RESUMEN

OBJECTIVE: Spike sorting of muscular and neural recordings requires separating action potentials that overlap in time (superimposed action potentials (APs)). We propose a new algorithm for resolving superimposed action potentials, and we test it on intramuscular EMG (iEMG) and intracortical recordings. METHODS: Discrete-time shifts of the involved APs are first selected based on a heuristic extension of the peel-off algorithm. Then, the time shifts that provide the minimal residual Euclidean norm are identified (Discrete Brute force Correlation (DBC)). The optimal continuous-time shifts are then estimated (High-Resolution BC (HRBC)). In Fusion HRBC (FHRBC), two other cost functions are used. A parallel implementation of the DBC and HRBC algorithms was developed. The performance of the algorithms was assessed on 11,000 simulated iEMG and 14,000 neural recording superpositions, including two to eight APs, and eight experimental iEMG signals containing four to eleven active motor units. The performance of the proposed algorithms was compared with that of the Branch-and-Bound (BB) algorithm using the Rank-Product (RP) method in terms of accuracy and efficiency. RESULTS: The average accuracy of the DBC, HRBC and FHRBC methods on the entire simulated datasets was 92.16±17.70, 93.65±16.89, and 94.90±15.15 (%). The DBC algorithm outperformed the other algorithms based on the RP method. The average accuracy and running time of the DBC algorithm on 10.5 ms superimposed spikes of the experimental signals were 92.1±21.7 (%) and 2.3±15.3 (ms). CONCLUSION AND SIGNIFICANCE: The proposed algorithm is promising for real-time neural decoding, a central problem in neural and muscular decoding and interfacing.

Asunto(s)

RESUMEN

It is necessary to decompose the intra-muscular EMG signal to extract motor unit action potential (MUAP) waveforms and firing times. Some algorithms were proposed in the literature to resolve superimposed MUAPs, including Peel-Off (PO), branch and bound (BB), genetic algorithm (GA), and particle swarm optimization (PSO). This study aimed to compare these algorithms in terms of overall accuracy and running time. Two sets of two-to-five MUAP templates (set1: a wide range of energies, and set2: a high degree of similarity) were used. Such templates were time-shifted, and white Gaussian noise was added. A total of 1000 superpositions were simulated for each template and were resolved using PO (also, POI: interpolated PO), BB, GA, and PSO algorithms. The generalized estimating equation was used to identify which method significantly outperformed, while the overall rank product was used for overall ranking. The rankings were PSO, BB, GA, PO, and POI in the first, and BB, PSO, GA, PO, POI in the second set. The overall ranking was BB, PSO, GA, PO, and POI in the entire dataset. Although the BB algorithm is generally fast, there are cases where the BB algorithm is too slow and it is thus not suitable for real-time applications.

Asunto(s)

RESUMEN

Dyslipidemia, the disorder of lipoprotein metabolism resulting in high lipid profile, is an important modifiable risk factor for coronary heart diseases. It is associated with more than four million worldwide deaths per year. Half of the children with dyslipidemia have hyperlipidemia during adulthood, and its prediction and screening are thus critical. We designed a new dyslipidemia diagnosis system. The sample size of 725 subjects (age 14.66 ±â€¯2.61 years; 48% male; dyslipidemia prevalence of 42%) was selected by multistage random cluster sampling in Iran. Single nucleotide polymorphisms (rs1801177, rs708272, rs320, rs328, rs2066718, rs2230808, rs5880, rs5128, rs2893157, rs662799, and Apolipoprotein-E2/E3/E4), and anthropometric, life-style attributes, and family history of diseases were analyzed. A framework for classifying mixed-type data in imbalanced datasets was proposed. It included internal feature mapping and selection, re-sampling, optimized group method of data handling using convex and stochastic optimizations, a new cost function for imbalanced data and an internal validation. Its performance was assessed using hold-out and 4-foldcross-validation. Four other classifiers namely as supported vector machines, decision tree, and multilayer perceptron neural network and multiple logistic regression were also used. The average sensitivity, specificity, precision and accuracy of the proposed system were 93%, 94%, 94% and 92%, respectively in cross validation. It significantly outperformed the other classifiers and also showed excellent agreement and high correlation with the gold standard. A non-invasive economical version of the algorithm was also implemented suitable for low- and middle-income countries. It is thus a promising new tool for the prediction of dyslipidemia.

RESUMEN

BACKGROUND AND OBJECTIVE: Cochlear implants (CIs) are electronic devices restoring partial hearing to deaf individuals with profound hearing loss. In this paper, a new plug-in for traditional IIR filter-banks (FBs) is presented for cochlear implants based on wavelet neural networks (WNNs). Having provided such a plug-in for commercially available CIs, it is possible not only to use available hardware in the market but also to optimize their performance compared with the-state-of-the-art. METHODS: An online database of Dutch diphone perception was used in our study. The weights of the WNNs were tuned using particle swarm optimization (PSO) on a training set (speech-shaped noise (SSN) of 2 dB SNR), while its performance was assessed on a test set in terms of objective and composite measures in the hold-out validation framework. The cost function was defined based on the combination of mean square error (MSE), short­time objective intelligibility (STOI) criteria on the training set. Variety of performance indices were used including segmental signal- to -noise ratio (SNRseg), MSE, STOI, log-likelihood ratio (LLR), weighted spectral slope (WSS), and composite measures Csig,Cbak and Covl. Meanwhile, the following CI speech processing techniques were used for comparison: traditional FBs, dual resonance nonlinear (DRNL) and simple dual path nonlinear (SPDN) models. RESULTS: The average SNRseg, MSE, and LLR values for the WNN in the entire data set were 2.496 ±â€¯2.794, 0.086 ±â€¯0.025 and 2.323 ±â€¯0.281, respectively. The proposed method significantly improved MSE, SNR, SNRseg, LLR, Csig Cbak and Covl compared with the other three methods (repeated-measures analysis of variance (ANOVA); P < 0.05). The average running time of the proposed algorithm (written in Matlab R2013a) on the training and test sets for each consonant or vowel on an Intel dual-core 2.10 GHz CPU with 2GB of RAM was 9.91 ±â€¯0.87 (s) and 0.19 ±â€¯0.01 (s), respectively. CONCLUSIONS: The proposed algorithm is accurate and precise and is thus a promising new plug-in for traditional CIs. Although the tuned algorithm is relatively fast, it is necessary to use efficient vectorized implementations for real-time CI speech signal processing.

Asunto(s)

RESUMEN

BACKGROUND: In this paper, we propose a nonlinear minimally supervised method based on autoencoding (AEN) of EMG for myocontrol. The proposed method was tested against the state-of-the-art (SOA) control scheme using a Fitts' law approach. METHODS: Seven able-bodied subjects performed a series of target acquisition myoelectric control tasks using the AEN and SOA algorithms for controlling two degrees-of-freedom (radial/ulnar deviation and flexion/extension of the wrist), and their online performance was characterized by six metrics. RESULTS: Both methods allowed a completion rate close to 100%, however AEN outperformed SOA for all other performance metrics, e.g. it allowed to perform the tasks on average in half the time with respect to SOA. Moreover, the amount of information transferred by the proposed method in bit/s was nearly twice the throughput of SOA. CONCLUSIONS: These results show that autoencoders can map EMG signals into kinematics with the potential of providing intuitive and dexterous control of artificial limbs for amputees.

Asunto(s)

RESUMEN

OBJECTIVE: The aim of this study was to reconstruct low-quality High-density surface EMG (HDsEMG) signals, recorded with 2-D electrode arrays, using image inpainting and surface reconstruction methods. METHODS: It is common that some fraction of the electrodes may provide low-quality signals. We used variety of image inpainting methods, based on partial differential equations (PDEs), and surface reconstruction methods to reconstruct the time-averaged or instantaneous muscle activity maps of those outlier channels. Two novel reconstruction algorithms were also proposed. HDsEMG signals were recorded from the biceps femoris and brachial biceps muscles during low-to-moderate-level isometric contractions, and some of the channels (5-25%) were randomly marked as outliers. The root-mean-square error (RMSE) between the original and reconstructed maps was then calculated. RESULTS: Overall, the proposed Poisson and wave PDE outperformed the other methods (average RMSE 8.7 µVrms ± 6.1 µVrms and 7.5 µVrms ± 5.9 µVrms) for the time-averaged single-differential and monopolar map reconstruction, respectively. Biharmonic Spline, the discrete cosine transform, and the Poisson PDE outperformed the other methods for the instantaneous map reconstruction. The running time of the proposed Poisson and wave PDE methods, implemented using a Vectorization package, was 4.6 ± 5.7 ms and 0.6 ± 0.5 ms, respectively, for each signal epoch or time sample in each channel. CONCLUSION: The proposed reconstruction algorithms could be promising new tools for reconstructing muscle activity maps in real-time applications. SIGNIFICANCE: Proper reconstruction methods could recover the information of low-quality recorded channels in HDsEMG signals.

Asunto(s)

RESUMEN

Cancer is a collection of diseases that involves growing abnormal cells with the potential to invade or spread to the body. Breast cancer is the second leading cause of cancer death among women. A method for 5-year breast cancer recurrence prediction is presented in this manuscript. Clinicopathologic characteristics of 579 breast cancer patients (recurrence prevalence of 19.3%) were analyzed and discriminative features were selected using statistical feature selection methods. They were further refined by Particle Swarm Optimization (PSO) as the inputs of the classification system with ensemble learning (Bagged Decision Tree: BDT). The proper combination of selected categorical features and also the weight (importance) of the selected interval-measurement-scale features were identified by the PSO algorithm. The performance of HPBCR (hybrid predictor of breast cancer recurrence) was assessed using the holdout and 4-fold cross-validation. Three other classifiers namely as supported vector machines, DT, and multilayer perceptron neural network were used for comparison. The selected features were diagnosis age, tumor size, lymph node involvement ratio, number of involved axillary lymph nodes, progesterone receptor expression, having hormone therapy and type of surgery. The minimum sensitivity, specificity, precision and accuracy of HPBCR were 77%, 93%, 95% and 85%, respectively in the entire cross-validation folds and the hold-out test fold. HPBCR outperformed the other tested classifiers. It showed excellent agreement with the gold standard (i.e. the oncologist opinion after blood tumor marker and imaging tests, and tissue biopsy). This algorithm is thus a promising online tool for the prediction of breast cancer recurrence.

RESUMEN

A cochlear implant is an implanted electronic device used to provide a sensation of hearing to a person who is hard of hearing. The cochlear implant is often referred to as a bionic ear. This paper presents an undecimated wavelet-based speech coding strategy for cochlear implants, which gives a novel speech processing strategy. The undecimated wavelet packet transform (UWPT) is computed like the wavelet packet transform except that it does not down-sample the output at each level. The speech data used for the current study consists of 30 consonants, sampled at 16 kbps. The performance of our proposed UWPT method was compared to that of infinite impulse response (IIR) filter in terms of mean opinion score (MOS), short-time objective intelligibility (STOI) measure and segmental signal-to-noise ratio (SNR). Undecimated wavelet had better segmental SNR in about 96% of the input speech data. The MOS of the proposed method was twice in comparison with that of the IIR filter-bank. The statistical analysis revealed that the UWT-based N-of-M strategy significantly improved the MOS, STOI and segmental SNR (P < 0.001) compared with what obtained with the IIR filter-bank based strategies. The advantage of UWPT is that it is shift-invariant which gives a dense approximation to continuous wavelet transform. Thus, the information loss is minimal and that is why the UWPT performance was better than that of traditional filter-bank strategies in speech recognition tests. Results showed that the UWPT could be a promising method for speech coding in cochlear implants, although its computational complexity is higher than that of traditional filter-banks.

RESUMEN

Microalbuminuria (MA) is an independent predictor of cardiovascular and renal disease, development of overt nephropathy, and cardiovascular mortality in patients with type 2 diabetes. Detecting MA is an important screening tool to identify people with high risk of cardiovascular and kidney disease. The gold standard to detect MA is measuring 24-h urine albumin excretion. A new method for MA diagnosis is presented in this manuscript which uses clinical parameters usually monitored in type 2 diabetic patients without the need of an additional measurement of urinary albumin. We designed an expert-based fuzzy MA classifier in which rule induction was performed by particle swarm optimization. A variety of classifiers was tested. Additionally, multiple logistic regression was used for statistical feature extraction. The significant features were age, diabetic duration, body mass index and HbA1C (the average level of blood sugar over the previous 3 months, which is routinely checked every 3 months for diabetic patients). The resulting classifier was tested on a sample size of 200 patients with type 2 diabetes in a cross-sectional study. The performance of the proposed classifier was assessed using (repeated) holdout and 10-fold cross-validation. The minimum sensitivity, specificity, precision and accuracy of the proposed fuzzy classifier system with feature extraction were 95%, 85%, 84% and 92%, respectively. The proposed hybrid intelligent system outperformed other tested classifiers and showed "almost perfect agreement" with the gold standard. This algorithm is a promising new tool for screening MA in type-2 diabetic patients.

Asunto(s)

RESUMEN

Recently developed techniques allow the analysis of surface EMG in multiple locations over the skin surface (high-density surface electromyography, HDsEMG). The detected signal includes information from a greater proportion of the muscle of interest than conventional clinical EMG. However, recording with many electrodes simultaneously often implies bad-contacts, which introduce large power-line interference in the corresponding channels, and short-circuits that cause near-zero single differential signals when using gel. Such signals are called 'outliers' in data mining. In this work, outlier detection (focusing on bad contacts) is discussed for monopolar HDsEMG signals and a new method is proposed to identify 'bad' channels. The overall performance of this method was tested using the agreement rate against three experts' opinions. Three other outlier detection methods were used for comparison. The training and test sets for such methods were selected from HDsEMG signals recorded in Triceps and Biceps Brachii in the upper arm and Brachioradialis, Anconeus, and Pronator Teres in the forearm. The sensitivity and specificity of this algorithm were, respectively, 96.9 ± 6.2 and 96.4 ± 2.5 in percent in the test set (signals registered with twenty 2D electrode arrays corresponding to a total of 2322 channels), showing that this method is promising.

Asunto(s)

RESUMEN

This paper presents a density-based method to automatically decompose single-channel intramuscular electromyogram (EMG) signals into their component motor unit action potential (MUAP) trains. In contrast to most previous decomposition methods, which require pre-setting and (or) tuning of multiple parameters, the proposed method takes advantage of the data-dependent strategies in the pattern recognition procedures. In this method, outliers (superpositions) are excluded prior to classification and MUAP templates are identified by an adaptive density-based clustering procedure. MUAP trains are then identified by a novel density-based classifier that incorporates MUAP shape and discharge time information. MUAP trains are merged by a fuzzy system that incorporates expert human knowledge. Finally, superimpositions are resolved to fill the gaps in the MUAP trains. The proposed decomposition algorithm has been experimentally tested on signals from low-force (≤30% maximal) isometric contractions of the vastus medialis obliquus, vastus lateralis, biceps femoris long-head and tibialis anterior muscles. Comparison with expert manual decomposition that had been verified using a rigorous statistical analysis showed that the algorithm identified 80% of the total 229 motor unit trains with an accuracy greater than 90%. The algorithm is robust and accurate, and therefore it is a promising new tool for decomposing single-channel multi-unit signals.

Asunto(s)

RESUMEN

If electromyography (EMG) decomposition is to be a useful tool for scientific investigation, it is essential to know that the results are accurate. Because of background noise, waveform variability, motor-unit action potential (MUAP) indistinguishability, and perplexing superpositions, accuracy assessment is not straightforward. This paper presents a rigorous statistical method for assessing decomposition accuracy based only on evidence from the signal itself. The method uses statistical decision theory in a Bayesian framework to integrate all the shape- and firing-time-related information in the signal to compute an objective a posteriori measure of confidence in the accuracy of each discharge in the decomposition. The assessment is based on the estimated statistical properties of the MUAPs and noise and takes into account the relative likelihood of every other possible decomposition. The method was tested on 3 pairs of real EMG signals containing 4-7 active MUAP trains per signal that had been decomposed by a human expert. It rated 97% of the identified MUAP discharges as accurate to within ± 0.5 ms with a confidence level of 99%, and detected six decomposition errors. Cross-checking between signal pairs verified all but two of these assertions. These results demonstrate that the approach is reliable and practical for real EMG signals.

Asunto(s)

RESUMEN

The Particle Swarm Optimization (PSO) algorithm is applied to the problem of "load sharing" among muscles acting on the same joint for the purpose of estimating their individual mechanical contribution based on their EMG and on the total torque. Compared to the previously tested Interior-Reflective Newton Algorithm (IRNA), PSO is more computationally demanding. The mean square error between the experimental and reconstructed torque is similar for the two algorithms. However, IRNA requires multiple initializations and tighter constraints found by trial-and-errors for the input variables to find a suitable optimum which is not the case for PSO whose initialization is random.

Asunto(s)

RESUMEN

This paper describes an interactive computer program for decomposing EMG signals into their component motor-unit potential (MUP) trains and for averaging MUP waveforms. The program is able to handle single- or multi-channel signals recorded by needle or fine-wire electrodes during low and moderate levels of muscular contraction. It includes advanced algorithms for template matching, resolving superimpositions, and waveform averaging, as well as a convenient user interface for manually editing and verifying the results. The program also provides the ability to inspect the discharges of individual motor units more closely by subtracting out interfering activity from other MUP trains. Decomposition accuracy was assessed by cross-checking pairs of signals recorded by nearby electrodes during the same contraction. The results show that 100% accuracy can be achieved for MUPs with peak-to-peak amplitudes greater than 2.5 times the rms signal amplitude. Examples are presented to show how decomposition can be used to investigate motor-unit recruitment and discharge behavior, to study motor-unit architecture, and to detect action potential blocking in doubly innervated muscle fibers.

Asunto(s)