An examination of the Runs Test, Reverse Arrangements Test, and modified Reverse Arrangements Test for assessing surface EMG signal stationarity

Evaluation of windowing techniques for intramuscular EMG-based diagnostic, rehabilitative and assistive devices

Objective.Intramuscular electromyography (iEMG) signals, invasively recorded, directly from the muscles are used to diagnose various neuromuscular disorders/diseases and to control rehabilitative and assistive robotic devices. iEMG signals are potentially being used in neurology, kinesiology, rehabilitation and ergonomics, to detect/diagnose various diseases/disorders. Electromyography-based classification and analysis systems are being designed and tested for the classification of various neuromuscular disorders and to control rehabilitative and assistive robotic devices. Many studies have explored parameters such as the pre-processing, feature extraction and selection of classifiers that can affect the performance and efficacy of iEMG-based classification systems. The pre-processing stage includes the removal of any unwanted noise from the original signal and windowing of the signal.Approach.This study investigated and presented the optimum windowing configurations for robust control and better performance results of an iEMG-based analysis system based on the stationarity rate (SR) and classification accuracy. Both disjoint and overlap, windowing techniques with varying window and overlap sizes have been investigated using a machine learning-based classification algorithm called linear discriminant analysis.Main results.The optimum window size ranges are from 200-300 ms for the disjoint and 225-300 ms for the overlap windowing technique, respectively. The inferred results show that for the overlap windowing technique the optimum range of overlap size is from 10%-30% of the length of window size. The mean classification accuracy (MCA) and mean stationarity rate (MSR) were found to be lower in the disjoint windowing technique compared to overlap windowing at all investigated overlap sizes. Statistical analysis (two-way analysis of variance test) showed that the MSR and MCA of the overlap windowing technique was significantly different at overlap sizes of 10%-30% (p-values < 0.05).Significance.The presented results can be used to achieve the best possible classification results and SR for any iEMG-based real-time diagnosis, detection and control system, which can enhance the performance of the system significantly.

The Profiles of Non-stationarity and Non-linearity in the Time Series of Resting-State Brain Networks

The linearity and stationarity of fMRI time series need to be understood due to their important roles in the choice of approach for brain network analysis. In this paper, we investigated the stationarity and linearity of resting-state fMRI (rs-fMRI) time-series data from the Midnight Scan Club datasets. The degree of stationarity (DS) and the degree of non-linearity (DN) were, respectively, estimated for the time series of all gray matter voxels. The similarity and difference between the DS and DN were assessed in terms of voxels and intrinsic brain networks, including the visual network, somatomotor network, dorsal attention network, ventral attention network, limbic network, frontoparietal network, and default-mode network. The test-retest scans were utilized to quantify the reliability of DS and DN. We found that DS and DN maps had overlapping spatial distribution. Meanwhile, the probability density estimate function of DS had a long tail, and that of DN had a more normal distribution. Specifically, stronger DS was present in the somatomotor, limbic, and ventral attention networks compared to other networks, and stronger DN was found in the somatomotor, visual, limbic, ventral attention, and default-mode networks. The percentage of overlapping voxels between DS and DN in different networks demonstrated a decreasing trend in the order default mode, ventral attention, somatomotor, frontoparietal, dorsal attention, visual, and limbic. Furthermore, the ICC values of DS were higher than those of DN. Our results suggest that different functional networks have distinct properties of non-stationarity and non-linearity owing to the complexity of rs-fMRI time series. Thus, caution should be taken when analyzing fMRI data (both resting-state and task-activation) using simplified models.

Feature extraction of the first difference of EMG time series for EMG pattern recognition

This paper demonstrates the utility of a differencing technique to transform surface EMG signals measured during both static and dynamic contractions such that they become more stationary. The technique was evaluated by three stationarity tests consisting of the variation of two statistical properties, i.e., mean and standard deviation, and the reverse arrangements test. As a result of the proposed technique, the first difference of EMG time series became more stationary compared to the original measured signal. Based on this finding, the performance of time-domain features extracted from raw and transformed EMG was investigated via an EMG classification problem (i.e., eight dynamic motions and four EMG channels) on data from 18 subjects. The results show that the classification accuracies of all features extracted from the transformed signals were higher than features extracted from the original signals for six different classifiers including quadratic discriminant analysis. On average, the proposed differencing technique improved classification accuracies by 2-8%.

An investigation of fMRI time series stationarity during motor sequence learning foot tapping tasks

BACKGROUND

Understanding complex brain networks using functional magnetic resonance imaging (fMRI) is of great interest to clinical and scientific communities. To utilize advanced analysis methods such as graph theory for these investigations, the stationarity of fMRI time series needs to be understood as it has important implications on the choice of appropriate approaches for the analysis of complex brain networks.

NEW METHOD

In this paper, we investigated the stationarity of fMRI time series acquired from twelve healthy participants while they performed a motor (foot tapping sequence) learning task. Since prior studies have documented that learning is associated with systematic changes in brain activation, a sequence learning task is an optimal paradigm to assess the degree of non-stationarity in fMRI time-series in clinically relevant brain areas. We predicted that brain regions involved in a "learning network" would demonstrate non-stationarity and may violate assumptions associated with some advanced analysis approaches. Six blocks of learning, and six control blocks of a foot tapping sequence were performed in a fixed order. The reverse arrangement test was utilized to investigate the time series stationarity.

RESULTS

Our analysis showed some non-stationary signals with a time varying first moment as a major source of non-stationarity. We also demonstrated a decreased number of non-stationarities in the third block as a result of priming and repetition.

COMPARISON WITH EXISTING METHODS

Most of the current literature does not examine stationarity prior to processing.

CONCLUSIONS

The implication of our findings is that future investigations analyzing complex brain networks should utilize approaches robust to non-stationarities, as graph-theoretical approaches can be sensitive to non-stationarities present in data.

Non-stationarity of EEG during wakefulness and anaesthesia: advantages of EEG permutation entropy monitoring

Monitors evaluating the electroencephalogram (EEG) to determine depth of anaesthesia use spectral analysis approaches for analysis windows up to 61.5 s as well as additional smoothing algorithms. Stationary EEG is required to reliably apply the index algorithms. Because of rapid physiological changes, artefacts, etc., the EEG may not always fulfil this requirement. EEG analysis using permutation entropy (PeEn) may overcome this issue, since PeEn can also be applied to practically nonstationary EEG. One objective was to determine the duration of EEG sequences that can be considered stationary at different anaesthetic levels. The second, more important objective was to test the reliability of PeEn to reflect the anaesthetic levels for short EEG segments. EEG was recorded from 15 volunteers undergoing sevoflurane and propofol anaesthesia at different anaesthetic levels and for each group 10 data sets were included. EEG stationarity was evaluated for EEG sample lengths from 4 to 116 s for each level. PeEn was calculated for these sequences using different parameter settings and analysis windows from 2 to 60 s. During wakefulness EEG can only be considered stationary for sequences up to 12 s. With increasing anaesthetic level the probability and duration of stationary EEG increases. PeEn is able to reliably separate consciousness from unconsciousness for EEG segments as short as 2 s. Especially during wakefulness a conflict between stationary EEG sequence durations and methods used for monitoring may exist. PeEn does not require stationarity and functions for EEG sequences as short as 2 s. These promising results seem to support the application of non-linear parameters, such as PeEn, to depth of anaesthesia monitoring.

Measures of radial correlation and trend for classification of breast masses in mammograms

In this paper, a novel approach for classification of breast masses is presented that quantifies the texture of masses without relying on accurate extraction of their contours. Two novel feature descriptors based on 2D extensions of the reverse arrangement (RA) and Mantel's tests were designed for this purpose. Measures of radial correlation and radial trend were extracted from the original gray-scale values as well as from the Gabor magnitude response of 146 regions of interest, including 120 benign masses and 26 malignant tumors. Four classifiers, Fisher-linear discriminant analysis, Bayesian, support vector machine, and an artificial neural network based on radial basis functions (ANN-RBF), were employed to predict the diagnosis, using stepwise logistic regression for feature selection and the leave-one-patient-out method for cross-validation. The ANN-RBF resulted in an area under the receiver operating characteristic curve of 0.93. The experimental results show the effectiveness of the proposed approach.

Isometric fatigue patterns in time and time-frequency domains of triceps surae muscle in different knee positions

The occurrence of fatigue in triceps surae (TS) muscles during sustained plantar flexion contraction is investigated by means of the RMS electromyogram (EMG) and the instantaneous median frequency (IMF) of the short time Fourier transform (STFT). Six male subjects realized a 40% maximal plantar flexion isometric voluntary contraction until fatigue in two knee positions. Electrodes were positioned on gastrocnemius medialis, gastrocnemius lateralis and soleus muscles. The torque (TO) and EMG signals were synchronized. The RMS and the median of the IMF values were obtained, respectively, for each 250 ms and 1s windows of signal. Each signal was segmented into 10 epochs, from which the mean values of IMF, RMS and TO were obtained and submitted to linear regressions to determine parameter trends. Friedman test with the Dunn's post hoc were used to test for differences among muscles activation for each knee position and among slopes of regression curves, as well as to observe changes in TS RMS values over time. The results indicate different activation strategies with the knee extended (KE) in contrast to knee flexed (KF). With the KE, the gastrocnemii showed typical fatigue behavior with significant (p<0.05) IMF reductions and RMS increases over time, while soleus showed concomitant RMS and IMF increases (p<0.05) suggesting an increased soleus contribution to the torque production. With KF, the gastrocnemii were under activated, increasing the role of soleus. Thus, time-frequency analysis represented an important tool for TS muscular fatigue evaluation, allowing differentiates the role of soleus muscle.

Time and time-frequency characterization of dual-axis swallowing accelerometry signals

Single-axis swallowing accelerometry has shown potential as a non-invasive clinical swallowing assessment tool. Previous swallowing accelerometry research has focused exclusively on the anterior-posterior vibration detected on the surface of the neck. However, hyolaryngeal motion during pharyngeal swallowing occurs in both the anterior-posterior and superior-inferior directions, suggesting that dual-axis accelerometry may be worthy of investigation. With this motivation, the present paper provides a characterization of dual-axis swallowing accelerometry signals from healthy adults in the time and time-frequency domains. Time-domain analysis revealed that signals in the two axes exhibited different probability density functions, and minimal cross-correlation and mutual information. Time-frequency analysis highlighted inter-axis dissimilarities in the scalograms, pseudo-spectra and temporal evolution of low- and high-frequency content. Therefore, it was concluded that the two axes contain different information about swallowing and that the superior-inferior axis should be further investigated in future swallowing accelerometry studies.