Edward A. Clancy

Worcester Polytechnic Institute, Worcester, Massachusetts, United States

Are you Edward A. Clancy?

Claim your profile

Publications (42)61.68 Total impact

  • Kishor Koirala · Meera Dasog · Pu Liu · Edward A. Clancy
    [Show abstract] [Hide abstract]
    ABSTRACT: Processed (i.e., rectified, smoothed) electromyogram (EMG) activity from skeletal muscles precedes mechanical tension by 50-100 ms. This property can be exploited to anticipate muscle mechanical activity. Thus, we investigated the ability of surface EMG to estimate joint torque at future times, up to 750 ms. EMG recorded from the biceps and triceps muscles of 54 subjects during constant-posture, force-varying contractions was related to elbow torque. Higher-order FIR models, combined with advanced EMG processing (whitening; four EMG channels per muscle), provided a nearly identical minimum error of 5.48 ± 2.21% MVCF (flexion maximum voluntary contraction) over the time advance range of 0-60 ms. Error grew for larger time advances. The more common method of filtering EMG amplitude with a Butterworth filter (2nd-order, 1.5 Hz cutoff frequency) produced a statistically inferior (p<10-6) minimum torque error of 6.90 ± 2.39% MVCF, with an error nadir at a time advance of 60 ms. Error was progressively poorer at all other time advances. Lower-order FIR models mimicked the poorer performance of the Butterworth models. The more advanced models provide lower estimation error, require no selection of an electromechanical delay term and maintain their lowest error over a substantial range of advance times.
    No preview · Article · Jun 2014 · IEEE Transactions on Neural Systems and Rehabilitation Engineering
  • [Show abstract] [Hide abstract]
    ABSTRACT: The reliability of clinical and scientific information provided by algorithms that automatically decompose the electromyogram (EMG) depends on the algorithms' accuracies. We used experimental and simulated data to assess the agreement and accuracy of three publicly available decomposition algorithms-EMGlab [1] (single channel data only), Fuzzy Expert [2] and Montreal [3]. Data consisted of quadrifilar needle EMGs from the tibialis anterior of 12 subjects at 10%, 20% and 50% maximum voluntary contraction (MVC); single channel needle EMGs from the biceps brachii of 10 controls and 10 patients during contractions just above threshold; and matched simulated data. Performance was assessed via agreement between pairs of algorithms for experimental data and accuracy with respect to the known decomposition for simulated data. For the quadrifilar experimental data, median agreements between the Montreal and Fuzzy Expert algorithms at 10%, 20% and 50% MVC were 95%, 86% and 64%, respectively. For the single channel control and patient data, median agreements between the three algorithm pairs were statistically similar at ~97% and ~92%, respectively. Accuracy on the simulated data exceeded this performance. Agreement/accuracy was strongly related to the Decomposability Index [3]. When agreement was high between algorithm pairs applied to simulated data, so was accuracy.
    No preview · Article · May 2014 · IEEE Transactions on Neural Systems and Rehabilitation Engineering
  • Pu Liu · Francois Martel · Denis Rancourt · Edward A. Clancy · D. Richard Brown
    [Show abstract] [Hide abstract]
    ABSTRACT: Existing commercial hand prostheses can be controlled from the electrical activity (electromyogram or EMG) of remnant muscle tissue within the forearm, but are limited in function to one degree of freedom of proportional control. In a pilot study (N=3 subjects), we used least squares estimation to identify a model between forearm electrical activity recorded by high-resolution (64 channel) electrode arrays (applied over the flexor and, separately, extensor muscles of the forearm) to force in the four fingertips. Average errors ranged from 4.21 to 10.20 %MVCF (flexion maximum voluntary contraction), depending on the muscle contraction task performed, number of EMG electrodes in the model and the electrode montage selected. Results suggest that, at least for intact subjects, 2-4 degrees of freedom of proportional control are available from the EMG signals of the forearm.
    No preview · Conference Paper · May 2014
  • Pu Liu · Donald R. Brown · Edward A. Clancy · Francois Martel · Denis Rancourt
    [Show abstract] [Hide abstract]
    ABSTRACT: Electromyogram (EMG) activity from the extensor and flexor muscles of the forearm was sensed with high-density surface electrode arrays and related to the force produced at the four fingertips during constant-posture, slowly force-varying contractions from three healthy subjects. Various electrode montages (spatial filters) and number of electrodes used in the system identification were studied. Average errors were small, ranging from 4.21 to 8.10 %MVCF (flexion maximum voluntary contraction), with errors trending lower when more EMG channels were used and when a monopolar electrode montage was selected. Results are supportive that multiple degrees of freedom of proportional control information are available from the surface EMG of the forearm, at least in intact subjects. Applications for future study include the control of prosthetic upper limb devices in amputees.
    No preview · Conference Paper · Dec 2013
  • Meera Dasog · Kishor Koirala · Pu Liu · Edward A. Clancy
    [Show abstract] [Hide abstract]
    ABSTRACT: Whitening the surface electromyogram (EMG) improves EMG amplitude (EMGσ) and EMG-torque estimation. Laboratory studies utilizing contraction levels up to maximum voluntary contraction (MVC) show that whitening is useful over a frequency band extending to 1000-2000 Hz. However, EMG electrode systems with such wide bandwidth are uncommon, particularly in real-time applications; and these contraction levels are also not common. Thus, we studied the influence of the frequency band over which whitening was performed vs. the resulting performance. Low-level, torque-varying contractions (average torque level of 18.5% flexion MVC) of the elbow were contrasted with medium-level 50% MVC constant-torque contractions. For each, the maximum whitening bandwidth was varied between 30-2000 Hz. The low-level contractions (which incorporate the contraction range of most daily tasks) showed that performance utilizing frequencies out to 400-500 Hz was not statistically different (p<0.01) than results out to the full available frequency (2000 Hz). For the medium-level (50% MVC) contractions, frequencies out to 800-900 Hz were statistically equivalent to the full bandwidth. These results suggest that conventional electrodes with a typical passband of ~500 Hz are appropriate when whitening data from contraction levels typically experienced in many applications. Wider bandwidths may be advantageous for strenuous activities.
    No preview · Article · Oct 2013 · IEEE transactions on neural systems and rehabilitation engineering: a publication of the IEEE Engineering in Medicine and Biology Society
  • Pu Liu · Lukai Liu · Francois Martel · Denis Rancourt · Edward A Clancy
    [Show abstract] [Hide abstract]
    ABSTRACT: Electromyogram (EMG)-torque modeling is of value to many different application areas, including ergonomics, clinical biomechanics and prosthesis control. One important aspect of EMG-torque modeling is the ability to account for the joint angle influence. This manuscript describes an experimental study which relates the biceps/triceps surface EMG of 12 subjects to elbow torque at seven joint angles (spanning 45-135°) during constant-posture, quasi-constant-torque contractions. Advanced EMG amplitude (EMGσ) estimation processors (i.e., whitened, multiple-channel) were investigated and three non-linear EMGσ-torque models were evaluated. When EMG-torque models were formed separately for each of the seven distinct joint angles, a minimum "gold standard" error of 4.23±2.2% MVCF90 resulted (i.e., error relative to maximum voluntary contraction at 90° flexion). This model structure, however, did not directly facilitate interpolation across angles. The best model which did so (i.e., parameterized the angle dependence), achieved an error of 4.17±1.7% MVCF90. Results demonstrated that advanced EMGσ processors lead to improved joint torque estimation. We also contrasted models that did vs. did not account for antagonist muscle co-contraction. Models that accounted for co-contraction estimated individual flexion muscle torques that were ∼29% higher and individual extension muscle torques that were ∼68% higher.
    No preview · Article · Aug 2013 · Journal of electromyography and kinesiology: official journal of the International Society of Electrophysiological Kinesiology
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Stroke affects 750,000 people annually, and 80% of stroke survivors are left with weakened limbs and hands. Repetitive hand movement is often used as a rehabilitation technique in order to regain hand movement and strength. In order to facilitate this rehabilitation, a robotic glove was designed to aid in the movement and coordination of gripping exercises. This glove utilizes a cable system to open and close a patients hand. The cables are actuated by servomotors, mounted in a backpack weighing 13.2lbs including battery power sources. The glove can be controlled in terms of finger position and grip force through switch interface, software program, or surface myoelectric (sEMG) signal. The primary control modes of the system provide: active assistance, active resistance and a preprogrammed mode. This project developed a working prototype of the rehabilitative robotic glove which actuates the fingers over a full range of motion across one degree-of-freedom, and is capable of generating a maximum 15N grip force.
    Full-text · Article · Jun 2013 · IEEE International Conference on Rehabilitation Robotics : [proceedings]
  • Kishor Koirala · Meera Dasog · Pu Liu · Edward A. Clancy
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper investigates the ability of surface electromyogram (EMG) to estimate joint torque at future times, up to 1 s. EMG was recorded from the biceps and triceps muscles of 54 subjects during constant-posture, force-varying contractions and related to the torque produced about the elbow. EMG to joint torque was predicted up to 80 ms into the future without any changes in the minimum least square error of 5.48% of maximum voluntary contraction for the best estimation model investigated: whitened, multiple-channel EMG used with a nonlinear model. Error progressively increased for prediction times above 80 ms.
    No preview · Conference Paper · Apr 2013
  • Lukai Liu · Edward A. Clancy · Paolo Bonato
    [Show abstract] [Hide abstract]
    ABSTRACT: The discharge rate of motor unit action potential sequences has been related to fatigue and neuromuscular diseases, but typically simple methods are used to do so. We adapted more advanced methods used to calculate heart rate to fit in the context of surface electromyogram discharge rate calculation. Simulation results with a deterministic discharge rate modulation model suggest that parameter fine-tuning is necessary to accurately and robustly estimate discharge rate.
    No preview · Conference Paper · Apr 2013
  • [Show abstract] [Hide abstract]
    ABSTRACT: The reliability of automated electromyogram (EMG) decomposition algorithms is important in clinical and scientific studies. In this paper, we analyzed the performance of two multi-channel decomposition algorithms -- Montreal and Fuzzy Expert using both experimental and simulated data. Comparison data consisted of quadrifiler needle EMG from the tibialis anterior muscle of 12 subjects (young and elderly) at three contraction levels (10, 20 and 50% MVC), and matched simulation data. Performance was assessed via agreement between the two algorithms for experimental data and accuracy with respect to the known decomposition for simulated data. For the experimental data, median agreement between the Montreal and Fuzzy Expert algorithms at 10, 20 and 50% MVC was 95.7, 86.4 and 64.8%, respectively. For the simulation data, median accuracy was 99.8%, 100% and 95.9% for Montreal, and 100%, 98% and 93.5% for Fuzzy Expert at the different contraction levels.
    No preview · Conference Paper · Apr 2013
  • Lukai Liu · Pu Liu · Edward A. Clancy · Erik Scheme · Kevin B. Englehart
    [Show abstract] [Hide abstract]
    ABSTRACT: Time and frequency domain features of the surface electromyogram (EMG) signal acquired from multiple channels have frequently been investigated for use in controlling upper-limb prostheses. A common control method is EMG-based motion classification. We propose the use of EMG signal whitening as a preprocessing step in EMG-based motion classification. Whitening decorrelates the EMG signal, and has been shown to be advantageous in other EMG applications including EMG amplitude estimation and EMG-force processing. In a study of ten intact subjects and five amputees with up to 11 motion classes and ten electrode channels, we found that the coefficient of variation of time domain features (mean absolute value, average signal length and normalized zero crossing rate) was significantly reduced due to whitening. When using these features along with autoregressive power spectrum coefficients, whitening added approximately five percentage points to classification accuracy when small window lengths (<100 ms) were considered.
    No preview · Article · Mar 2013 · IEEE transactions on neural systems and rehabilitation engineering: a publication of the IEEE Engineering in Medicine and Biology Society
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The time delay between two surface electromyograms (EMGs) acquired along the conduction path is used to estimate mean action potential conduction velocity. Modeling the linear impulse response between "upstream" and "downstream" EMG signals permits an estimate of the distribution of velocities, providing more information. In this work, we analyzed EMG from bipolar electrodes placed on the tibialis anterior of 36 subjects, using an inter-electrode distance of 10 mm. Regularized least squares was used to fit the coefficients of a finite impulse response model. We trained the model on one recording, then tested on two others. The optimum correlation between the model-predicted and actual EMG averaged 0.70. We also compared estimation of the mean conduction delay from the peak time of the impulse response to the "gold standard" peak time of the cross-correlation between the upstream and downstream EMG signals. Optimal models differed from the gold standard by 0.02 ms, on average. Model performance was influenced by the regularization parameters. The impulse responses, however, incorrectly contained substantive power at very low time delays, causing delay distribution estimates to exhibit high probabilities at very short conduction delays. Unrealistic distribution estimates resulted. Larger inter-electrode spacing may be required to alleviate this limitation.
    Full-text · Article · Feb 2013 · Medical & Biological Engineering
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Mean electromyogram (EMG) conduction delay is often estimated as the average time delay between two surface EMG recordings arranged along the conduction path. It has previously been shown that the complete distribution of conduction delays can be estimated from the impulse response relating the "upstream" EMG recording to the "downstream" recording. In this work, we examined regularized least squares methods for estimating the impulse response, namely the pseudo-inverse with small singular values discarded and post hoc lowpass filtering. Performance was evaluated by training the model to one recording, then testing on others. Correlation between model-predicted EMG and measured EMG was assessed for 36 subjects, using EMG recordings with 5 mm inter-electrode spacing. The best correlation was 0.86, on average, for both regularization methods. We additionally compared the mean conduction delay computed from the "gold standard" cross-correlation method to the peak time of the impulse response. The best models differed by 0.01 ms, on average, for both regularization methods. Nonetheless, the impulse responses exhibited excessive energy near zero time, causing delay distribution estimates to exhibit high probabilities at unphysiological short time delays. Inter-electrode spacing larger than 5 mm may be required to alleviate this limitation.
    Full-text · Article · Aug 2012 · Conference proceedings: ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Conference
  • Lukai Liu · Pu Liu · Edward A Clancy · Erik Scheme · Kevin B Englehart
    [Show abstract] [Hide abstract]
    ABSTRACT: The electromyogram (EMG) signal has been used as the command input to myoelectric prostheses. A common control scheme is based on classifying the EMG signals from multiple electrodes into one of several distinct classes of user intent/function. In this work, we investigated the use of EMG whitening as a preprocessing step to EMG pattern recognition. Whitening is known to decorrelate the EMG signal, with improved performance shown in the related applications of EMG amplitude estimation and EMG-torque processing. We reanalyzed the EMG signals recorded from 10 electrodes placed circumferentially around the forearm of 10 intact subjects and 5 amputees. The coefficient of variation of two time-domain features-mean absolute value and signal length-was significantly reduced after whitening. Pre-whitened classification models using these features, along with autoregressive power spectrum coefficients, added approximately five percentage points to their classification accuracy. Improvement was best using smaller window durations (<100 ms).
    No preview · Article · Aug 2012 · Conference proceedings: ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Conference
  • Source
    John Meklenburg · Edward A Clancy · Radouil Tzekov
    [Show abstract] [Hide abstract]
    ABSTRACT: Oscillatory potentials (OPs) are typically isolated from the electroretinogram (ERG) via linear, time-invariant, bandpass filtering. The use of a highpass cutoff frequency that is too low results in a- and b-wave contamination of the OP signal, while a cutoff frequency that is too high removes significant OP signal energy. Two methods for automated highpass cutoff frequency estimation were developed and evaluated. An OP amplitude analysis method exploited a trend in variation of maximum OP amplitude with cutoff frequency. A second method fit a time-varying exponential model to the rising edge of the b-wave and selected a cutoff frequency based on minimizing the error between the residual ERG signal (the signal formed by subtracting the OP signal from the original ERG signal) and the exponential fit. The performance of each method was evaluated at 11 luminances (0.001-100 scot cd·s/m(2)) in ten wild-type adult mice by comparing the automated selections to expert-selected highpass cutoff frequencies. It was noted that cutoff frequency selection was not critical at the lower luminance levels, but strongly influenced the OP signal shape for higher luminances. At the highest luminance, errors between the OP amplitude and exponential model versus expert selection were -6.3 ± 13 and -8.2 ± 7.3 Hz, respectively. ANOVAs showed that estimations made by the OP amplitude analysis method were generally statistically indistinguishable from the expert identifications. Furthermore, both OP amplitude analysis and exponential fitting error analysis provided excellent fits to the manual selections for the four highest stimulus luminance values.
    Full-text · Article · Jul 2012 · Documenta Ophthalmologica
  • Source
    Edward A Clancy · Lukai Liu · Pu Liu · Daniel V Zandt Moyer
    [Show abstract] [Hide abstract]
    ABSTRACT: The surface electromyogram (EMG) from biceps and triceps muscles of 33 subjects was related to elbow torque, contrasting EMG amplitude (EMGσ) estimation processors, linear/nonlinear model structures, and system identification techniques. Torque estimation was improved by 1) advanced EMGσ processors (i.e., whitened, multiple-channel signals); 2) longer duration training sets (52 s versus 26 s); and 3) determination of model parameters via pseudoinverse and ridge regression methods. Dynamic, nonlinear parametric models that included second- or third-degree polynomial functions of EMGσ outperformed linear models and Hammerstein/Weiner models. A minimum error of 4.65 ± 3.6% maximum voluntary contraction (MVC) flexion was attained using a third-degree polynomial, 28th-order dynamic model, with model parameters determined using the pseudoinverse method with tolerance 5.6 × 10 (-3) on 52 s of four-channel whitened EMG data. Similar performance (4.67 ± 3.7% MVC flexion error) was realized using a second-degree, 18th-order ridge regression model with ridge parameter 50.1.
    Full-text · Article · Jan 2012 · IEEE transactions on bio-medical engineering
  • Source
    Lukai Liu · Pu Liu · Edward A. Clancy · Kevin B. Englehart
    [Show abstract] [Hide abstract]
    ABSTRACT: The surface electromyogram (EMG) signal collected from multiple channels has frequently been investigated for use in controlling upper-limb prostheses. One common control method is EMG-based motion classification. Time and frequency features derived from the EMG have been investigated. We propose the use of EMG signal whitening as a preprocessing step in EMG-based motion classification. Whitening decorrelates the EMG signal, and has been shown to be advantageous in other EMG applications. In a ten-subject study of up to 11 motion classes and ten electrode channels, we found that whitening improved classification accuracy by approximately 5% when small window length durations (
    Full-text · Article · Apr 2011
  • Source
    Pu Liu · Lukai Liu · Francois Martel · Denis Rancourt · Edward A. Clancy
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper describes an experimental study which relates the simultaneous biceps/triceps surface electromyogram (EMG) of 12 subjects to elbow torque at seven joint angles during constant-posture, quasi-constant-torque contractions. Advanced EMG amplitude (EMGσ) estimation processors were investigated, and an EMG-torque model considering agonist and antagonist co-contractions was evaluated at each joint angle. Preliminary results show that advanced (i.e., whitened, multiple- channel) EMGσ processors lead to improved joint torque estimation and that the EMGσ torque relationship may only change by a scaling factor as a function of joint angle. I. INTRODUCTION A significant literature has developed around the problem of relating the surface EMG to muscle tensions and joint torque. However, most investigators have not accounted for muscle co-contractions by assuming that an agonist muscle can be contracted while the antagonist muscle is inhibited (1), (2). Also, there are clear advances in EMGσ processing techniques over the last few years (3), yet little have been incorporated into EMG-torque estimation. The present study investigated the EMG-torque problem by modeling agonist- antagonist co-contractions over a wide range of joint torques at seven different angles, and also applied advanced EMGσ processing techniques (whitening, multiple-channel combination).
    Full-text · Article · Apr 2011
  • Source
    Edward A. Clancy · Mark V. Bertolina · Roberto Merletti · Dario Farina
    [Show abstract] [Hide abstract]
    ABSTRACT: The amplitude and mean power spectral frequency (MNF) of the electromyogram (EMG) of flexor digitorum superficialis and extensor carpi radialis muscles were monitored during cyclic, force-varying, constant-posture, submaximal, grip-force contractions until endurance. These contractions are reminiscent of work tasks associated with the risk of repetitive stress injuries. Based on recommendations from a prior cross-comparison study of these data, the cyclic grip-force contractions were temporally aligned cycle-by-cycle to the achieved grip force profile, then EMG amplitude was computed using signal whitening (500 ms window) and MNF was computed using the short-time Fourier transform (500 ms window). In addition, brief (8 s) constant-force (static) contractions were interspersed within the cyclic contractions every 5 min. MNF was tracked during these periods. All subjects reported a marked increase in pain/discomfort/fatigue during the contraction trials, until self-selecting to discontinue contractions after 30–90 min. Discomfort returned to near-baseline levels during the ensuing 45 min recovery (rest) period. No statistical trend was found in either EMG amplitude or MNF during the cyclic contractions or the recovery period. Initial MNF and MNF slope were monitored during the 8 s interspersed static contractions. These parameters also did not follow any consistent trend. These results indicate limitations in the use of these EMG descriptors for assessment of fatigue during long-duration, force-varying contractions.
    Full-text · Article · Oct 2008 · Journal of Electromyography and Kinesiology
  • Source
    Edward A Clancy · Hongfang Xia · Anita Christie · Gary Kamen
    [Show abstract] [Hide abstract]
    ABSTRACT: Multiple-channels of electromyogram activity are frequently transduced via electrodes, then combined electronically to form one electrophysiologic recording, e.g. bipolar, linear double difference and Laplacian montages. For high quality recordings, precise gain and frequency response matching of the individual electrode potentials is achieved in hardware (e.g., an instrumentation amplifier for bipolar recordings). This technique works well when the number of derived signals is small and the montages are pre-determined. However, for array electrodes employing a variety of montages, hardware channel matching can be expensive and tedious, and limits the number of derived signals monitored. This report describes a method for channel matching based on the concept of equalization filters. Monopolar potentials are recorded from each site without precise hardware matching. During a calibration phase, a time-varying linear chirp voltage is applied simultaneously to each site and recorded. Based on the calibration recording, each monopolar channel is digitally filtered to "correct" for (equalize) differences in the individual channels, and then any derived montages subsequently created. In a hardware demonstration system, the common mode rejection ratio (at 60 Hz) of bipolar montages improved from 35.2+/-5.0 dB (prior to channel equalization) to 69.0+/-5.0 dB (after equalization).
    Preview · Article · Oct 2007 · Journal of Neuroscience Methods

Publication Stats

903 Citations
61.68 Total Impact Points

Institutions

  • 2005-2014
    • Worcester Polytechnic Institute
      • • Department of Electrical and Computer Engineering
      • • Department of Biomedical Engineering
      Worcester, Massachusetts, United States
  • 2004
    • Harvard University
      Cambridge, Massachusetts, United States
  • 1997-1999
    • Liberty Mutual Research Institute for Safety
      Boston, Massachusetts, United States
  • 1994-1999
    • Massachusetts Institute of Technology
      • Department of Electrical Engineering and Computer Science
      Cambridge, Massachusetts, United States