large feature set inevitably causes the increase of the computational cost. This paper proposes a boosted cascade chain to obtain both high detection performance and high computational efficiency. Sixteen features that are widely used in seizure detection are implemented. Considering the sequential characteristics of EEG signals, the features are extracted on each 1-second segment and its former three segments. Thus, a total of 64 features are used to construct a feature pool. Based on the feature pool, Real AdaBoost is used to select a group of effective features, on which weak classifiers are learned to assemble a strong classifier. The strong classifier is transformed to a cascade classifier by reordering the weak classifiers and learning a threshold for each weak classifier. The cascade classifier still has the similar classification strength to the original strong classifier. More importantly, it is able to reject easy non-seizure samples by the first a few weak classifiers in the cascade, thus high computational efficiency can be obtained. To evaluate our method, 90.6-hour EEG signals from four patients are tested. The experimental results show that our method can achieve an average accuracy of 95.31% and an average detection rate of 91.29% with the false positive rate of 4.68%. On average, only about 4 features are used. Compared with support vector machine (SVM), our method is much more efficient with the similar detection performance.

This paper presents a Matlab-based software (MathWorks inc.) called BioSigPlot for the visualization of multi-channel biomedical signals, particularly for the EEG. This tool is designed for researchers on both engineering and medicine who have to collaborate to visualize and analyze signals. It aims to provide a highly customizable interface for signal processing experimentation in order to plot several kinds of signals while integrating the common tools for physician. The main advantages compared to other existing programs are the multi-dataset displaying, the synchronization with video and the online processing. On top of that, this program uses object oriented programming, so that the interface can be controlled by both graphic controls and command lines. It can be used as EEGlab plug-in but, since it is not limited to EEG, it would be distributed separately. BioSigPlot is distributed free of charge (http://biosigplot.sourceforge.net), under the terms of GNU Public License for non-commercial use and open source development.

Transcranial Magnetic Stimulation (TMS) is a non-invasive technique that produces excitatory depolarization in the neurons located in the cerebral cortex. In order to proficiently stimulate a specific cerebral area it is of main importance the correct positioning and maintaining of the magnetic coil, while avoiding the mismatch due to subject head-coil relative movement. Hot Spot Hound is a novel robot-assisted experimental platform for enhancing TMS stimulation performance. It integrates a commercial optoelectronic neuronavigation system (E.M.S. srl, Italy) for gathering data on the relative pose (position and orientation) of subject's scalp and TMS coil, and a service robotic arm designed for human interaction (LWR system by KUKA, Germany). Besides integrating and synchronizing different platform sub-systems, we implement a control strategy to center the stimulation point and compensate for involuntary subject movements. Specifically, the proposed control maintains constant over time the homogeneous transformation matrix between the pose of the coil and of the head. The value of position stiffness has been chosen in order to assure the better compromise between coil position and orientation error, and the safety of the experimental subject. In this paper we show that Hot Spot Hound can provide coil position and orientation errors well beyond the ones achieved by manual experimenter, while assuring safety in the physical interaction with the stimulated subject.

This paper discusses a single-channel speech enhancement method for cochlear implant listeners. It is assumed that the Fourier Transform coefficients of speech and background noise have different statistical distributions. A statistical-model-based method is adopted to update the signal-to-noise ratio and estimate the background noise so that the musical noise and speech distortion induced by traditional spectral subtraction method can be effectively reduced. This enhancement method was evaluated on seven postlingually deaf Chinese cochlear implant listeners in comparison with other two speech enhancement methods. Test materials were Mandarin sentences corrupted by three different types of background noise. Experimental results showed that the proposed speech enhancement method could benefit the speech intelligibility of Chinese cochlear implant listeners. The results suggest that different noise types may affect the performance of different speech enhancement algorithms.

Dynamic imaging is challenging in MRI and acceleration techniques are usually needed to acquire dynamic scene. K-t sparse is an acceleration technique based on compressed sensing, it acquires fewer amounts of data in k-t space by pseudo random ordering of phase encodes and reconstructs dynamic scene by exploiting sparsity of k-t space in transform domain. Another recently introduced technique accelerates dynamic MRI scans by acquiring k-space data in aliased form. K-space aliasing technique uses multiple RF excitation pulses to deliberately acquire aliased k-space data. During reconstruction a simple Fourier transformation along time frames can unaliase the acquired aliased data. This paper presents a novel method to combine k-t sparse and k-space aliasing to achieve higher acceleration than each of the individual technique alone. In this particular combination, a very critical factor of compressed sensing, the ratio of the number of acquired phase encodes to the number of total phase encode (n/N) increases therefore compressed sensing component of reconstruction performs exceptionally well. Comparison of k-t sparse and the proposed technique for acceleration factors of 4, 6 and 8 is demonstrated in simulation on cardiac data.

This paper presents the results of our study on finding a lower complexity and yet a robust seizure prediction method using intracranial electroencephalogram (iEEG) recordings. We compare two classifiers: a low-complexity Adaboost and the more complex support vector machine (SVM). Adaboost is a linear classier using decision stumps, and SVM uses a nonlinear Gaussian kernel. Bipolar and/or time-differential spectral power features of different sub-bands are extracted from the iEEG signal. Adaboost is used to simultaneously classify as well as rank the features. Eliminating the low discriminating features reduces computational complexity and power consumption. The top features selected by Adaboost were also used as a feature set for SVM classification. The outputs of classifiers are regularized by applying a moving-average window and a threshold is used to generate alarms. The proposed methods were applied on 8 invasive recordings selected from the EPILEPSIAE database, the European database of EEG seizure recordings. Doublecross validation is used by separating data sets for training and optimization from testing. The key conclusion is that Adaboost performs slightly better than SVM using a reduced feature set on average with significantly less complexity resulting in a sensitivity of 77.1% (27 of 35 seizures in 873 h recordings) and a false alarm rate of 0.18 per hour.

Muscle spindles are found throughout our skeletal muscle tissue and continuously provide us with a sense of our limbs' position and motion (proprioception). This paper advances a model for generating artificial muscle spindle signals for a prosthetic limb, with the aim of one day providing amputees with a sense of feeling in their artificial limb. By utilising the Opensim biomechanical modelling package the relationship between a joint's angle and the length of surrounding muscles is estimated for a prosthetic limb. This is then applied to the established Mileusnic model to determine the associated muscle spindle firing pattern. This complete system model is then reduced to allow for a computationally efficient hardware implementation. This reduction is achieved with minimal impact on accuracy by selecting key mono-articular muscles and fitting equations to relate joint angle to muscle length. Parameter values fitting the Mileusnic model to human spindles are then proposed and validated against previously published human neural recordings. Finally, a model for fusimotor signals is also proposed based on data previously recorded from reduced animal experiments.

This paper describes a method for automated optimization of total hip arthroplasty (THA) planning incorporating joint functionalities. The optimal planning is formulated as maximum a posterior (MAP) estimation, which ensures the best-balance of joint functionalities and bone-implant spatial relations based on their statistical models derived from the training datasets prepared by an experienced surgeon. According to the performance evaluation, four of the six functionalities of the automatically optimized plans were almost equivalent to those of surgeon's plans, and two of them were improved. We consider these results showed a potential usefulness of the proposed method.

This paper presents an approach to detection and segmentation of liver tumors in 3D computed tomography (CT) images. The automatic detection of tumor can be formulized as novelty detection or two-class classification issue. The method can also be used for tumor segmentation, where each voxel is to be assigned with a correct label, either a tumor class or nontumor class. A voxel is represented with a rich feature vector that distinguishes itself from voxels in different classes. A fast learning algorithm Extreme Learning Machine (ELM) is trained as a voxel classifier. In automatic liver tumor detection, we propose and show that ELM can be trained as a one-class classifier with only healthy liver samples in training. It results in a method of tumor detection based on novelty detection. We compare it with two-class ELM. To extract the boundary of a tumor, we adopt the semi-automatic approach by randomly selecting samples in 3D space within a limited region of interest (ROI) for classifier training. Our approach is validated on a group of patients' CT data and the experiment shows good detection and encouraging segmentation results.

Epileptic seizure detection and epilepsy diagnosis based on feature extraction and classification using electroencephalography (EEG) signals is an important area of research. In this paper, we present a simple and effective approach based on signal decomposition, using a novel variation of the Empirical Mode Decomposition called Empirical Mode Decomposition-Modified Peak Selection (EMD-MPS). EMD-MPS allows time-scale based de-trending of signals, allowing signals to be separated directly into a de-trended component, and a trend, according to a frequency separation criterion. Features are extracted from the decomposed components, and a simple classifier, namely the 1-NN classifier is used for three classification tasks. The technique is tested on a publicly available EEG database, and a classification accuracy of 99% for epilepsy diagnosis task, and 100% and 98.2% for two seizure detection tasks is obtained. These results are better than, or comparable to previous results using the same EEG database, but have been obtained with a simpler and computationally fast signal analysis and classification method.

In this paper, we propose a rectangular tetrahedral adaptive mesh based corotated finite element model for interactive soft tissue simulation. Our approach consists of several computation reduction techniques. They are as follows: 1) an efficient calculation approach for computing internal forces of nodes of elastic objects to take advantage of the rectangularity of the tetrahedral adaptive mesh; 2) fast shape matching approach by using a new scaling of polar decomposition; 3) an approach for the reduction of the number of times of shape matching by using the hierarchical structure. We implemented the approach into our surgery simulator and compared the accuracy of the deformation and the computation time among 1) proposed approach, 2) L-FE), and 3) NL-FEM. Finally, we show the effectiveness of our proposed approach.

This paper presents the classification of motor tasks, using surface electromyography (sEMG) to control a virtual prosthetic hand for rehabilitation of amputees. Two types of classifiers are compared: k-Nearest Neighbor (k-NN) and Bayesian (Discriminant Analysis). Motor tasks are divided into four groups correlated. The volunteers were people without amputation and several analyzes of each of the signals were conducted. The online simulations use the sliding window technique and for feature extraction RMS (Root Mean Square), VAR (Variance) and WL (Waveform Length) values were used. A model is proposed for reclassification using cross-validation in order to validate the classification, and a visualization in Sammon Maps is provided in order to observe the separation of the classes for each set of motor tasks. Finally, the proposed method can be implemented in a computer interface providing a visual feedback through an virtual hand prosthetic developed in Visual C++ and MATLAB commands.

The electrocardiogram (ECG) is being commonly used as a diagnostic tool to distinguish different types of atrial tachyarrhythmias. The inherent complexity and mechanistic and clinical inter-relationships often brings about diagnostic difficulties to treating physicians and primary health care professionals creating frequent misdiagnoses and cross classifications using visual criteria. The current paper presents a methodology for ECG based pattern analysis for detection of atrial flutter, atrial fibrillation and normal sinus rhythm beats. ECG is an inherently non-linear and non-stationary signal; its variation may contain indicators of current disease, or warnings about impending cardiac diseases. Routinely used time domain and frequency domain methods will not be able to capture the hidden information present in the ECG beats. In the present study, we have used non-linear features of higher order spectra (HOS) to differentiate the normal, atrial fibrillation and atrial flutter ECG beats. The bispectrum features were subjected to independent component analysis (ICA) for data reduction. The ICA coefficients were subsequently subjected to K-nearest-neighbor (KNN), classification and regression tree (CART) and neural network (NN) classifiers to evaluate the best automated classifier. We have obtained an average accuracy of 97.65%, sensitivity and specificity of 98.75% and 99.53% respectively using ten-fold cross validation. Overall, the results show that application of higher order spectra statistics is useful for the classification of atrial tachyarrhythmias with reasonably high accuracies. Further validation of the proposed technique will yield acceptable results for clinical implementation.

Studies by Rechtschaffen and Kales (R&K), rely on 30-sec epochs to score sleep stages. In this paper, we introduce a new approach based on three consecutive and non-consecutive 6-sec sub-epochs for the detection of the wake stage and stage 1 sleep. The Relative Spectral Energy Band (RSEB) is used as a feature extraction from the electroencephalographic (EEG) signal. Spectral estimation is performed using non-parametric and parametric methods. We then compared the performance of the conventional 30-sec epochs with the three consecutive and non-consecutive 6-sec epochs. The outcomes of this study showed that while the accuracy varies between subjects, the non-parametric method proved to be more effective with stage 1 sleep detection and the parametric method was more effective for wake stage detection. The non-consecutive sub-epoch method was more effective and consecutive method was least effective in non-parametric stage 1 detection. Alternatively, the 30-second epoch method was most effective for parametric wake stage detection.

This paper describes a robust method of Instantaneous Heart Rate (IHR) and R-peak detection from noisy electrocardiogram (ECG) signals. Generally, the IHR is calculated from the R-wave interval. Then, the R-waves are extracted from the ECG using a threshold. However, in wearable bio-signal monitoring systems, noise increases the incidence of misdetection and false detection of R-peaks. To prevent incorrect detection, we introduce a short-term autocorrelation (STAC) technique and a small-window autocorrelation (SWAC) technique, which leverages the similarity of QRS complex waveforms. Simulation results show that the proposed method improves the noise tolerance of R-peak detection.

Optic disc segmentation from retinal fundus image is a fundamental but important step in many applications such as automated glaucoma diagnosis. Very often, one method might work well on many images but fail on some other images and it is difficult to have a single method or model to cover all scenarios. Therefore, it is important to combine results from several methods to minimize the risk of failure. For this purpose, this paper computes confidence scores for three methods and combine their results for an optimal one. The experimental results show that the combined result from three methods is better than the results by any individual method. It reduces the mean overlapping error by 7.4% relatively compared with best individual method. Simultaneously, the number of failed cases with large overlapping errors is also greatly reduced. This is important to enhance the clinical deployment of the automated disc segmentation.

Recently effective connectivity studies have gained significant attention among the neuroscience community as Electroencephalography (EEG) data with a high time resolution can give us a wider understanding of the information flow within the brain. Among other tools used in effective connectivity analysis Granger Causality (GC) has found a prominent place. The GC analysis, based on strictly causal multivariate autoregressive (MVAR) models does not account for the instantaneous interactions among the sources. If instantaneous interactions are present, GC based on strictly causal MVAR will lead to erroneous conclusions on the underlying information flow. Thus, the work presented in this paper applies an extended MVAR (eMVAR) model that accounts for the zero lag interactions. We propose a constrained adaptive Kalman filter (CAKF) approach for the eMVAR model identification and demonstrate that this approach performs better than the short time windowing-based adaptive estimation when applied to information flow analysis.