Pattern recognition of number gestures based on a wireless surface EMG system

Using surface electromyography (sEMG) signal for efficient recognition of hand gestures has attracted increasing attention during the last decade, with most previous work being focused on recognition of upper arm and gross hand movements and some work on the classification of individual finger movements such as finger typing tasks. However, relatively few investigations can be found in the literature for automatic classification of multiple finger movements such as finger number gestures. This paper focuses on the recognition of number gestures based on a 4-channel wireless sEMG system. We investigate the effects of three popular feature types (i.e. Hudgins’ time–domain features (TD), autocorrelation and cross-correlation coefficients (ACCC) and spectral power magnitudes (SPM)) and four popular classification algorithms (i.e. k-nearest neighbor (k-NN), linear discriminant analysis (LDA), quadratic discriminant analysis (QDA) and support vector machine (SVM)) in offline recognition. Motivated by the good performance of SVM, we further propose combining the three features and employing a new classification method, multiple kernel learning SVM (MKL-SVM). Real sEMG results from six subjects show that all combinations, except k-NN or LDA using ACCC features, can achieve above 91% average recognition accuracy, and the highest accuracy is 97.93% achieved by the proposed MKL-SVM method using the three feature combination (3F). Referring to the offline recognition results, we also implement a real-time recognition system. Our results show that all six subjects can achieve a real-time recognition accuracy higher than 90%. The number gestures are therefore promising for practical applications such as human–computer interaction (HCI).     

Training a multivariable myoelectric mapping function to estimate fatigue

The mapping index (MI) is a fatigue assessment index that uses multiple time-domain myoelectric features to train an artificial neural network (ANN) to track the progression of fatigue. This work showed that mapping functions trained using data from independent subjects and contraction conditions to yield a generalized mapping index (GMI) can assess fatigue as well as functions trained with subject and contraction-specific data to yield MI. Surface myoelectric signals were collected from nine healthy participants during isometric, cyclic and random fatiguing contractions. Two datasets were collected: one for tuning the functions and the other for testing. The performance of fatigue indices was evaluated using a newly proposed piece-wise linear signal to noise ratio. ANN based indices were compared to normalized spectral moments (NSM) and mean frequency (MF). GMI performed as well as MI and outperformed NSM and MF demonstrating that subject and contraction-specific baseline data is not needed in order to train a mapping function which can effectively assess fatigue.     

Gender dependent EMGs of runners resolved by time/frequency and principal pattern analysis.

A promising approach for the analysis of surface electromyograms is to use wavelets to determine the spectral distribution of the signal intensity at any time. The authors have recently proposed using non-linearly scaled wavelets to obtain intensity patterns, which reflect the spectral distribution at any given time point. Further analysis of intensity-patterns is greatly facilitated by representing them as linear combinations of a base set of principal-patterns. The weight with which each principal-pattern contributes to the intensity-pattern can be represented on a set of orthogonal axes that span a previously introduced pattern space. The purpose of the present study was to show how to use pattern space to discriminate and classify male and female runners based on the electromyograms of five muscles of the limb. The results showed that there were significant gender specific differences, which allowed more than a 95% correct classification of the subjects as males or females. Classification was possible irrespective of the shod condition while running. Gender specific differences occurred at well-defined time periods during the movement. Common to both genders was that spectral changes did not parallel the changes in total signal intensity.     

Bispectrum-based features classification for myoelectric control

Surface electromyographic signals provide useful information about motion intentionality. Therefore, they are a suitable reference signal for control purposes. A continuous classification scheme of five upper limb movements applied to a myoelectric control of a robotic arm is presented. This classification is based on features extracted from the bispectrum of four EMG signal channels. Among several bispectrum estimators, this paper is focused on arithmetic mean, median, and trimmed mean estimators, and their ensemble average versions. All bispectrum estimators have been evaluated in terms of accuracy, robustness against outliers, and computational time. The median bispectrum estimator shows low variance and high robustness properties. Two feature reduction methods for the complex bispectrum matrix are proposed. The first one estimates the three classic means (arithmetic, harmonic, and geometric means) from the module of the bispectrum matrix, and the second one estimates the same three means from the square of the real part of the bispectrum matrix. A two-layer feedforward network for movement's classification and a dedicated system to achieve the myoelectric control of a robotic arm were used. It was found that the classification performance in real-time is similar to those obtained off-line by other authors, and that all volunteers in the practical application successfully completed the control task.     

The effects of four time-varying factors on the mean frequency of a myoelectric signal.

Daily activities involve dynamic muscle contractions that yield nonstationary myoelectric signals (MESs). The purpose of this work was to determine the individual effects of four time-varying factors (the number and firing rate of active motor units, muscle force and joint angle) on the mean frequency of a MES. Previous theoretical and experimental work revealed that although changes in the number and firing rate of active motor units contribute to the nonstationarities of the signal, they do not significantly affect the mean frequency. In the experimental work, 12 subjects performed 25 static contractions, one for each force (20, 30, 40, 50 and 60% of maximum voluntary contraction) and elbow joint angle (50, 70, 90, 110 and 130 degrees extension) combination. A MES was recorded from the surface of the biceps brachii during each contraction. The results indicated that muscle force only weakly affects the mean frequency. Also shown was that alteration in muscle geometry resulting from changes in elbow joint angle does significantly affect the mean frequency. Knowing this is important for the assessment of muscle fatigue during dynamic contractions.     

High density electromyography data of normally limbed and transradial amputee subjects for multifunction prosthetic control

Pattern recognition based control of powered upper limb myoelectric prostheses offers a means of extracting more information from the available muscles than conventional methods. By identifying repeatable patterns of muscle activity across multiple muscle sites rather than relying on independent EMG signals it is possible to provide more natural, reliable control of myoelectric prostheses. The purposes of this study were to (1) determine if participants can perform distinctive muscle activation patterns associated with multiple wrist and hand movements reliably and (2) to show that high density EMG can be applied individually to determine the electrode location of a clinically acceptable number of electrodes (maximally eight) to classify multiple wrist and hand movements reliably in transradial amputees. Eight normally limbed subjects (five female, three male) and four transradial amputee subjects (two traumatic and congenital) subjects participated in this study, which examined the classification accuracies of a pattern recognition control system. It was found that tasks could be classified with high accuracy (85-98%) with normally limbed subjects (10-13 tasks) and with amputees (4-6) tasks. In healthy subjects, reducing the number of electrodes to eight did not affect accuracy significantly when those electrodes were optimally placed, but did reduce accuracy significantly when those electrodes were distributed evenly. In the amputee subjects, reducing the number of electrodes up to 4 did not affect classification accuracy or the number of tasks with high accuracy, independent of whether those remaining electrodes were evenly distributed or optimally placed. The findings in healthy subjects suggest that high density EMG testing is a useful tool to identify optimal electrode sites for pattern recognition control, but its use in amputees still has to be proven. Instead of just identifying the electrode sites where EMG activity is strong, clinicians will be able to choose the electrode sites that provide the most important information for classification.     

The specificity of Ca2+ signalling

A calcium signal is a sudden increase in concentration of calcium ions (Ca2+) in the cytosol. Such signals are crucial for the control of many important functions of the body. In the brain, for example, Ca2+ signals are responsible for memory, in muscle cells they switch on contraction, whereas in gland cells they are responsible for regulation of secretion. In many cases Ca2+ signals can control several different processes in the same cell. As an example, we shall deal with one particular cell type, namely the pancreatic acinar cell, which is responsible for the secretion of the enzymes essential for the digestion of food. In this cell, Ca2+ signals do not only control the normal enzyme secretion, but also regulate growth (cell division) and programmed cell death (apoptosis). Until recently, it was a mystery how the same type of signal could regulate such diverse functions in one and the same cell. Recent technical advances have shown that different patterns of Ca2+ signals can be created, in space and time, which allow specific cellular responses to be elicited.     

Myoelectric signal measurement during prolonged computer terminal work.

Myoelectric signal (MES) behaviour was studied during prolonged, sustained, low level contractions using a portable system with limited data storage capacity. A pre-processing technique is described which overcomes memory and data storage limitations in a portable multichannel MES data logger. This technique for data reduction was used to study MES behaviour in four muscle groups during prolonged computer terminal work. Myoelectric signal parameters were recorded from eighteen individuals while they performed computer work both without breaks, and with "microbreaks" (short rest breaks of 30 seconds duration) at twenty minute intervals. Myoelectric signal (MES) data were collected from the cervical paraspinal extensors, the lumbar erector spinae, the upper trapezius, and the forearm extensors while participants performed their usual computer work activities. No significant slope for either amplitude or mean frequency was determined in either the break or no break trials over an eighty minute recording period. Instead, most data sets revealed a cyclic trend in terms of frequency and amplitude parameters of the MES. Characteristic values were compared between trials when subjects did and did not take microbreaks. It was found that the overall median value of mean frequency was higher for the "break" than the "no break" protocol only in the cervical extensors, although the clinical significance of this finding is not well understood. By far, the most interesting finding of this work was the discovery of a cyclic trend in the mean frequency of the myoelectric signals studied. This trend was present even when participants did not take breaks. The trend is a potential indicator of the cyclic recruitment of motor units during sustained postural contractions, and is the primary area to be investigated in future studies by the authors.     

Identifying COVID-19 by using spectral analysis of cough recordings: a distinctive classification study

Sound signals from the respiratory system are largely taken as tokens of human health. Early diagnosis of respiratory tract diseases is of great importance because, if delayed, it exerts irreversible effects on human health. The Coronavirus pandemic, which is deeply shaking the world, has revealed the importance of this diagnosis even more. During the pandemic, it has become the focus of researchers to differentiate symptoms from similar diseases such as influenza. Among these symptoms, the difference in cough sound played a distinctive role in research. Clinical data collected under the supervision of doctors in a reliable environment were used as the dataset consisting of 16 subjects suspected of COVID-19 with a specific patient demographic. Using the polymerase chain reaction test, the suspected subjects were divided into two groups as negative and positive. The negative and positive labels represent the patients with non-COVID and with a COVID-19 cough, respectively. Using the 3D plot or waterfall representation of the signal frequency spectrum, the salient features of the cough data are revealed. In this way, COVID-19 can be differentiated from other coughs by applying effective feature extraction and classification techniques. Power spectral density based on short-time Fourier transform and mel-frequency cepstral coefficients (MFCC) were chosen as the efficient feature extraction method. From among the classification techniques, the support vector machine (SVM) algorithm was applied to the processed signals in order to identify and classify COVID-19 cough. In terms of results evaluation, the cough of subjects with COVID-19 was detected with 95.86% classification accuracy thanks to the radial basis function (RBF) kernel function of SVM and the MFCC method. The diagnosis of COVID-19 coughs was performed with 98.6% and 91.7% sensitivity and specificity, respectively.     

Structure and polymorphism of human telomere-associated DNA

We have analyzed the DNA sequences associated with four different human telomeres. Two are members of distinct repeated sequence families which are located mainly but not exclusively at telomeres. Two are unique in the genome, one deriving from the long arm telomere of chromosome 7 and the other from the pseudoautosomal telomere. One telomere-associated repeated sequence has a polymorphic distribution among the chromosome ends, being present at a different combination of ends in different individuals. These data thus identify a new source of human genetic variation and indicate that the canonical features of the organization of telomere-associated DNA are widely conserved in evolution.     

Automatic selection of uncontaminated electromyogram as applied to respiratory muscle fatigue

An automatic procedure for detecting artifacts in the electromyogram (EMG) has been developed and applied to a study of respiratory muscle fatigue. Signal segments are characterized by a set of features, the normal variations of which have been estimated in a training session. From the features are calculated a classification variable, which expresses the degree of deviation from normal conditions. A deviation larger than a certain threshold value designates a segment as disturbed. The study deals with the choice of features, the selection of a suitable segment length, and the determination of an optimal classification threshold. The four features chosen include measures of amplitude symmetry, extreme excursions in the signal tracing, the signal-to-noise ratio, and the shape of the EMG power spectrum. Recordings from three subjects were used for the evaluation of the method. The results indicate that a segment length of 250 ms is appropriate. Accepting a 10% rate of false detections, the average rate of missed detections was 2.2%.     

Time-dependent power spectral density estimation of surface electromyography during isometric muscle contraction: Methods and comparisons

This paper studies the time-dependent power spectral density (PSD) estimation of nonstationary surface electromyography (SEMG) signals and its application to fatigue analysis during isometric muscle contraction. The conventional time-dependent PSD estimation methods exhibit large variabilities in estimating the instantaneous SEMG parameters so that they often fail to identify the changing patterns of short-period SEMG signals and gauge the extent of fatigue in specific muscle groups. To address this problem, a time-varying autoregressive (TVAR) model is proposed in this paper to describe the SEMG signal, and then the recursive least-squares (RLS) and basis function expansion (BFE) methods are used to estimate the model coefficients and the time-dependent PSD. The instantaneous parameters extracted from the PSD estimation are evaluated and compared in terms of reliability, accuracy, and complexity. Experimental results on synthesized and real SEMG data show that the proposed TVAR-model-based PSD estimators can achieve more stable and precise instantaneous parameter estimation than conventional methods.     

Study of the scapular muscle latency and deactivation time in people with and without shoulder impingement

Changes in muscle activities are commonly associated with shoulder impingement and theoretically caused by changes in motor program strategies. The purpose of this study was to assess for differences in latencies and deactivation times of scapular muscles between subjects with and without shoulder impingement. Twenty-five healthy subjects and 24 subjects with impingement symptoms were recruited. Glenohumeral kinematic data and myoelectric activities using surface electrodes from upper trapezius (UT), lower trapezius (LT), serratus anterior (SA) and anterior fibers of deltoid were collected as subjects raised and lowered their arm in response to a visual cue. Data were collected during unloaded, loaded and after repetitive arm raising motion conditions. The variables were analyzed using 2 or 3 way mixed model ANOVAs. Subjects with impingement demonstrated significantly earlier contraction of UT while raising in the unloaded condition and an earlier deactivation of SA across all conditions during lowering of the arm. All subjects exhibited an earlier activation and delayed deactivation of LT and SA in conditions with a weight held in hand. The subjects with impingement showed some significant differences to indicate possible differences in motor control strategies. Rehabilitation measures should consider appropriate training measures to improve movement patterns and muscle control.     

Differential affinity and cooperativity functions of the amino-terminal 70 residues of lambda integrase

The site-specific recombinase (Int) of bacteriophage lambda is a heterobivalent DNA-binding protein that binds two different classes of DNA-binding sites within its recombination target sites. The several functions of Int are apportioned between a large carboxy-terminal domain that cleaves and ligates DNA at each of its four "core-type" DNA-binding sites and a small amino-terminal domain, whose primary function is binding to each of its five "arm-type" DNA sites, which are distant from the core region. Int bridges between the two classes of binding sites are facilitated by accessory DNA-bending proteins that along with Int comprise higher-order recombinogenic complexes. We show here that although the 64 amino-terminal residues of Int bind efficiently to a single arm site, this protein cannot form doubly bound complexes on adjacent arm sites. However, 1-70 Int does show the same cooperative binding to adjacent arm sites as the full length protein. We also found that 1-70 Int specifies cooperative interactions with the accessory protein Xis when the two are bound to their adjacent cognate sites P2 and X1, respectively. To complement the finding that these two amino-terminal domain functions (along with arm DNA binding) are all specified by residues 1-70, we determined that Thr75 is the first residue of the minimal carboxy-terminal domain, thereby identifying a specific interdomain linker region. We have measured the affinity constants for Int binding to each of the five arm sites and the cooperativity factors for Int binding to the two pairs of adjacent arm sites, and we have identified several DNA structural features that contribute to the observed patterns of Int binding to arm sites. Taken together, the results highlight several interesting features of arm DNA binding that invite speculation about additional levels of complexity in the regulation of lambda site-specific recombination.     

Identification of highly specific electroencephalographic patterns in patients with neuroepithelial tumors of the III ventricle by algorithmic methods

An algorithm was suggested for identifying highly specific electroencephalographic (EEG) patterns in neurooncologic patients. The algorithm provides selection of patients with their further classification into main and control groups based on the already existing database of EEG indicators; requests to it; generation of mono-indicator candidates for EEG-patterns on the basis of a 4-dipole table for selecting and verifying sensitive and specific EEG patterns and outlining the best ones. Our material included 368 patients with basal-diencephalic tumors. Algorithmic methods revealed new EEG patterns in patients with different anatomical and topographical variants of neuroepithelial tumors in the III ventricle. We think it reasonable to use the revealed syndromes to improve diagnosis and identify pathophysiological basis of clinical syndromes.     

Classification and lateralization of temporal lobe epilepsies with and without hippocampal atrophy based on whole-brain automatic MRI segmentation

Brain images contain information suitable for automatically sorting subjects into categories such as healthy controls and patients. We sought to identify morphometric criteria for distinguishing controls (n = 28) from patients with unilateral temporal lobe epilepsy (TLE), 60 with and 20 without hippocampal atrophy (TLE-HA and TLE-N, respectively), and for determining the presumed side of seizure onset. The framework employs multi-atlas segmentation to estimate the volumes of 83 brain structures. A kernel-based separability criterion was then used to identify structures whose volumes discriminate between the groups. Next, we applied support vector machines (SVM) to the selected set for classification on the basis of volumes. We also computed pairwise similarities between all subjects and used spectral analysis to convert these into per-subject features. SVM was again applied to these feature data. After training on a subgroup, all TLE-HA patients were correctly distinguished from controls, achieving an accuracy of 96 ± 2% in both classification schemes. For TLE-N patients, the accuracy was 86 ± 2% based on structural volumes and 91 ± 3% using spectral analysis. Structures discriminating between patients and controls were mainly localized ipsilaterally to the presumed seizure focus. For the TLE-HA group, they were mainly in the temporal lobe; for the TLE-N group they included orbitofrontal regions, as well as the ipsilateral substantia nigra. Correct lateralization of the presumed seizure onset zone was achieved using hippocampi and parahippocampal gyri in all TLE-HA patients using either classification scheme; in the TLE-N patients, lateralization was accurate based on structural volumes in 86 ± 4%, and in 94 ± 4% with the spectral analysis approach. Unilateral TLE has imaging features that can be identified automatically, even when they are invisible to human experts. Such morphometric image features may serve as classification and lateralization criteria. The technique also detects unsuspected distinguishing features like the substantia nigra, warranting further study.     

On the effect of thermal agents in the response of the brachial biceps at different contraction levels

The objective of this study was to assess electromyographic features of the brachial biceps muscle after the application of cryotherapy and short-wave diathermy. Sixty healthy volunteers participated in the study and were equally divided into three groups: cryotherapy – application of ice packs for 30 min; short-wave diathermy for 20 min; and control. The thermal agents were applied to the anterior and posterior regions of the non-dominant arm. The electromyographic (EMG) signal from the brachial biceps was recorded before and after the application of thermal agents during flexion of the elbow joint at 25%, 50%, 75% of a maximum voluntary isometric contraction defined at least two days before the actual experiments (MVIC_bl). The volunteers also were asked to execute a free MVIC before and after the application of the thermal agents (MVIC_free). A linear regression model with mixed effects (random and fixed) was used. Intra-group analysis showed a reduction in root mean square (RMS) at MVIC_free, with no change in the median frequency of the EMG signal at any contraction level for the short-wave diathermy group. An increase on RMS values and a decrease on median frequencies were found after the application of cryotherapy for all contraction levels. The results imply that cryotherapy plays an important role on changing neuromuscular responses at various levels of muscle contraction. Therapists should be aware of that and carefully consider its use prior to activities in which neuromuscular precision is required.     

Altered patterns of muscle activation during performance of four functional tasks in patients with shoulder disorders: interpretation from voluntary response index.

Altered motor control of the shoulder muscles during performance of a specific motor task in patients with shoulder disorders (SDs) has been an interesting subject to researchers. This study compared shoulder muscle activation patterns by surface electromyography (sEMG), including the upper trapezius (UT), lower trapezius (LT), and serratus anterior (SA) muscles, during four functional tasks in 25 patients with SDs and controls. A voluntary response index (VRI) was calculated, including magnitude and similarity index (SI), to quantify sEMG patterns during four functional tasks. Responsiveness and clinically meaningful levels of discrimination between patients and control for EMG magnitude and SI were determined. An altered pattern of motor control during four functional tasks was evident in the patients, in which greater EMG amplitude and abnormal EMG patterns were found. For SI among four functional tasks, normal subjects ranged from 0.80 to 1.00 while patients ranged from 0.70 to 0.99. High probabilities (97%) of discrimination between patients and normal subjects were found by SI method during an overhead height task (patients: 0.85-0.96, normal subjects: 0.95-1.00). Our results also suggest that an individual can be estimated to be abnormal when lower SI values are observed during the four functional tasks.     

Evidence for composite cost functions in arm movement planning: an inverse optimal control approach

An important issue in motor control is understanding the basic principles underlying the accomplishment of natural movements. According to optimal control theory, the problem can be stated in these terms: what cost function do we optimize to coordinate the many more degrees of freedom than necessary to fulfill a specific motor goal? This question has not received a final answer yet, since what is optimized partly depends on the requirements of the task. Many cost functions were proposed in the past, and most of them were found to be in agreement with experimental data. Therefore, the actual principles on which the brain relies to achieve a certain motor behavior are still unclear. Existing results might suggest that movements are not the results of the minimization of single but rather of composite cost functions. In order to better clarify this last point, we consider an innovative experimental paradigm characterized by arm reaching with target redundancy. Within this framework, we make use of an inverse optimal control technique to automatically infer the (combination of) optimality criteria that best fit the experimental data. Results show that the subjects exhibited a consistent behavior during each experimental condition, even though the target point was not prescribed in advance. Inverse and direct optimal control together reveal that the average arm trajectories were best replicated when optimizing the combination of two cost functions, nominally a mix between the absolute work of torques and the integrated squared joint acceleration. Our results thus support the cost combination hypothesis and demonstrate that the recorded movements were closely linked to the combination of two complementary functions related to mechanical energy expenditure and joint-level smoothness.     

Identification of isometric contractions based on High Density EMG maps

Identification of motion intention and muscle activation strategy is necessary to control human–machine interfaces like prostheses or orthoses, as well as other rehabilitation devices, games and computer-based training programs. Pattern recognition from sEMG signals has been extensively investigated in the last decades, however, most of the studies did not take into account different strengths and EMG distributions associated to the intended task. The identification of such quantities could be beneficial for the training of the subject or the control of assistive devices. Recent studies have shown the need to improve pattern-recognition classification by reducing sensitivity to changes in the exerted strength, muscle-electrode shifts and bad contacts. Surface High Density EMG (HD-EMG) obtained from 2-dimensional arrays can provide much more information than electrode pairs for inferring not only motion intention but also the strategy adopted to distribute the load between muscles as well as changes in the spatial distribution of motor unit action potentials within a single muscle because of it.The objectives of this study were: (a) the automatic identification of four isometric motor tasks associated with the degrees of freedom of the forearm: flexion–extension and supination–pronation and (b) the differentiation among levels of voluntary contraction at low-medium efforts. For this purpose, monopolar HD-EMG maps were obtained from five muscles of the upper-limb in healthy subjects. An original classifier is proposed, based on: (1) Two steps linear discriminant analysis of the EMG information for each type of contraction, and (2) features extracted from HD-EMG maps and related to its intensity and distribution in the 2D space. The classifier was trained and tested with different effort levels. Spatial distribution-based features by themselves are not sufficient to classify the type of task or the effort level with an acceptable accuracy; however, when calculated with the “isolated masses” method proposed in this study and combined with intensity-base features, the performance of the classifier is improved. The classifier is capable of identifying the tasks even at 10% of Maximum Voluntary Contraction, in the range of effort level developed by patients with neuromuscular disorders, showing that intention end effort of motion can be estimated from HD-EMG maps and applied in rehabilitation.