Motor unit filter prelearning strategies for decomposition of compound muscle action potentials in high-density surface electromyograms.

We analyze the efficiency of motor unit (MU) filter prelearning from high-density surface electromyographic (HDEMG) recordings of voluntary muscle contractions in the identification of the motor unit firing patterns during elicited muscle contractions. Motor unit filters are assessed from 10 s long low level isometric voluntary contractions by gradient-based optimization of three different cost functions and then applied to synthetic HDEMG recordings of elicited muscle contractions with dispersion of motor unit firings ranging from 13 ms to 1 ms. We demonstrate that the number of identified MUs and the precision of MU identification depend significantly on the selected cost function. Regardless the selected cost function and MU synchronization level, the median precision of motor unit identification in elicited contraction is ≥ 95 % and is comparable to the precision in voluntary contractions. On the other hand, median miss rate increases significantly from < 1 % to ~ 3 % with the tested level of MU synchronization.Clinical Relevance-The identification of MU firings from HDEMG in elicited muscle contractions provides a new tool for in vivo investigation of motor excitability in humans.

Tutorial: Analysis of motor unit discharge characteristics from high-density surface EMG signals.

Recent work demonstrated that it is possible to identify motor unit discharge times from high-density surface EMG (HDEMG) decomposition. Since then, the number of studies that use HDEMG decomposition for motor unit investigations has increased considerably. Although HDEMG decomposition is a semi-automatic process, the analysis and interpretation of the motor unit pulse trains requires a thorough inspection of the output of the decomposition result. Here, we report guidelines to perform an accurate extraction of motor unit discharge times and interpretation of the signals. This tutorial includes a discussion of the differences between the extraction of global EMG signal features versus the identification of motor unit activity for physiological investigations followed by a comprehensive guide on how to acquire, inspect, and decompose HDEMG signals, and robust extraction of motor unit discharge characteristics.

On the Impact of Muscle Shortening on Non-Negative Matrix Factorization of Dynamic Surface Electromyograms.

We analyzed the sensitivity of Non-negative Matrix Factorization (NMF) of dynamic surface electromyograms (EMG) to muscle shortening. We first identified Motor unit action potentials (MUAPs) by decomposing experimentally recorded EMG signals during slow shortening of biceps brachii muscle in five young healthy males. We then used these MUAPs to generate different synthetic EMG signals with different muscle shortening and excitation profiles. Afterwards, we applied NMF to the synthetic EMG signals and calculated Pearson correlation coefficient (CC) between the extracted NMF components and a) muscle shortening and b) muscle excitation profiles. The results demonstrated good match between NMF components and muscle excitation profiles, but only when the muscle excitation level changed for at least 10 % during the muscle shortening. During constant muscle excitation, the resulting NMF components correlated significantly with the muscle shortening profiles. These results demonstrate that NMF components reflect not only the muscle excitation profiles but also muscle shortening profiles. Therefore, the results of NMF analysis of dynamic EMG signals need to be interpreted with caution.

On the impact of spike segmentation on motor unit identification in dynamic surface electromyograms.

We discuss the adaptation of preexisting Convolution Kernel Compensation (CKC) surface electromyogram (EMG) decomposition technique to dynamic muscle contractions. In particular, three different algorithms for segmentation of motor unit (MU) spike trains into MU firings are discussed and mutually compared on synthetic dynamic surface EMG. The first segmentation algorithm employs a priori knowledge of the regularity of MU firings. The second one builds on K-means classification of MU spikes, whereas the third one combines both the regularity of MU firings and the previously introduced Pulse-to-Noise Ratio (PNR). On average, 5.5 ± 0.6 MUs were identified with sensitivity of 88.4 % ± 17.0 %, 83.8 % ± 16.7 % and 90.7 % ± 15.1 % for the first, the second and the third segmentation algorithm, respectively, demonstrating the feasibility of MU identification in moderate dynamic muscle contractions. In our tests, the third segmentation approach demonstrated superior accuracy in MU identification.

Simulations of high-density surface electromyograms in dynamic muscle contractions.

We describe the extension of pre-existing cylindrical volume conductor model to synthetic high-density surface electromyograms (hdEMG), simulated during dynamic contractions of fusiform skeletal muscles. Its modular structure comprises two main parts. First, dynamic changes of motor unit action potentials (MUAPs) during 36 discrete steps of muscle shortening are simulated. Second, the increase in depth of simulated motor units (MUs) due to shortening and thickening of muscle fibers is simulated. MU firing patterns are generated with the model proposed by Fuglevand et al. and convolved with simulated MUAPs. In this way, the hdEMG simulator can be used to generate dynamic hdEMG of arbitrary muscle shortening, thickening and excitation profiles. In order to demonstrate the value of the aforementioned simulator we independently analyze the impact of muscle shortening and muscle thickening on MU identification by Convolution Kernel Compensation (CKC) technique.

Blind source identification from the multichannel surface electromyogram.

The spinal circuitries combine the information flow from the supraspinal centers with the afferent input to generate the neural codes that drive the human skeletal muscles. The muscles transform the neural drive they receive from alpha motor neurons into motor unit action potentials (electrical activity) and force. Thus, the output of the spinal cord circuitries can be examined noninvasively by measuring the electrical activity of skeletal muscles at the surface of the skin i.e. the surface electromyogram (EMG). The recorded multi-muscle EMG activity pattern is generated by mixing processes of neural sources that need to be identified from the recorded signals themselves, with minimal or no a priori information available. Recently, multichannel source separation techniques that rely minimally on a priori knowledge of the mixing process have been developed and successfully applied to surface EMG. They act at different scales of information extraction to identify: (a) the activation signals shared by synergistic skeletal muscles, (b) the specific neural activation of individual muscles, separating it from that of nearby muscles i.e. from crosstalk, and (c) the spike trains of the active motor neurons. This review discusses the assumptions made by these methods, the challenges and limitations, as well as examples of their current applications.

Accurate identification of motor unit discharge patterns from high-density surface EMG and validation with a novel signal-based performance metric.

OBJECTIVE: A signal-based metric for assessment of accuracy of motor unit (MU) identification from high-density surface electromyograms (EMG) is introduced. This metric, so-called pulse-to-noise-ratio (PNR), is computationally efficient, does not require any additional experimental costs and can be applied to every MU that is identified by the previously developed convolution kernel compensation technique. APPROACH: The analytical derivation of the newly introduced metric is provided, along with its extensive experimental validation on both synthetic and experimental surface EMG signals with signal-to-noise ratios ranging from 0 to 20 dB and muscle contraction forces from 5% to 70% of the maximum voluntary contraction. MAIN RESULTS: In all the experimental and simulated signals, the newly introduced metric correlated significantly with both sensitivity and false alarm rate in identification of MU discharges. Practically all the MUs with PNR > 30 dB exhibited sensitivity >90% and false alarm rates <2%. Therefore, a threshold of 30 dB in PNR can be used as a simple method for selecting only reliably decomposed units. SIGNIFICANCE: The newly introduced metric is considered a robust and reliable indicator of accuracy of MU identification. The study also shows that high-density surface EMG can be reliably decomposed at contraction forces as high as 70% of the maximum.

Non-invasive characterization of motor unit behaviour in pathological tremor.

This paper presents the fully automatic identification of motor unit spike trains from high-density surface electromyograms (EMG) in pathological tremor. First, a mathematical derivation is provided to theoretically prove the possibility of decomposing noise-free high-density surface EMG signals into motor unit spike trains with high correlation, which are typical of tremor contractions. Further, the proposed decomposition method is tested on simulated signals with different levels of noise and on experimental signals from 14 tremor-affected patients. In the case of simulated tremor with central frequency ranging from 5 Hz to 11 Hz and signal-to-noise ratio of 20 dB, the method identified ∼8 motor units per contraction with sensitivity in spike timing identification ≥ 95% and false alarm and miss rates ≤ 5%. In experimental signals, the number of identified motor units varied substantially (range 0-21) across patients and contraction types, as expected. The behaviour of the identified motor units was consistent with previous data obtained by intramuscular EMG decomposition. These results demonstrate for the first time the possibility of a fully non-invasive investigation of motor unit behaviour in tremor-affected patients. The method provides a new means for physiological investigations of pathological tremor.

Changes in motor unit characteristics after eccentric elbow flexor exercise.

Morphological evidence suggests that fast-twitch fibers are prone to disruption of their membrane structures by eccentric exercise. However, it is unclear how this is reflected in the discharge rate and action potential propagation of individual motor units, especially at high contraction levels. High-density surface electromyograms were recorded from biceps brachii muscle and decomposed to individual motor unit action potentials at isometric contraction levels between 10% and 75% of maximal voluntary contraction (MVC) before intermittent maximal elbow flexor eccentric exercise, and two hours (2H), two days (2D) and four days (4D) post-exercise. Maximal voluntary force decreased by 21.3±5.6% 2H and by 12.6±11.1% 2D post-exercise. Motor unit discharge rate increased and mean muscle fiber conduction velocity decreased, at the highest isometric contraction levels only (50% and 75% of MVC) at 2H post-exercise. These results indicate that eccentric exercise can disturb the function of motor units active at high contraction levels in the early stages after exercise, which seems to be compensated by the central nervous system with an increase in neural drive during submaximal isometric contractions.

Accuracy assessment of CKC high-density surface EMG decomposition in biceps femoris muscle.

The aim of this study was to assess the accuracy of the convolution kernel compensation (CKC) method in decomposing high-density surface EMG (HDsEMG) signals from the pennate biceps femoris long-head muscle. Although the CKC method has already been thoroughly assessed in parallel-fibered muscles, there are several factors that could hinder its performance in pennate muscles. Namely, HDsEMG signals from pennate and parallel-fibered muscles differ considerably in terms of the number of detectable motor units (MUs) and the spatial distribution of the motor-unit action potentials (MUAPs). In this study, monopolar surface EMG signals were recorded from five normal subjects during low-force voluntary isometric contractions using a 92-channel electrode grid with 8 mm inter-electrode distances. Intramuscular EMG (iEMG) signals were recorded concurrently using monopolar needles. The HDsEMG and iEMG signals were independently decomposed into MUAP trains, and the iEMG results were verified using a rigorous a posteriori statistical analysis. HDsEMG decomposition identified from 2 to 30 MUAP trains per contraction. 3 ± 2 of these trains were also reliably detected by iEMG decomposition. The measured CKC decomposition accuracy of these common trains over a selected 10 s interval was 91.5 ± 5.8%. The other trains were not assessed. The significant factors that affected CKC decomposition accuracy were the number of HDsEMG channels that were free of technical artifact and the distinguishability of the MUAPs in the HDsEMG signal (P < 0.05). These results show that the CKC method reliably identifies at least a subset of MUAP trains in HDsEMG signals from low force contractions in pennate muscles.

Noninvasive analysis of motor unit behavior in pathological tremor.

A robust surface EMG decomposition tool, referred to as tremor-optimized Convolution Kernel Compensation (CKC) technique, is described. This technique modifies and extends the previously published CKC method in order to circumvent the typical assumption on regularity and asynchrony of motor unit firings in normal condition and adapt to the discharge patterns in pathological tremor. The results on synthetic and experimental surface EMG signals demonstrate high performance of decomposition. In the case of simulated surface EMG with 20 dB SNR, excitation level of 20% maximum voluntary contraction (MVC) and simulated tremor frequency of 8 Hz, the newly proposed method identified 8 ± 2 motor units with sensitivity of motor unit discharge identification ≥ 95 % and false alarm and miss rates ≤ 5%. The performance worsened with increasing noise power, with 5 ± 2 motor units identified at 10 dB SNR and 3 ± 1 at 0 dB SNR. In 24 recordings of high-density surface EMG signals from four tremor-affected patients, the modified CKC technique identified 134 motor units (6 ± 4 motor units per contraction).

A modelling study on transmission of the central oscillator in tremor by a motor neuron pool.

In spite of decades of intense research, pathological tremors still constitute unknown disorders. This study addresses, based on a multi-scale model, the behavior of an entire pool of motor neurons in tremor, under the hypothesis that tremor is an oscillation of central origin commonly projected to all motor neurons that innervate a muscle. Our results show that under such conditions both paired discharges and enhanced motor neuron synchronization, two of the characteristic landmarks of tremor, emerge. Moreover, coherence and correlation analyses suggest that the central tremor oscillator is transmitted linearly by the motor neuron pool given that a small set (7 or 8) of motor neurons are sampled.

On the selection of the cost function for gradient-based decomposition of surface electromyograms.

Recently, gradient Convolution Kernel Compensation method was introduced for blind assessment of sparse pulse sequences (PS) out of their convolutive mixtures. This method employs multichannel recordings, is fully automatic and is minimally biased by assumptions about underlying mixing process. In the first step, the unknown mixing channels (convolution kernels) are compensated, whereas in the second step the gradient algorithm is used to blindly optimize the estimated PSs. This paper discusses the selection of the cost functions for aforementioned gradient-based optimization and provides analytical framework for their mutual comparison. Theoretical derivations are validated on both synthetic signals with random mixing matrices and experimental surface electromyograms from abductor pollicis brevis muscle. The analytical derivations agree very well with the results obtained from numerical simulations and establish theoretical guidelines for developing new gradient-based decomposition methods.

Building interactive virtual environments for simulated training in medicine using VRML and Java/JavaScript.

Medicine is a difficult thing to learn. Experimenting with real patients should not be the only option; simulation deserves a special attention here. Virtual Reality Modelling Language (VRML) as a tool for building virtual objects and scenes has a good record of educational applications in medicine, especially for static and animated visualisations of body parts and organs. However, to create computer simulations resembling situations in real environments the required level of interactivity and dynamics is difficult to achieve. In the present paper we describe some approaches and techniques which we used to push the limits of the current VRML technology further toward dynamic 3D representation of virtual environments (VEs). Our demonstration is based on the implementation of a virtual baby model, whose vital signs can be controlled from an external Java application. The main contributions of this work are: (a) outline and evaluation of the three-level VRML/Java implementation of the dynamic virtual environment, (b) proposal for a modified VRML Timesensor node, which greatly improves the overall control of system performance, and (c) architecture of the prototype distributed virtual environment for training in neonatal resuscitation comprising the interactive virtual newborn, active bedside monitor for vital signs and full 3D representation of the surgery room.

Correlation-based decomposition of surface electromyograms at low contraction forces.

The paper studies a surface electromyogram (SEMG) decomposition technique suitable for identification of complete motor unit (MU) firing patterns and their motor unit action potentials (MUAPs) during low-level isometric voluntary muscle contractions. The algorithm was based on a correlation matrix of measurements, assumed unsynchronised (uncorrelated) MU firings, exhibited a very low computational complexity and resolved the superimposition of MUAPs. A separation index was defined that identified the time instants of an MU's activation and was eventually used for reconstruction of a complete MU innervation pulse train. In contrast with other decomposition techniques, the proposed approach worked well also when the number of active MUs was slightly underestimated, if the MU firing patterns partly overlapped and if the measurements were noisy. The results on synthetic SEMG show 100% accuracy in the detection of innervation pulses down to a signal-to-noise ratio (SNR) of 10 dB, and 93+/-4.6% (mean+/-standard deviation) accuracy with 0 dB additive noise. In the case of real SEMG, recorded with an array of 61 electrodes from biceps brachii of five subjects at 10% maximum voluntary contraction, seven active MUs with a mean firing rate of 14.1 Hz were identified on average.

Optical triple sensor for measuring pH, oxygen and carbon dioxide.

A triple sensor unit consisting of opto-chemical sensors for measurement of pH, oxygen and carbon dioxide in bioreactors is presented. The pH and the CO2 sensor are based on the color change of a pH-sensitive dye immobilized on a polymeric support. The resulting changes in absorption are monitored through optical fibers. The oxygen sensor is based on the quenching of the fluorescence of a metal-organic dye. All three sensors are fully LED compatible. The sensitive membranes consist of plastic films and can be stored and replaced conveniently. The sensors are sterilizable with hydrogen peroxide and ethanol. In addition, the pH sensor is steam sterilizable. Accuracy, resolution and reproducibility fulfill the requirements for use in biotechnological applications. Calibration procedures for each sensor are presented. The working principle and the performance of all three sensors are described, with particular emphasis given to their application in bioreactors.