A deep learning image analysis method for renal perfusion estimation in pseudo-continuous arterial spin labelling MRI.

Accurate segmentation of renal tissues is an essential step for renal perfusion estimation and postoperative assessment of the allograft. Images are usually manually labeled, which is tedious and prone to human error. We present an image analysis method for the automatic estimation of renal perfusion based on perfusion magnetic resonance imaging. Specifically, non-contrasted pseudo-continuous arterial spin labeling (PCASL) images are used for kidney transplant evaluation and perfusion estimation, as a biomarker of the status of the allograft. The proposed method uses machine/deep learning tools for the segmentation and classification of renal cortical and medullary tissues and automates the estimation of perfusion values. Data from 16 transplant patients has been used for the experiments. The automatic analysis of differentiated tissues within the kidney, such as cortex and medulla, is performed by employing the time-intensity-curves of non-contrasted T1-weighted MRI series. Specifically, using the Dice similarity coefficient as a figure of merit, results above 93%, 92% and 82% are obtained for whole kidney, cortex, and medulla, respectively. Besides, estimated cortical and medullary perfusion values are considered to be within the acceptable ranges within clinical practice.

Validation of the filling factor index to study the filling process of the sEMG signal in the quadriceps.

INTRODUCTION: The EMG filling factor is an index to quantify the degree to which an EMG signal has been filled. Here, we tested the validity of such index to analyse the EMG filling process as contraction force was slowly increased. METHODS: Surface EMG signals were recorded from the quadriceps muscles of healthy subjects as force was gradually increased from 0 to 40% MVC. The sEMG filling process was analyzed by measuring the EMG filling factor (calculated from the non-central moments of the rectified sEMG). RESULTS: (1) As force was gradually increased, one or two prominent abrupt jumps in sEMG amplitude appeared between 0 and 10% of MVC force in all the vastus lateralis and medialis. (2) The jumps in amplitude were originated when a few large-amplitude MUPs, clearly standing out from previous activity, appeared in the sEMG signal. (3) Every time an abrupt jump in sEMG amplitude occurred, a new stage of sEMG filling was initiated. (4) The sEMG was almost completely filled at 2-12% MVC. (5) The filling factor decreased significantly upon the occurrence of an sEMG amplitude jump, and increased as additional MUPs were added to the sEMG signal. (6) The filling factor curve was highly repeatable across repetitions. CONCLUSIONS: It has been validated that the filling factor is a useful, reliable tool to analyse the sEMG filling process. As force was gradually increased in the vastus muscles, the sEMG filling process occurred in one or two stages due to the presence of abrupt jumps in sEMG amplitude.

M-wave changes caused by brief voluntary and stimulated isometric contractions.

INTRODUCTION: Under isometric conditions, the increase in muscle force is accompanied by a reduction in the fibers' length. The effects of muscle shortening on the compound muscle action potential (M wave) have so far been investigated only by computer simulation. This study was undertaken to assess experimentally the M-wave changes caused by brief voluntary and stimulated isometric contractions. METHODS: Two different methods of inducing muscle shortening under isometric condition were adopted: (1) applying a brief (1 s) tetanic contraction and (2) performing brief voluntary contractions of different intensities. In both methods, supramaximal stimulation was applied to the brachial plexus and femoral nerves to evoke M waves. In the first method, electrical stimulation (20 Hz) was delivered with the muscle at rest, whereas in the second, stimulation was applied while participants performed 5-s stepwise isometric contractions at 10, 20, 30, 40, 50, 60, 70, and 100% MVC. The amplitude and duration of the first and second M-wave phases were computed. RESULTS: The main findings were: (1) on application of tetanic stimulation, the amplitude of the M-wave first phase decreased (~ 10%, P < 0.05), that of the second phase increased (~ 50%, P < 0.05), and the M-wave duration decreased (~ 20%, P < 0.05) across the first five M waves of the tetanic train and then plateaued for the subsequent responses; (2) when superimposing a single electrical stimulus on muscle contractions of increasing forces, the amplitude of the M-wave first phase decreased (~ 20%, P < 0.05), that of the second phase increased (~ 30%, P < 0.05), and M-wave duration decreased (~ 30%, P < 0.05) as force was raised from 0 to 60-70% MVC force. CONCLUSIONS: The present results will help to identify the adjustments in the M-wave profile caused by muscle shortening and also contribute to differentiate these adjustments from those caused by muscle fatigue and/or changes in Na+-K+ pump activity.

EMG probability density function: a new way to look at EMG signal filling from single motor unit potential to full interference pattern.

An analytical derivation of the EMG signal's amplitude probability density function (EMG PDF) is presented and used to study how an EMG signal builds-up, or fills, as the degree of muscle contraction increases. The EMG PDF is found to change from a semi-degenerate distribution to a Laplacian-like distribution and finally to a Gaussian-like distribution.We present a measure, the EMG filling factor, to quantify the degree to which an EMG signal has been built-up. This factor is calculated from the ratio of two non-central moments of the rectified EMG signal. The curve of the EMG filling factor as a function of the mean rectified amplitude shows a progressive and mostly linear increase during early recruitment, and saturation is observed when the EMG signal distribution becomes approximately Gaussian. Having presented the analytical tools used to derive the EMG PDF, we demonstrate the usefulness of the EMG filling factor and curve in studies with both simulated signals and real signals obtained from the tibialis anterior muscle of 10 subjects. Both simulated and real EMG filling curves start within the 0.2 to 0.35 range and rapidly rise towards 0.5 (Laplacian) before stabilizing at around 0.637 (Gaussian). Filling curves for the real signals consistently followed this pattern (100% repeatability within trials in 100% of the subjects). The theory of EMG signal filling derived in this work provides (a) an analytically consistent derivation of the EMG PDF as a function of motor unit potentials and motor unit firing patterns; (b) an explanation of the change in the EMG PDF according to degree of muscle contraction; and (c) a way (the EMG filling factor) to quantify the degree to which an EMG signal has been built-up.

Long-Range Traffic Monitoring Based on Pulse-Compression Distributed Acoustic Sensing and Advanced Vehicle Tracking and Classification Algorithm.

We introduce a novel long-range traffic monitoring system for vehicle detection, tracking, and classification based on fiber-optic distributed acoustic sensing (DAS). High resolution and long range are provided by the use of an optimized setup incorporating pulse compression, which, to our knowledge, is the first time that is applied to a traffic-monitoring DAS system. The raw data acquired with this sensor feeds an automatic vehicle detection and tracking algorithm based on a novel transformed domain that can be regarded as an evolution of the Hough Transform operating with non-binary valued signals. The detection of vehicles is performed by calculating the local maxima in the transformed domain for a given time-distance processing block of the detected signal. Then, an automatic tracking algorithm, which relies on a moving window paradigm, identifies the trajectory of the vehicle. Hence, the output of the tracking stage is a set of trajectories, each of which can be regarded as a vehicle passing event from which a vehicle signature can be extracted. This signature is unique for each vehicle, allowing us to implement a machine-learning algorithm for vehicle classification purposes. The system has been experimentally tested by performing measurements using dark fiber in a telecommunication fiber cable running in a buried conduit along 40 km of a road open to traffic. Excellent results were obtained, with a general classification rate of 97.7% for detecting vehicle passing events and 99.6% and 85.7% for specific car and truck passing events, respectively.

The filling factor of the sEMG signal at low contraction forces in the quadriceps muscles is influenced by the thickness of the subcutaneous layer.

Introduction: It has been shown that, for male subjects, the sEMG activity at low contraction forces is normally "pulsatile", i.e., formed by a few large-amplitude MUPs, coming from the most superficial motor units. The subcutaneous layer thickness, known to be greater in females than males, influences the electrode detection volume. Here, we investigated the influence of the subcutaneous layer thickness on the type of sEMG activity (pulsatile vs. continuous) at low contraction forces. Methods: Voluntary surface EMG signals were recorded from the quadriceps muscles of healthy males and females as force was gradually increased from 0% to 40% MVC. The sEMG filling process was examined by measuring the EMG filling factor, computed from the non-central moments of the rectified sEMG signal. Results: 1) The sEMG activity at low contraction forces was "continuous" in the VL, VM and RF of females, whereas this sEMG activity was "pulsatile" in the VL and VM of males. 2) The filling factor at low contraction forces was lower in males than females for the VL (p = 0.003) and VM (p = 0.002), but not for the RF (p = 0.54). 3) The subcutaneous layer was significantly thicker in females than males for the VL (p = 0.001), VM (p = 0.001), and RF (p = 0.003). 4) A significant correlation was found in the vastus muscles between the subcutaneous layer thickness and the filling factor (p < 0.05). Discussion: The present results indicate that the sEMG activity at low contraction forces in the female quadriceps muscles is "continuous" due to the thick subcutaneous layer of these muscles, which impedes an accurate assessment of the sEMG filling process.

Automatic jitter measurement in needle-detected motor unit potential trains.

In an active motor unit (MU), the time intervals between the firings of its muscle fibers vary across successive MU activations. This variability is called jitter and is increased in pathological processes that affect the neuromuscular junctions or terminal axonal segments of MUs. Traditionally, jitter has been measured using single fiber electrodes (SFEs) and a difficult and subjective manual technique. SFEs are expensive and reused, implying a potential risk of patient infection; so, they are being gradually substituted by safer, disposable, concentric needle electrodes (CNEs). As CNEs are larger, voltage contributions from individual fibers of a MU are more difficult to detect, making jitter measurement more difficult. This paper presents an automatic method to estimate jitter from trains of motor unit potentials (MUPs), for both SFE and CNE records. For a MUP train, segments of MUPs generated by single muscle fibers (SF MUP segments) are found and jitter is measured between pairs of these segments. Segments whose estimated jitter values are not reliable, according to several SF MUP segment characteristics, are excluded. The method has been tested in several simulation studies that use mathematical models of muscle fiber potentials. The results are very satisfactory in terms of jitter estimation error (less than 10% in most of the cases studied) and mean number of valid jitter estimates obtained per simulated train (greater than 1.0 in many of the cases and less than 0.5 only in the most complicated). A preliminary study with real signals was also performed, using 19 MUP trains from 3 neuropathic patients. Jitter measurements obtained by the automatic method were compared with those extracted from a commercial system (Keypoint) and the edition and supervision of an expert electromyographer. From these measurements 63% were taken from equivalent interval pair sites within the time span of the MUP trains and, as such, were considered as compatible measurements. Differences in jitter of these compatible measurements were very low (mean value of 1.3 µs, mean of absolute differences of 2.97 µs, 25% and 75% percentile intervals of -0.85 and 3.82 µs, respectively). Although new tests with larger number of real recordings are still required, the method seems promising for clinical practice.

The influence of the reference electrode location on the M-wave characteristics in the quadriceps.

INTRODUCTION: In the compound muscle action potential (M wave) recorded using the belly-tendon configuration, the contribution of the tendon electrode is assumed to be negligible compared to the belly electrode. We tested this assumption by placing the reference electrode at a distant (contralateral) site, which allowed separate recording of the belly and tendon contributions. METHODS: M waves were recorded at multiple selected sites over the right quadriceps heads and lower leg using two different locations for the reference electrode: the ipsilateral (right) and contralateral (left) patellar tendon. The general parameters of the M wave (amplitude, area, duration, latency, and frequency) were measured. RESULTS: (1) The tendon potential had a small amplitude (<30%) compared to the belly potential; (2) Changing the reference electrode from the ipsilateral to the contralateral patella produced moderate changes in the M wave recorded over the innervation zone, these changes affecting significantly the amplitude of the M-wave second phase (p = 0.006); (3) Using the contralateral reference system allowed recording of short-latency components occurring immediately after the stimulus artefact, which had the same latency and amplitude (p = 0.18 and 0.25, respectively) at all recording sites over the leg. CONCLUSIONS: The potential recorded at the "tendon" site after femoral nerve stimulation is small (compared to the belly potential), but not negligible, and makes a significant contribution to the second phase of belly-tendon M wave. Adopting a distant (contralateral) reference allowed recording of far-field components that may aid in the understanding of the electrical formation of the M wave.

Effects of muscle shortening on single-fiber, motor unit, and compound muscle action potentials.

Even under isometric conditions, muscle contractions are associated with some degree of fiber shortening. The effects of muscle shortening on extracellular electromyographic potentials have not been characterized in detail. Moreover, the anatomical, biophysical, and detection factors influencing the muscle-shortening effects have been neither identified nor understood completely. Herein, we investigated the effects of muscle shortening on the amplitude and duration characteristics of single-fiber, motor unit, and compound muscle action potentials. We found that, at the single-fiber level, two main factors influenced the muscle-shortening effects: (1) the electrode position and distance relative to the myotendinous zone and (2) the electrode distance to the maxima of the dipole field arising from the stationary dipole created at the fiber-tendon junction. Besides, at the motor unit and muscle level, two additional factors were involved: (3) the overlapping between the propagating component of some fibers with the non-propagating component of other fibers and (4) the spatial spreading of the fiber-tendon junctions. The muscle-shortening effects depend critically on the electrode longitudinal distance to the myotendinous zone. When the electrode was placed far from the myotendinous zone, muscle shortening resulted in an enlargement and narrowing of the final (negative) phase of the potential, and this enlargement became less pronounced as the electrode approached the fiber endings. For electrode locations close to the myotendinous zone, muscle shortening caused a depression of both the main (positive) and final (negative) phases of the potential. Beyond the myotendinous zone, muscle shortening led to a decrease of the final (positive) phase. The present results provide reference information that will help to identify changes in MUPs and M waves due to muscle shortening, and thus to differentiate these changes from those caused by muscle fatigue.

Masked least-squares averaging in processing of scanning-EMG recordings with multiple discharges.

Removing artifacts from nearby motor units is one of the main objectives when processing scanning-EMG recordings. Methods such as median filtering or masked least-squares smoothing (MLSS) can be used to eliminate artifacts in recordings with just one discharge of the motor unit potential (MUP) at each location. However, more effective artifact removal can be achieved if several discharges per position are recorded. In this case, processing usually involves averaging the discharges available at each position and then applying a median filter in the spatial dimension. The main drawback of this approach is that the median filter tends to distort the signal waveform. In this paper, we present a new algorithm that operates on multiple discharges simultaneously and in the spatial dimension. We refer to this algorithm as the multi-masked least-squares smoothing (MMLSS) algorithm: an extension of the MLSS algorithm for the case of multiple discharges. The algorithm is tested using simulated scanning-EMG signals in different recording conditions, i.e., at different levels of muscle contraction and for different numbers of discharges per position. The results demonstrate that the algorithm eliminates artifacts more effectively than any previously available method and does so without distorting the waveform of the signal. Graphical abstract The raw scanning-EMG signal, which can be composed by several discharges of the MU, is processed by the MMLSS algorithm so as to eliminate the artifact interference. Firstly, artifacts are detected for each discharge from the raw signal, obtaining a multi-discharge validity mask that indicates the samples that have been corrupted by artifacts. Secondly, a least-squares smoothing procedure simultaneously operating in the spatial dimension and among the discharges is applied to the raw signal. This second step is performed using only the not contaminated samples according to the validity mask. The resulting MMLSS-processed scanning-EMG signal is clean of artifact interference.

Recovery of the first and second phases of the M wave after prolonged maximal voluntary contractions.

INTRODUCTION: We compared the recovery of muscle electrical properties after maximal voluntary contractions (MVCs) of 1 and 3 min duration by examining separately the first and second phases of the muscle compound action potential (M wave). METHODS: M waves were evoked by supramaximal single shocks to the femoral nerve throughout the 30-min recovery following 1-min and 3-min MVCs. The amplitude, duration, and area of the M-wave first and second phases, along with peak-to-peak amplitude and total area, were measured from the knee extensors. RESULTS: (1) The amplitude of the M-wave first phase increased to the same extent (and had the same time course of recovery) after the 1 and 3-min MVCs, whereas the amplitude of the second phase increased more markedly after the 1-min than after the 3-min MVC (P < 0.05). (2) The first phase remained enlarged for 2 min after exercise, whereas the augmentation of the second phase only lasted for 30 s. (3) After 30 min of recovery, the amplitude, area, and duration of both the first and second phases were decreased compared to control values (P < 0.05). CONCLUSIONS: The similar enlargement of the M-wave first phase after the 1 and 3-min MVCs suggests that the extracellular K+ concentration attained after these contractions was similar. The mechanisms responsible for the long-term decreases in M-wave amplitude and duration are unknown at present, but are likely due to a decrease in the amplitude of individual transmembrane potentials and an increase in conduction velocity.

A new method for measurement of motor unit action potential duration based on correlation, a pilot study.

We present a new, automatic, correlation-based method for measuring the duration of motor unit action potentials (MUAPs). The method seeks to replicate the way an expert elctromyographer uses his or her eyes, calculating the start and end of the MUAP waveform on the basis of the degree of similarity of non-excluded discharges. We analysed 68 potentials from normal deltoid muscles during slight contraction. For each MUAP, two experienced electromyographers manually determined start and end marker positions, which were used as gold standard duration positions (GSP) in our subsequent tests. The novel method was compared with Nandedkar's method and a wavelet transform-based method. To compare the three methods, the differences between the automatic marker positions and GSPs were statistically evaluated using one-factor ANOVA, the estimated mean square error, and a Chi-square test on the numbers of automatic marker placements with gross errors. All these parameters showed smaller values for the novel method and in most of the cases were statistically significant. In addition, the parameters of the new method were subjected to a sensitivity study, showing its good performance within a range of clinically useful parameter values. The new automatic method determined start and end markers in a more accurate and reliable manner than both of the acknowledged state-of-the art methods used in our comparison study. Graphical abstract The description of a new automatic duration measurement algorithm based on the similarity among discharges of the same MUAP. This method gave better results than the Nandedkar method and a highly regarded wavelet-based method. The new correlation-based method also had the lowest rate of gross aberrant errors in automatic placements.

Exact inter-discharge interval distribution of motor unit firing patterns with gamma model.

Inter-discharge interval distribution modeling of the motor unit firing pattern plays an important role in electromyographic decomposition and the statistical analysis of firing patterns. When modeling firing patterns obtained from automatic procedures, false positives and false negatives can be taken into account to enhance performance in estimating firing pattern statistics. Available models of this type, however, are only approximate and use Gaussian distributions, which are not strictly suitable for modeling renewal point processes. In this paper, the theory of point processes is used to derive an exact solution to the distribution when a gamma distribution is used to model the physiological firing pattern. Besides being exact, the solution provides a way to model the skewness of the inter-discharge distribution, and this may make it possible to obtain a better fit with available experimental data. In order to demonstrate potential applications of the model, we use it to obtain a maximum likelihood estimator of firing pattern statistics. Our tests found this estimator to be reliable over a wide range of firing conditions, whether dealing with real or simulated firing patterns, the proposed solution had better agreement than other models. Graphical Abstract Model of the MU firing pattern generation and detection: fT,1(τ), IDI PDF of the physiological firing pattern; fT(τ), IDI PDF after modeling undetected firings (false negatives); fS(τ), IDI PDF after modeling classification errors (false positives).

Sliding window averaging in normal and pathological motor unit action potential trains.

OBJECTIVE: To evaluate the performance of a recently proposed motor unit action potential (MUAP) averaging method based on a sliding window, and compare it with relevant published methods in normal and pathological muscles. METHODS: Three versions of the method (with different window lengths) were compared to three relevant published methods in terms of signal analysis-based merit figures and MUAP waveform parameters used in the clinical practice. 218 MUAP trains recorded from normal, myopathic, subacute neurogenic and chronic neurogenic muscles were analysed. Percentage scores of the cases in which the methods obtained the best performance or a performance not significantly worse than the best were computed. RESULTS: For signal processing figures of merit, the three versions of the new method performed better (with scores of 100, 86.6 and 66.7%) than the other three methods (66.7, 25 and 0%, respectively). In terms of MUAP waveform parameters, the new method also performed better (100, 95.8 and 91.7%) than the other methods (83.3, 37.5 and 25%). CONCLUSIONS: For the types of normal and pathological muscle studied, the sliding window approach extracted more accurate and reliable MUAP curves than other existing methods. SIGNIFICANCE: The new method can be of service in quantitative EMG.

A masked least-squares smoothing procedure for artifact reduction in scanning-EMG recordings.

Scanning-EMG is an electrophysiological technique in which the electrical activity of the motor unit is recorded at multiple points along a corridor crossing the motor unit territory. Correct analysis of the scanning-EMG signal requires prior elimination of interference from nearby motor units. Although the traditional processing based on the median filtering is effective in removing such interference, it distorts the physiological waveform of the scanning-EMG signal. In this study, we describe a new scanning-EMG signal processing algorithm that preserves the physiological signal waveform while effectively removing interference from other motor units. To obtain a cleaned-up version of the scanning signal, the masked least-squares smoothing (MLSS) algorithm recalculates and replaces each sample value of the signal using a least-squares smoothing in the spatial dimension, taking into account the information of only those samples that are not contaminated with activity of other motor units. The performance of the new algorithm with simulated scanning-EMG signals is studied and compared with the performance of the median algorithm and tested with real scanning signals. Results show that the MLSS algorithm distorts the waveform of the scanning-EMG signal much less than the median algorithm (approximately 3.5 dB gain), being at the same time very effective at removing interference components. Graphical Abstract The raw scanning-EMG signal (left figure) is processed by the MLSS algorithm in order to remove the artifact interference. Firstly, artifacts are detected from the raw signal, obtaining a validity mask (central figure) that determines the samples that have been contaminated by artifacts. Secondly, a least-squares smoothing procedure in the spatial dimension is applied to the raw signal using the not contaminated samples according to the validity mask. The resulting MLSS-processed scanning-EMG signal (right figure) is clean of artifact interference.

Influence of timing variability between motor unit potentials on M-wave characteristics.

The transient enlargement of the compound muscle action potential (M wave) after a conditioning contraction is referred to as potentiation. It has been recently shown that the potentiation of the first and second phases of a monopolar M wave differed drastically; namely, the first phase remained largely unchanged, whereas the second phase underwent a marked enlargement and shortening. This dissimilar potentiation of the first and second phases has been suggested to be attributed to a transient increase in conduction velocity after the contraction. Here, we present a series of simulations to test if changes in the timing variability between motor unit potentials (MUPs) can be responsible for the unequal potentiation (and shortening) of the first and the second M-wave phases. We found that an increase in the mean motor unit conduction velocity resulted in a marked enlargement and narrowing of both the first and second M-wave phases. The enlargement of the first phase caused by a global increase in motor unit conduction velocities was apparent even for the electrode located over the innervation zone and became more pronounced with increasing distance to the innervation zone, whereas the potentiation of the second phase was largely independent of electrode position. Our simulations indicate that it is unlikely that an increase in motor unit conduction velocities (accompanied or not by changes in their distribution) could account for the experimental observation that only the second phase of a monopolar M wave, but not the first, is enlarged after a brief contraction. However, the combination of an increase in the motor unit conduction velocities and a spreading of the motor unit activation times could potentially explain the asymmetric potentiation of the M-wave phases.

Limitations of Spectral Electromyogramic Analysis to Determine the Onset of Neuromuscular Fatigue Threshold during Incremental Ergometer Cycling.

Recently, a new method has been proposed to detect the onset of neuromuscular fatigue during an incremental cycling test by assessing the changes in spectral electromyographic (sEMG) frequencies within individual exercise periods of the test. The method consists on determining the highest power output that can be sustained without a significant decrease in spectral frequencies. This study evaluated the validity of the new approach by assessing the changes in spectral indicators both throughout the whole test and within individual exercise periods of the test. Fourteen cyclists performed incremental cycle ergometer rides to exhaustion with bipolar surface EMG signals recorded from the vastus lateralis. The mean and median frequencies (Fmean and Fmedian, respectively) of the sEMG power spectrum were calculated. The main findings were: (1) Examination of spectral indicators within individual exercise periods of the test showed that neither Fmean nor Fmedian decreased significantly during the last (most fatiguing) exercise periods. (2) Examination of the whole incremental test showed that the behaviour of Fmean and Fmedian with increasing power output was highly inconsistent and varied greatly among subjects. (3) Over the whole incremental test, half of the participants exhibited a positive relation between spectral indicators and workload, whereas the other half demonstrated the opposite behavior. Collectively, these findings indicate that spectral sEMG indexes do not provide a reliable measure of the fatigue state of the muscle during an incremental cycling test. Moreover, it is concluded that it is not possible to determine the onset of neuromuscular fatigue during an incremental cycling test by examining spectral indicators within individual exercise periods of the test. Key pointsThe behaviour of spectral EMG indicators during the incremental test exhibited a high heterogeneity among individuals, with approximately half of the participants showing a positive relation between spectral indicators and workload and the other half showing the opposite behaviour.None of the spectral EMG indicators examined (Fmean nor Fmedian) decreased significantly between the ventilatory threshold and the highest power output.Examination of spectral indicators within individual exercise periods of the test showed that neither Fmean nor Fmedian decreased significantly during the last (most fatiguing) exercise periods.

Ultrasound Image Discrimination between Benign and Malignant Adnexal Masses Based on a Neural Network Approach.

The discrimination between benign and malignant adnexal masses in ultrasound images represents one of the most challenging problems in gynecologic practice. In the study described here, a new method for automatic discrimination of adnexal masses based on a neural networks approach was tested. The proposed method first calculates seven different types of characteristics (local binary pattern, fractal dimension, entropy, invariant moments, gray level co-occurrence matrix, law texture energy and Gabor wavelet) from ultrasound images of the ovary, from which several features are extracted and collected together with the clinical patient age. The proposed technique was validated using 106 benign and 39 malignant images obtained from 145 patients, corresponding to its probability of appearance in general population. On evaluation of the classifier, an accuracy of 98.78%, sensitivity of 98.50%, specificity of 98.90% and area under the curve of 0.997 were calculated.

Averaging methods for extracting representative waveforms from motor unit action potential trains.

In the context of quantitative electromyography (EMG), it is of major interest to obtain a waveform that faithfully represents the set of potentials that constitute a motor unit action potential (MUAP) train. From this waveform, various parameters can be determined in order to characterize the MUAP for diagnostic analysis. The aim of this work was to conduct a thorough, in-depth review, evaluation and comparison of state-of-the-art methods for composing waveforms representative of MUAP trains. We evaluated nine averaging methods: Ensemble (EA), Median (MA), Weighted (WA), Five-closest (FCA), MultiMUP (MMA), Split-sweep median (SSMA), Sorted (SA), Trimmed (TA) and Robust (RA) in terms of three general-purpose signal processing figures of merit (SPMF) and seven clinically-used MUAP waveform parameters (MWP). The convergence rate of the methods was assessed as the number of potentials per MUAP train (NPM) required to reach a level of performance that was not significantly improved by increasing this number. Test material comprised 78 MUAP trains obtained from the tibialis anterioris of seven healthy subjects. Error measurements related to all SPMF and MWP parameters except MUAP amplitude descended asymptotically with increasing NPM for all methods. MUAP amplitude showed a consistent bias (around 4% for EA and SA and 1-2% for the rest). MA, TA and SSMA had the lowest SPMF and MWP error figures. Therefore, these methods most accurately preserve and represent MUAP physiological information of utility in clinical medical practice. The other methods, particularly WA, performed noticeably worse. Convergence rate was similar for all methods, with NPM values averaged among the nine methods, which ranged from 10 to 40, depending on the waveform parameter evaluated.

Inter-discharge interval distribution of motor unit firing patterns with detection errors.

Inter-discharge interval (IDI) distribution analysis of motor unit firing patterns is a valuable tool in EMG decomposition and analysis. However, the firing pattern obtained by EMG decomposition may have detection errors: false positives (incorrectly classified firings) and false negatives (missed firings). In this paper, the mathematical derivation of an IDI distribution model that accommodates false positives and false negatives of the detection process is presented. An approximation of the general model to adapt to specific EMG decomposition conditions is also presented. To illustrate the usefulness of the model, the obtained distribution is used to derive the maximum likelihood estimates of the statistics of motor unit firing patterns, the IDI mean and standard deviation, and estimates of the false negative and false positive ratios. Results obtained from simulation experiments and tests with real motor unit firing patterns show an enhanced estimation performance when compared to previously available algorithms. Goodness-of-fit tests applied to estimations for real data corrupted with false positives showed that the model-driven estimations fitted the uncorrupted data better than EFE estimations: 82% versus 52% not rejectable, respectively, when false positives were about 10% of IDIs. With about 5% false positives, the not rejectable estimations were 85% versus 70%.