Single-channel Surface Research Articles

A 3D surface coil with deep learning‐based noise reduction for parotid gland imaging at 7T

AbstractBackgroundBackground: Parotid gland neoplasms occur near the facial nerve. Hence, it is crucial to determine whether the malignant neoplasms involve the facial nerve and whether sacrifice of the nerve in surgery is necessary. Furthermore, while 20% of all neoplasms are malignant, the most common benign neoplasm, pleomorphic adenoma, has a risk for malignant transformation, making early detection and treatment essential. 7T magnetic resonance imaging offers increased signal‐to‐noise ratio (SNR) and sensitivity.AimIn this work, we address imaging the parotid gland since it remains challenging at 7T because of its spatial location.Materials and MethodsHere, we present a novel three‐dimensional surface coil (3D Coil) architecture that offers increased depth penetration and SNR compared to the single channel surface coil. We further developed a deep learning (DL)‐based noise reduction method that receives inputs from three elements of the 3D Coil.ResultsThe 3D coil with DL‐based denoising method offers twice the SNR compared to the single channel surface coil for parotid gland imaging at 7T.Discussion and ConclusionThe proposed 3D Coil and DL‐based noise reduction method offers a promising way of achieving higher SNR for parotid salivary gland imaging at 7T, paving the road for clinical applications.

Improved deep residual shrinkage network for a multi-cylinder heavy-duty engine fault detection with single channel surface vibration

The health monitoring and fault diagnosis of heavy-duty engines are increasingly important for energy storage ecosystem. During operation, vibration characters corresponding to the specific fault need to be extracted from the overall system vibration. Faulty characteristics emanating from one single cylinder are also mixed with those from other cylinders. Besides, the change of working condition brings strong nonlinearities in surface vibration. To solve these problems, an improved deep residual shrinkage network (IDRSN) is developed for detecting diverse engine faults at various degrees using single channel surface vibration signal. Within IDRSN, a wide convolution kernel is utilized in first convolution layer to capture the long-term fault-related impacts and eliminate the short-time random impact. The residual network module is adopted to enhance the focus the relevant components of vibration signals. Mini-batch training strategy is used to improve the model stability. Meanwhile, Gradient-weighted class activation map is adopted to assess the consistency between the learned knowledge and the fault-related information. The IDRSN is implemented to diagnosing a diesel engine under various faults, faulty degrees and operating speeds. Comparisons with existing models are analyzed in terms of hyper-parameters, training samples, noise resistance, and visualization. Results demonstrate the proposed IDRSN's superior performance on fault diagnosis accuracy, stability, anti-noise performance, and anti-interference performance. An average accuracy rate of 98.38 % was achieved by the proposed IDRSN, in comparison to 96.64 % and 93.56 % achieved by the DRSN and the wide-kernel deep convolutional neural network respectively. These results highlight the proposed IDRSN's superiority in diagnosing multiple faults under various working conditions, offering a low-cost, highly effective, and applicable approach for complex fault diagnosis tasks.

Whole-body fasciculation detection in amyotrophic lateral sclerosis using motor unit MRI

ObjectiveCompare fasciculation rates between amyotrophic lateral sclerosis (ALS) patients and healthy controls in body regions relevant for diagnosing ALS using motor unit MRI (MUMRI) at baseline and 6 months follow-up, and relate this to single-channel surface EMG (SEMG). MethodsTongue, biceps brachii, paraspinals and lower legs were assessed with MUMRI and biceps brachii and soleus with SEMG in 10 healthy controls and 10 patients (9 typical ALS, 1 primary lateral sclerosis [PLS]). ResultsMUMRI-detected fasciculation rates in typical ALS patients were higher compared to healthy controls for biceps brachii (2.40 ± 1.90 cm-3min−1vs. 0.04 ± 0.10 cm-3min−1, p = 0.004), paraspinals (1.14 ± 1.61 cm-3min−1vs. 0.02 ± 0.02 cm-3min−1, p = 0.016) and lower legs (1.42 ± 1.27 cm-3min−1vs. 0.13 ± 0.10 cm-3min−1, p = 0.004), but not tongue (1.41 ± 1.94 cm-3min−1vs. 0.18 ± 0.18 cm-3min−1, p = 0.556). The PLS patient showed no fasciculation. At baseline, 6/9 ALS patients had increased fasciculation rates compared to healthy controls in at least 2 body regions. At follow-up every patient had increased fasciculation rates in at least 2 body regions. The MUMRI-detected fasciculation rate correlated with SEMG-detected fasciculation rates (τ = 0.475, p = 0.006). ConclusionMUMRI can non-invasively image fasciculation in multiple body regions and appears sensitive to disease progression in individual patients. SignificanceMUMRI has potential as diagnostic tool for ALS.

Investigation of Motor Units Activity: Comparison of Single Channel Surface EMG Deconvolution and Blind Source Separation of Multichannel Data

Investigation of Motor Units Activity: Comparison of Single Channel Surface EMG Deconvolution and Blind Source Separation of Multichannel Data

Relative contribution of neuromuscular activation, muscle size, and muscle quality to maximum strength output of the thigh muscles in young individuals.

This study aimed to investigate the relationship between maximal muscle strength and neuromuscular activation, muscle size, and quality of quadriceps (QF) and hamstring muscles (HM). The study included 24 young men and women. The neuromuscular activation parameter was recorded using a single-channel surface electromyography (EMG) with the root mean square (RMS) during maximal isometric knee extension and flexion from four muscles: rectus femoris and vastus lateralis for QF; biceps femoris and semitendinosus for HM. In addition, the peak torque was measured during the same session. B-mode ultrasonographic transverse images were obtained from the anterior, lateral, and posterior thighs. Furthermore, we calculated the muscle thickness (MT) and echo intensity (EI) of the four muscles as indicators of muscle size and quality. The averaged MT, EI, and absolute RMS of QF were calculated by averaging the values of the rectus femoris and vastus lateralis, and that of HM was calculated by averaging the values of the biceps femoris and semitendinosus. The knee extension peak torque was correlated with EI (r=-0.61, P < 0.01) and RMS (r=0.53, P < 0.01) in the QF. In contrast, the knee flexion peak torque was correlated with RMS (r=0.53, P < 0.05) but not with MT and EI in HM. In addition, EI and RMS in QF, and RMS in HM were selected as the major determinants of muscle strength in the stepwise regression analysis. These results suggest that muscle strength is moderately associated with different factors related to the thigh muscles in young individuals.

Prediction Method for Large Grasping Force using Single-Channel Surface Electromyography: A Pilot Study

Frequent grasping with a large force in the workplace leads to work-related musculoskeletal disorders (WMSDs). Thus, the frequency and force of grasping in each worker should be monitored for preventing these WMSDs. The frequency of grasping can be counted by observation and existing methods using vision-based devices or wearable sensors. Grasping force measurement generally requires force sensors between the hand and objects. Almost force sensors are limited to the surface and shape of the object in measurement; thus, measurement method without force sensors is necessary to monitor grasping force in various workplaces. Previous studies developed measurement methods for grasping force using electromyography (EMG). However, almost existing methods require multiple-channels EMG for grasping force measurement. In addition, many methods were not applied to large grasping forces. Therefore, the objective of this study is to develop the prediction method for large grasping force using single-channel EMG. The proposed method predicts grasping force by machine learning and features calculated from single channel surface EMG (sEMG) signal on flexor digitorum superficialis (FDS) muscle. As a pilot study, the proposed method was tested by sEMG data of different large grasping forces (5 to 60 kgf). The results showed that the correlation between the actual and predicted grasping force was greater than 0.8. These results indicate that the proposed method could measure large grasping force by only single-channel sEMG.

Investigation of the causes of land surface temperature changes in Khorramabad

development of remote sensing applications has led to their use in a wide range of environmental studies. One of these aspects is urban studies and especially land surface temperature. In this study, the changes in land surface temperature in the Khorramabad city in Iran were investigated and the causes of land temperature changes were investigated. For this purpose, Landsat satellite images were processed in four periods of 2000, 2007, 2014 and 2021 and were recovered using a single-channel surface temperature algorithm. Temperatures were high in 2000 due to the type of roofs of buildings and the dirt around the city. Decreased in 2007 and 2014 due to the use of roofs that reflect light. In 2021, due to severe manipulations around the city and the destruction of vegetation and change it into built-up bare soil cover caused the temperature to rise again in the suburbs.

Simulation Study of Radio Frequency Safety and the Optimal Size of a Single-Channel Surface Radio Frequency Coil for Mice at 9.4 T Magnetic Resonance Imaging.

The optimized size of a single-channel surface radio frequency (RF) coil for mouse body images in a 9.4 T magnetic resonance imaging (MRI) system was determined via electromagnetic-field analysis of the signal depth according to the size of a single-channel coil. The single-channel surface RF coils used in electromagnetic field simulations were configured to operate in transmission/reception mode at a frequency of 9.4 T–400 MHz. Computational analysis using the finite-difference time-domain method was used to assess the single-channel surface RF coil by comparing single-channel surface RF coils of varying sizes in terms of |B1|-, |B1+|-, |B1−|- and |E|-field distribution. RF safety for the prevention of burn injuries to small animals was assessed using an analysis of the specific absorption rate. A single-channel surface RF coil with a 20 mm diameter provided optimal B1-field distribution and RF safety, thus confirming that single-channel surface RF coils with ≥25 mm diameter could not provide typical B1-field distribution. A single-channel surface RF coil with a 20 mm diameter for mouse body imaging at 9.4 T MRI was recommended to preserve the characteristics of single-channel surface RF coils, and ensured that RF signals were applied correctly to the target point within RF safety guidelines.

Single Channel Surface Electromyogram Deconvolution is a Useful Pre-Processing for Myoelectric Control.

Myoelectric control requires fast and stable identification of a movement from data recorded from a comfortable and straightforward system. We consider a new real-time pre-processing method applied to a single differential surface electromyogram (EMG): deconvolution, providing an estimation of the cumulative firings of motor units. A 2 channel-10 class finger movement problem has been investigated on 10 healthy subjects. We have compared raw EMG and deconvolution signals, as sources of information for two specific classifiers (based on either Support Vector Machines or k-Nearest Neighbours), with classical time-domain input features selected using Mutual Component Analysis. Using the proposed pre-processing technique, classification performances statistically improve. For example, the true positive rates of the best-tested configurations were 80.9% and 86.3% when using the EMG and its deconvoluted signal, respectively. Even considering the limited dataset and range of classification approaches investigated, our preliminary results indicate the potential usefulness of the deconvolution pre-processing. Deconvolution of EMG is a fast pre-processing that could be easily embedded in different myoelectric control applications.

Single-channel surface electromyography signal classification with variational mode decomposition and entropy feature for lower limb movements recognition

Currently, many researchers tend to use multi-channel surface electromyography (sEMG) signals to improve the accuracy of lower limb movement recognition. However, the collection of multi-channel sEMG signals will reduce the usability of wearable devices for lower limbs based on sEMG signals in amputees, patients with impaired muscle function, and the disabled. How to effectively use single-channel sEMG signals to achieve better recognition performance is a difficult problem to improve the usability of wearable devices based on sEMG signals. In this research, we proposed a precise feature extraction method for single-channel sEMG signals to achieve accurate recognition of lower limb movements. The single-channel sEMG signal was decomposed into multiple variational modal functions (VMF) through variational mode decomposition (VMD), and entropy features were extracted from VMFs to highlight the prominent information of the sEMG signal. Entropy features with statistical differences were selected by the Kruskal-Wallis test. Four lower limb movements were recognized through machine learning. Moreover, the recognition performance exhibited by the proposed method on the sEMG signal of two different muscles was evaluated. The sEMG signals of four lower limb movements from twenty subjects recorded by the wearable sEMG signal sensor were employed to test the proposed method. The experimental results showed that the accuracy of the proposed method for the sEMG signals of two different muscles reached 95.82% and 97.44%. This research concluded that the proposed method is promising to improve the usability of wearable devices based on sEMG signals in amputees, patients with impaired muscle function, and the disabled.

Electromyography Signal Analysis and Classification using Time-Frequency Representations and Deep Learning.

Analysis and classification of electromyography (EMG) signals are crucial for rehabilitation and motor control. This study investigates electromyogram (EMG) time-frequency representations and then creates conventional and deep learning models for EMG signal classification. Firstly, a dataset of single-channel surface EMG signals has been recorded for four subjects to differentiate between forearm flexion and extension. Then, different time-frequency EMG representations have been used to build conventional and deep learning models for EMG classification. We compared the performance of pre-trained convolutional neural network models, namely GoogLeNet, SqueezeNet and AlexNet, and achieved accuracies of 92.71%, 90.63% and 87.5%, respectively. Also, data augmentation techniques on the levels of raw EMG signals and their time- frequency representations helped improve the accuracy of GoogLeNet to 96.88%. Furthermore, our approach demonstrated superior performance on another publicly available 10-class EMG dataset, and also using traditional classifiers trained on hand-crafted features.

Single-channel surface electromyography (sEMG) based control of a multi-functional prosthetic hand

Patients suffering from upper limb amputation experience a significant reduction in their ability to perform various tasks. The currently available myoelectric prostheses seek to reinstate the lost abilities of the amputees. However, most users abandon their devices due to high cost, large size, heavy weight, complexity, and limited functional control. This paper presents a multifunctional prosthetic hand that can perform six hand activities deploying a single-channel surface electromyography (sEMG) sensor. EMG signals for fifteen subjects (five amputees and ten intact) were acquired for six contraction levels of forearm muscles using the designed sensor. These levels were further classified to recognize six predefined hand gestures by a fuzzy logic classifier. The proposed system showed an excellent success (> 95%) and other performance parameters above 96%. The intended classification-based control scheme was further realized in real-time to achieve six grip patterns for the developed prosthetic prototype. The hand was tested on five subjects (two amputees and three intact) showed a percentage of success above 91% for accomplishing the dexterous grasping operations. The proposed approach is modest, efficient, and provides a low-cost solution to amputees with multiple degrees of freedom with a single channel EMG system.

An applied method for railway wheel profile measurements due to wear using image processing techniques

Accurate wear prediction on railway wheels and evolution of railway wheels profile can affect the maintenance planning. The objective of this paper is to provide a new applied method for measuring the railway wheel profile with photographing from the railway wheel to measure by image processing techniques. The aim of this new applied method is to measure the wheel profile using images taken from railway wheels and compare it with the original plan. For this purpose, images were taken by using a camera. In this study, all automatic correction options were turned off and brightness and contrast were in normal conditions. The pixel data is converted to double-type data and are placed in the range of zero and one. Then, the input image which is usually as a three-channel or RGB image is converted to a single channel or gray surface image. Images taken from wheel profiles are processed using image processing techniques. Then the lines, curves and shapes in the image are extracted as cross-sectional and continuous curves. The new applied method results by image processing method obtained show good agreement with those achieved in field measurements.

A Minimal Setup for Spontaneous Smile Quantification Applicable for Valence Detection.

Tracking emotional responses as they unfold has been one of the hallmarks of applied neuroscience and related disciplines, but recent studies suggest that automatic tracking of facial expressions have low validation. In this study, we focused on the direct measurement of facial muscles involved in expressions such as smiling. We used single-channel surface electromyography (sEMG) to evaluate the muscular activity from the Zygomaticus Major face muscle while participants watched music videos. Participants were then tasked with rating each video with regard to their thoughts and responses to each of them, including their judgment of emotional tone (“Valence”), personal preference (“Liking”) and rating of whether the video displayed strength and impression (“Dominance”). Using a minimal recording setup, we employed three ways to characterize muscular activity associated with spontaneous smiles. The total time spent smiling (ZygoNum), the average duration of smiles (ZygoLen), and instances of high valence (ZygoTrace). Our results demonstrate that Valence was the emotional dimension that was most related to the Zygomaticus activity. Here, the ZygoNum had higher discriminatory power than ZygoLen for Valence quantification. An additional investigation using fractal properties of sEMG time series confirmed previous studies of the Facial Action Coding System (FACS) documenting a smoother contraction of facial muscles for enjoyment smiles. Further analysis using ZygoTrace responses over time to the video events discerned “high valence” stimuli with a 76% accuracy. Additional validation of this approach came against previous findings on valence detection using features derived from a single channel EEG setup. We discuss these results in light of both the recent replication problems of facial expression measures, and in relation to the need for methods to reliably assess emotional responses in more challenging conditions, such as Virtual Reality, in which facial expressions are often covered by the equipment used.

Appropriate window width for the “clustering index method” in the tibialis anterior muscle

We previously reported a new quantitative analysis of single-channel surface electromyography (EMG), the "clustering index method" (CI method), in the tibialis anterior muscle, which achieved sufficiently good sensitivity to detect neurogenic or myogenic abnormalities. The window width is a fundamental parameter of the CI method, and was arbitrarily set at 15 ms in that study. In this study, we searched for the most appropriate window width using expanded patient data. The data from our previous study were reanalyzed, and new patients were enrolled. Window width in the CI method was changed from 5 to 27.5 ms with a step of 2.5 ms. For each window width, Z-score values of individual subjects were calculated and the diagnostic yield was investigated. We enrolled 67 controls, 29 subjects with neurogenic disorders, and 39 with myogenic disorders. When the window width was set at 22.5 ms, the highest sensitivity was achieved both for neurogenic (97%) and myogenic (72%) disorders, with a specificity of 97%. Seven of 10 patients with inclusion body myositis were also abnormal. Reliable results were obtained by collecting 15 epochs per subject. There are two conflicting effects that appear to be best balanced at a window width of 22.5 ms: a wider width decreases the chance that a motor unit potential (MUP) is divided into two adjacent windows, and a narrower width reduces the possibility that an MUP firing at a low-frequency is counted twice by the differential sequences. CI is promising as a non-invasive method to diagnose neuromuscular disorders.

Development and optimization of a receive-only surface array with purely geometrical decoupling for rat brain MRI at 2 T

Magnetic resonance imaging (MRI) experiments on small animals, as well as in humans, require specific radiofrequency (RF) coil designs in order to maximize field homogeneity during excitation and signal-to-noise ratio (SNR) during reception. Therefore, the use of dedicated transmit-only and receive-only coil geometries has been preferred as standard technology configuration. Among the receive coil geometries, the surface coil has been widely used to increase the SNR in localized regions of interest. With the advent of MRI-phased arrays, the high sensitivity of surface coils can be achieved in extended regions, further allowing parallel imaging acquisition methodologies to accelerate acquisition time. This work describes the development and characterization of a two-channel receive-only RF coil set specifically designed for MRI acquisition of rat brain in a 2 T system for neuroscience studies. We have developed and compared the performance of one common surface coil with a two-channel phased array with purely geometric decoupling between coil elements (i.e., without using low input impedance preamplifiers), both actively detuned during reception and having exactly the same dimensions. The results have shown that the single-channel surface coil achieved up to 5 times higher SNR at 5 mm depth compared to a volume coil commonly used for rat brain MRI in our laboratory, while the two-channel array achieved 4 times higher SNR at the same depth, with the additional advantage of producing an extended area with higher sensitivity. In vivo brain images of anesthetized rats confirmed the SNR gain for both receive-only single-channel surface coil and two-channel array in comparison with a volume coil. We have verified that only using the partial overlapping technique does not provide sufficient isolation to eliminate mutual coupling interaction between elements, since residual coupling can still affect the coil performance.

A machine learning approach to epileptic seizure prediction using Electroencephalogram (EEG) Signal

This study investigates the properties of the brain electrical activity from different recording regions and physiological states for seizure detection. Neurophysiologists will find the work useful in the timely and accurate detection of epileptic seizures of their patients. We explored the best way to detect meaningful patterns from an epileptic Electroencephalogram (EEG). Signals used in this work are 23.6 s segments of 100 single channel surface EEG recordings collected with the sampling rate of 173.61 Hz. The recorded signals are from five healthy volunteers with eyes closed and eyes open, and intracranial EEG recordings from five epilepsy patients during the seizure-free interval as well as epileptic seizures. Feature engineering was done using; i) feature extraction of each EEG wave in time, frequency and time-frequency domains via Butterworth filter, Fourier Transform and Wavelet Transform respectively and, ii) feature selection with T-test, and Sequential Forward Floating Selection (SFFS). SVM and KNN learning algorithms were applied to classify preprocessed EEG signal. Performance comparison was based on Accuracy, Sensitivity and Specificity. Our experiments showed that SVM has a slight edge over KNN.

Differentiation of Epileptic and Psychogenic Nonepileptic Seizures Using Single-Channel Surface Electromyography.

Epileptic seizures (ES) and psychogenic nonepileptic seizures (PNES) are difficult to differentiate when based on a patient's self-reported symptoms. This study proposes review of objective data captured by a surface electromyography (sEMG) wearable device for classification of events as ES or PNES. This may help clinicians accurately identify ES and PNES. Seventy-one subjects were prospectively enrolled across epilepsy monitoring units at VA Epilepsy Centers of Excellence. Subjects were concomitantly monitored using video EEG and a wearable sEMG epilepsy monitor, the Sensing Portable sEmg Analysis Characterization (SPEAC) System. Three epileptologists independently classified ES and PNES that contained upper extremity motor activity based on video EEG. The sEMG data from those events were automatically processed to provide a seizure score for event classification. After brief training (60 minutes), the sEMG data were reviewed by a separate group of four epileptologists to independently classify events as ES or PNES. According to video EEG review, 17 subjects experienced 34 events (15 ES and 19 PNES with upper extremity motor activity). The automated process correctly classified 87% of ES (positive predictive value = 88%, negative predictive value = 76%) and 79% of PNES, and the expert reviewers correctly classified 77% of ES (positive predictive value = 94%, negative predictive value = 84%) and 96% of PNES. The automated process and the expert reviewers correctly classified 100% of tonic-clonic seizures as ES, and 71 and 50%, respectively, of non-tonic-clonic ES. Automated and expert review, particularly in combination, of sEMG captured by a wearable seizure monitor (SPEAC System) may be able to differentiate ES (especially tonic-clonic) and PNES with upper extremity motor activity.

Locomo-Net: A Low -Complex Deep Learning Framework for sEMG-Based Hand Movement Recognition for Prosthetic Control.

Background: The enhancement in the performance of the myoelectric pattern recognition techniques based on deep learning algorithm possess computationally expensive and exhibit extensive memory behavior. Therefore, in this paper we report a deep learning framework named ‘Low-Complex Movement recognition-Net’ (LoCoMo-Net) built with convolution neural network (CNN) for recognition of wrist and finger flexion movements; grasping and functional movements; and force pattern from single channel surface electromyography (sEMG) recording. The network consists of a two-stage pipeline: 1) input data compression; 2) data-driven weight sharing. Methods: The proposed framework was validated on two different datasets- our own dataset (DS1) and publicly available NinaPro dataset (DS2) for 16 movements and 50 movements respectively. Further, we have prototyped the proposed LoCoMo-Net on Virtex-7 Xilinx field-programmable gate array (FPGA) platform and validated for 15 movements from DS1 to demonstrate its feasibility for real-time execution. Results: The effectiveness of the proposed LoCoMo-Net was verified by a comparative analysis against the benchmarked models using the same datasets wherein our proposed model outperformed Twin- Support Vector Machine (SVM) and existing CNN based model by an average classification accuracy of 8.5 % and 16.0 % respectively. In addition, hardware complexity analysis is done to reveal the advantages of the two-stage pipeline where approximately 27 %, 49 %, 50 %, 23 %, and 43 % savings achieved in lookup tables (LUT’s), registers, memory, power consumption and computational time respectively. Conclusion: The clinical significance of such sEMG based accurate and low-complex movement recognition system can be favorable for the potential improvement in quality of life of an amputated persons.

Automated Processing of Single-Channel Surface Electromyography From Generalized Tonic-Clonic Seizures to Inform Semiology.

Advances in surface electromyography (sEMG) monitoring allow for long-term data collection in a natural environment, giving objective information that may identify risk of sudden unexpected death in epilepsy and guide clinical decision-making. Generalized tonic-clonic seizure semiology, namely motor tonic and clonic phase duration, may be an important factor in determining the level of seizure control and risk of sudden unexpected death in epilepsy. This study demonstrates a quantitative analysis of sEMG collected with a dedicated wearable device. During routine monitoring, 23 generalized tonic-clonic seizures from 19 subjects were simultaneously recorded with video-EEG and sEMG. A continuous wavelet-transform was used to determine the frequency components of sEMG recorded during generalized tonic-clonic seizures. An automated process, incorporating a variant of cross-validation, was created to identify ideal frequencies and magnitude ranges for tonic and clonic phases and determine phase durations. Phase durations determined using sEMG analysis were compared with phase durations determined by independent epileptologists' review of video-EEG. Cross-validation revealed that the optimal frequency bands for tonic and clonic phases are 150 to 270 Hz and 12 to 70 Hz, respectively. The average difference in phase duration calculated using the two methods for tonic and clonic phases and total seizure duration were -0.42 ± 4.94, -5.12 ± 9.68, and -5.11 ± 11.33 seconds, respectively (results presented are TsEMG - TvEEG, μ ± σ). The automated processing of sEMG presented here accurately identified durations of tonic, clonic, and total motor durations of generalized tonic-clonic seizures similar to durations identified by epileptologists' review of video-EEG.