Ensemble averaging of locomotor electromyographic patterns using interpolation

Neuromuscular alterations during walking in persons with moderate knee osteoarthritis

This paper compared the neuromuscular responses during walking between those with early-stage knee osteoarthritis (OA) to asymptomatic controls. The rationale for studying those with mild to moderate knee OA was to determine the alterations in response to dynamic loading that might be expected before severe pain, joint space narrowing and joint surface changes occur. We used pattern recognition techniques to explore both amplitude and shape changes of the surface electromyograms recorded from seven muscles crossing the knee joint of 40 subjects with knee OA and 38 asymptomatic controls during a walking task. The principal patterns for each muscle grouping explained over 83% of the variance in the waveforms. This result supported the notion that the main neuromuscular patterns were similar between asymptomatic controls and those with OA, reflecting the specific roles of the major muscles during walking. ANOVA revealed significant (p<0.05) differences in the principal pattern scores reflecting both amplitude and shape alterations in the OA group and among muscles. These differences captured subtle changes in the neuromuscular responses of the subjects with OA throughout different phases of the gait cycle and most likely reflected changes in the mechanical environment (joint loading, instability) and pain. The subjects with OA attempted to increase activity of the lateral sites and reduce activity in the medial sites, having minimal but prolonged activity during late stance. Therefore, alterations in neuromuscular responses were found even in this high functioning group with moderate knee OA.

Speed dependence of averaged EMG profiles in walking

Electromyogram (EMG) profiles strongly depend on walking speed and, in pathological gait, patients do not usually walk at normal speeds. EMG data was collected from 14 muscles in two groups of healthy young subjects who walked at five different speeds ranging from 0.75 to 1.75 ms(-1). We found that average EMG profiles varied in a predictable way with speed. The average EMG profile for each muscle at any speed could be estimated in a simple way from two functions, one constant and one proportionally increasing with walking speed. By taking into account the similarity among profiles within functional groups, the number of basic functions could be reduced further. Any average EMG profile among the 14 leg muscles studied at all speeds in the measured range could be predicted from six constant and ten speed-dependent basic patterns. These results can be interpreted in terms of a central pattern generator for human walking.

Application of the LaGrange polynomial in skeletal muscle fatigue analysis

The percentage of decrement in torque and the number of serial contractions are mutually exclusive methodological controls in the study of muscular fatigue. This paper examines the feasibility of using the LaGrange polynomial in the analysis of voluntary muscular fatigue patterns. Twenty-one men (ages 20-60 years) reported to the orthopedic biomechanics laboratory on 2 days separated by 4 months. During both sessions, participants completed three maximal isokinetic (180 deg x s(-1)) contractions of the knee extensors to serve as baseline, before starting the fatigue protocol. The fatigue protocol consisted of serial contractions until a 50% strength decrement was reached. The LaGrange polynomial was first used to interpolate the individual fatigue pattern for each participant into 15 data points (trials). Data analysis was then conducted on these 15 data points. Intraclass correlation analysis of variance showed that the reliability of baseline torque was very good (.93). Baseline torque, the average of three trials, exhibited a 5.4 Nm (6%) increase from the first to second test session (p < .05). The mean level of torque, average of the 15-point fatigue pattern, also increased 7.5 Nm (15%) on the second test session (p < .05). The classic torque deficit for the first trial of a fatigue series was preserved by the interpolation method. Serial contractions resulted in an average decrease in torque of 29.5 Nm (50%) from the first to last trial (p <.05). The interpolation method also retained the linear and quadratic trend components commonly observed for isometric and isokinetic fatigue patterns. The two trend components accounted for 94.7% of the total trial variance. It was concluded that the LaGrange polynomial used to interpolate fatigue patterns to fewer data points was successful.

EMG signals detection and processing for on-line control of functional electrical stimulation

The surface EMG signal detected from voluntarily activated muscles can be used as a control signal for functional neuromuscular electrical stimulation. A proper positioning of the recording electrodes in relation to the stimulation electrodes, and a proper processing of the recorded signals is required to reduce the stimulus artefact and the non-voluntary contribution (M-wave). Six orientations and six locations of the recording electrodes were investigated in the present work. A comb filter (with and without a blanking windowing) was applied to remove the signal components synchronously correlated to the stimulus. An operative definition of the signal to noise ratio and an efficiency index were implemented. It resulted that when the recording electrodes were located within the two stimulation electrodes the best orientation was perpendicular to the longitudinal line. However the best absolute indexes were obtained when the recording electrodes were located externally of the stimulation electrodes, and in that case the best orientation was longitudinal. Concerning the filtering procedure, the use of a blanking window before the application of the comb filter, gave the best performance.

Dominant pattern extraction from 3-D kinematic data

A new method for the extraction of a repeating pattern in cyclic biomechanical data is proposed--singular value decomposition pattern analysis (SVDPA). This method is based on the recent work of Kanjilal and Palit [14], [15] and can be applied to both contiguous and repeated trials without being constrained to be strictly periodic. SVDPA is a data-driven approach that does not use a preselected set of basis functions; but instead utilizes a data matrix with a special structure to identify repeating patterns. Several important features of SVDPA are described including its close relationship to the Kahunen-Loève transform. The dominant pattern is defined as the average energy component (AEC). The AEC is obtained from the SVD of the data matrix and is equivalent to the optimal [maximal signal-to-noise ratio (SNR)] ensemble average pattern. The degree of periodicity and SNR for the AEC are defined explicitly from the singular values of the data matrix. We illustrate the usefulness of SVDPA for dominant pattern extraction by applying it to the quasiperiodic three-dimensional trajectory of a marker attached to the trunk during treadmill locomotion. The AEC obtained for the normalized trajectory and error estimates at each point suggests that SVDPA could be a useful tool for the extraction of the fine details from cyclic biomechanical data.

Electromyographic signals during gait: criteria for envelope filtering and number of strides

The use of linear envelopes to represent the electromyographic (EMG) measurements obtained during locomotion has become common practice. Guidelines for designing envelope filters and specifying the minimum number of strides needed to produce valid EMG profiles have been developed. Electromyograms from eight major muscles of the lower leg are measured from five normal young adults during self-selected slow, free and fast walking speeds. 30 strides per task are measured. The 'ideal' EMG profile is defined from the ensemble average of the rectified EMG signal. An error measure is defined and used as a criterion to assess the appropriateness of various cut-off frequencies for envelope filters and the number of strides required for establishing a good EMG profile. It is found that between six and ten strides are needed to form a representative profile, and an envelope filter with a minimum cut-off frequency of approximately 9 Hz is necessary.

The effects of signal conditioning on the statistical analyses of gait EMG

Ensemble averaged EMG profiles generated for leg muscles during gait have been used to clinically assess disease or injury. Several of the methods that have been reported for conditioning gait EMG signals were compared using data collected from clinically normal subjects walking on a treadmill. Specifically investigated were the effects of filtering and the quantity of data averaged upon several statistical tests that measure the variability of, or differences between, EMG profiles. Our results suggest that the variance ratio (VR) provides a reasonable test of data variability because of its modest sensitivity to both the degree of filtering and the amount of data averaged. They also suggest that of the comparison statistics: Pearson's r, the Kolmogorov-Smirnov T test and the ANOVA F ratio, the T test was the most reliable in detecting differences between given profiles for all test conditions. However, recognition of this ability of the T test must be tempered by the knowledge that while obvious EMG signal differences did exist, observable functional differences in gait did not. The relationship between statistically similar/dissimilar EMG patterns and clinically functional/dysfunctional gait patterns needs to be established. In addition, since all of the test statistics studied were affected to some degree by filtering and averaging, care should be used when comparing statistical results from separate studies unless it is known that the studies were conducted under similar conditions, including data processing. To that end, we recommend that at least 20 strides be used in the averaging process since the statistics we tested have reached or are asymptotically approaching their final values by this point.

The selection of a standard convention for analyzing gait data based on the analysis of relevant biomechanical factors

The gait cycle is usually analyzed using one of two conventions: heel contact initiates the cycle and stance is followed by swing phases or toe-off initiates the cycle and swing then stance phases are analyzed. The purpose of the study is to demonstrate that based on a consideration of relevant biomechanical measures, a convention identifying toe-off as the beginning of the gait cycle and presenting swing followed by stance phases is the superior convention. Net joint torques and electromyographic (EMG) data from selected muscles in the lower extremity were obtained from four subjects while walking and running. Data were collected for consecutive swing, stance and swing phases to compare the variables at the swing-to-stance and stance-to-swing transitions. Larger joint torques were observed at the swing-to-stance transition at the hip and knee for both gaits compared to the other transition. EMG results showed greater activation levels for five of the six muscles at the swing-to-stance transition. Results indicated that the subjects needed to prepare for the initiation of stance and the application of relatively large external forces and moments. Further, the transition from stance to swing did not seem to be as critical a point in the gait cycle since the movements and EMG were relatively low. This being the case, the stance-to-swing transition should be used as the beginning and ending of the gait cycle (toe-off initiating the cycle) and the more meaningful transition of swing-to-stance phases should occur in the middle of the analysis.

Interpretation and parameterization of dynamic trunk isoinertial movements using an ensemble-averaging technique

A procedure for averaging trunk dynamometric positions, velocities, and torques is described which enables identification and parameterization of kinematic movement patterns. This technique is accurate and presents the data in a very compact way, which is ideal for storage and analysis. It can be applied to many types of movement analyses. Application of the technique to the analysis of flexion-extension movement patterns in two low back pain patients is described.

Pattern analysis of electromyographic linear envelopes exhibited by subjects with uninjured and injured knees during free and fast speed walking

We wished to determine whether anterior cruciate ligament (ACL) injury caused changes in patterns of activity of individual or groups of muscles that control the knee joint. The electromyographic (EMG) patterns of six muscles in 26 individuals with uninjured knees and 20 individuals with ruptured ACL were studied during free and fast speed walking. A previously developed clustering technique was used for analysis. This technique involves making the Fourier transform of the average linear envelope (LE) of each muscle for each subject and using the amplitude and phase angles of the lower frequency harmonics as features to describe a pattern. These features are then grouped to subdivide the population of EMG patterns into different types. The results of analyses on single muscles showed that there exists a typical or "normal" pattern for each muscle which most uninjured and some injured subjects exhibit as well as several atypical patterns which mostly injured subjects exhibit, and that the atypical patterns are much more evident at fast walking speed. The characteristics of atypical patterns with respect to normal include time shifts in the peak of major phases of activity, the absence of a second phase of activity, and the existence of additional phases of activity. Synergy analysis showed that if one muscle has an atypical pattern, then several do; i.e., ACL injury induces major changes in the control strategy of the knee. The implications are that for rehabilitation programs one must focus on the training and strengthening of more than one muscle, and that for reconstructive procedures the changes in mechanics of the joint can possibly induce a significant change in its control strategy.

A quantitative and qualitative description of electromyographic linear envelopes for synergy analysis

The muscular synergy patterns of human locomotion can be described by the phasic activity of electromyographic linear envelopes (LE) and the interphasic spatio-temporal relations. To represent the phasic activity, the LE is modeled as the summation of Gaussian pulses of various lengths. The parameters of interest are the temporal features: time, duration, and amplitude of the phases of activity. A maximum likelihood approach to the parameter estimation for a mixture of normal distributions is adopted for extracting the temporal features. Based on the derived temporal features, a set of relational descriptors can be defined to describe the spatio-temporal relations between the multichannel phasic activities. The strength of this approach is not only that the phasic activity of LE can be quantitatively represented accurately, but also that the resulting synergy patterns can be easily interpreted by observers.

Clustering analysis and pattern discrimination of EMG linear envelopes

A technique has been developed for performing pattern analysis of EMG activities generated during locomotion. In this development it was found that the shapes of the EMG linear envelopes (LE) are mainly determined by their phase spectra; their magnitude spectra are much less important. Autoregressive (AR) parametric models and discrete Fourier transform (DFT) approaches were tested and compared. The latter was proved to be a better way to describe the EMG LE's. Feature extraction and clustering were performed by doing DFT of EMG LE's, extracting part of the phase and magnitude spectra (in less important degree) as features, and using the percent powers to weigh the corresponding harmonics. The approach was applied to the clustering analysis of EMG LE's of normal and anterior cruciate ligament (ACL) injured subjects during walking.

Temporal feature extraction and clustering analysis of electromyographic linear envelopes in gait studies

A technique for automatically clustering linear envelopes of the EMG during gait has been developed which uses a temporal feature representation and a maximum peak matching scheme. This new technique provides a viable way to define compact and meaningful EMG waveform features. The envelope matching is performed by dynamic programming, providing qualitatively the largest numbers of matched peaks and quantitatively a minimum distance measurement. The resulting averaged EMG profiles have low statistical variation and can serve as templates for EMG comparison and further classification.

EMG profiles of knee joint musculature during walking: changes induced by anterior cruciate ligament deficiency

A tear of the anterior cruciate ligament (ACL) disrupts the delicate balance of static stabilizers of the knee, leading to significant alterations in joint kinematics. Little is known about the dynamic compensatory responses of the patient to these kinematic alterations. This lack of quantitative information on the muscle synergy patterns has limited the surgeon's ability to evaluate various operative and rehabilitative techniques. Twelve subjects with documented ACL deficiency for at least 1 year and 15 normal participants were studied. Each subject was asked to walk at free and fast speeds on a 12 m walkway. The right and left foot contact patterns and the linear envelopes from the surface electromyogram (EMG) patterns of the gastrocnemius, medial and lateral hamstrings, rectus femoris, and vastus lateralis were measured. Significant differences were found in the muscle synergy patterns during walking. During the swing-to-stance transition, the ACL-deficient subjects showed significantly less activity in the quadriceps and gastrocnemius muscles and more activity in the biceps femoris than in the normal group. During early swing, the vastus lateralis is more active than normal, and during midstance and terminal stance, the hamstrings appear to be less active than normal subjects. These dynamic compensatory mechanisms suggest that use of the hamstring tendons in reconstructive procedures may alter important compensatory mechanisms about the knee joint. Application of dynamic EMG techniques to the study of reconstructive procedures should provide additional information that will assist the clinician in the rational choice of a surgical procedure.

Helical motion analysis of the knee--II. Kinematics of uninjured and injured knees during walking and pivoting

The knee kinematics of eight individuals with uninjured knees and of seven individuals with ruptured anterior cruciate ligaments have been investigated during walking and pivoting. The kinematics were measured using a six degree of freedom goniometer and quantitated using helical motion analysis. The helical motion variables reveal clearly that the knee is definitely neither a hinge nor a planar joint and its dynamic behavior changes over the stride. Ligamentous loss results in more adduction and external rotation during certain periods of the stride. Also, the range of translation of the tibia in the medial/lateral direction is reduced, and its average translation is more medial.

Helical motion analysis of the knee--I. Methodology for studying kinematics during locomotion

A technique for investigating the three-dimensional kinematics of knee motion during dynamic functional tasks has been developed. It involves the combined usage of a six degree of freedom goniometer and helical motion analysis. A detailed procedure for coordinate system alignment and calibration must be followed. Once established this entire procedure is routinely implementable. Ensemble averages from multiple walking strides reveal that this technique is sensitive enough to differentiate between the kinematics of an uninjured and injured knee.

Normative childhood EMG gait patterns

The population characteristics of the linear envelopes of the electromyograms measured from seven lower extremity muscles in children were studied during locomotion. The variability and changes in pattern with respect to walking speed and age were investigated using statistical properties and analysis of variance of the envelopes. All muscles studied showed changes in their patterns that were associated with different walking speeds. Some changes concerned the relative intensity of existing phases of activity whereas others concerned the existence of additional phases of activity. The most remarkable trend was that all envelopes tended to have more consistent patterns as speed increased. With regard to age, within the span of 4-11 years, only the two thigh muscles studied demonstrated appreciable differences.