Yasushi Itoh

Mechanical behaviour of condenser microphone in mechanomyography

Condenser microphones (MIC) have been widely used in mechanomyography, together with accelerometers and piezoelectric contact sensors. The aim of the present investigation was to clarify the mechanical variable (acceleration, velocity or displacement) indicated by the signal from a MIC transducer using a mechanical sinusoidal vibration system. In addition, the mechanomyogram (MMG) was recorded simultaneously with a MIC transducer and accelerometer (ACC) during voluntary contractions to confirm the mechanical variable reflected by the actual MMG and to examine the influence of motion artifact on the MMG. To measure the displacement-frequency response, mechanical sinusoidal vibrations of 3 to 300 Hz were applied to the MIC transducer with different sizes of air chambers (5, 10, 15 and 20 mm in diameter and 15, 20 or 25 mm long). The MIC transducer showed a linear relationship between the output amplitude and the vibration displacement, however, its frequency response declined with decreasing diameter and decreasing length of the air chamber. In fact, the cut-off frequency (−3 dB) of the MIC transducer with the 5-mm-diameter chamber was 10, 8 and 4 Hz for the length 15, 20 and 25 mm, respectively. The air chamber with at least a diameter of 10 mm and a length of 15 mm is recommended for the MIC transducer. The sensitivity of this MIC transducer arrangement was 92 mVμm−1 when excited at 100 Hz. During voluntary contraction, the amplitude spectral density function of the MMG from the MIC transducer resembled that of the double integral of the ACC transducer signal. The angle of the MIC transducer was delayed by 180° in relation to the ACC transducer signal. The sensitivity of the MIC transducer was reduced to one-third because of the peculiar volume change of air chamber when the MMG was detected on the surface of the skin. In addition, the MIC transducer was contaminated by a smaller motion artifact than that from the ACC transducer. The maximal peak amplitude of the MIC and ACC transducer signal with the motion artifact was 7.7 and 12.3 times as much as the RMS amplitude of each signal without the motion artifact, respectively. These findings suggest that the MIC transducer acts as a displacement meter in the MMG. The MIC transducer seems to be a possible candidate for recording the MMG during dynamic muscle contractions as well as during sustained contractions.

Technical aspects of mechnomyography recording with piezoelectric contact sensor.

The piezoelectric contact sensor has been widely utilised in mechanomyography (MMG). The authors aim to clarify the mechanical variables (i.e. acceleration, velocity or displacement) reflected by the MMG signal detected with a piezoelectric contact sensor (PEC), and compare the results with those obtained simultaneously by an accelerometer (ACC). To measure the acceleration-frequency response, a mechanical sinusoidal excitation of 5 to 300 Hz at a constant magnitude of 0.01 G was applied to the two transducers. The acceleration-frequency response of the ACC transducer was confirmed to be almost flat. The PEC without any restriction of the transducer housing (including the combined seismic mass) demonstrated a similar response to the ACC transducer. The PEC transducer output with restricted housing decreased with increasing sinusoidal frequency and an attenuation slope of −40 dB/decade and phase angle of −180 degrees. The voluntary MMG signal during isometric knee extension was recorded simultaneously with the two transducers. The amplitude spectral density distribution of the MMG from the PEC transducer was narrow and the mean frequency was approximately one-half that obtained from the ACC tranducer. The amplitude spectral density distribution with the PEC transducer resembled that of the double integral over time of the ACC transducer signal. The phase angle of the PEC transducer signal was different from that of the ACC transducer signal by approximately −180°. These results suggest that the PEC transducer acts as a displacement meter of muscle vibration. In addition, differences in the MMG frequency components relating to the transducer type must be taken into consideration when investigating the mechanical activity of muscle.

Relationship between mechanomyogram and force during voluntary contractions reinvestigated using spectral decomposition

Abstract A mechanomyogram (MMG) is considered to represent the pressure waves resulting from the lateral expansion of contracting muscle fibers. However, the actual MMG recording appears not only to reflect lateral changes of active fibers, but also to include the effect of their longitudinal shortening, because the fiber orientation, particularly in pennate muscles, is not parallel with the MMG transducer attached at the skin surface. In the present investigation, a spectral decomposition method was developed to eliminate the interference due to fiber longitudinal movement from the MMG recording. The MMG was recorded over the belly of the rectus femoris muscle, which is a pennate muscle. Vibration over the tibial tuberosity (VTT) was used as a measure of the integrated longitudinal movement of the muscle fibers. The lateral and longitudinal components included in the MMG were separated by a spectral decomposition method that is based on the coherence function of the MMG and VTT. The MMG/force relationship was compared between the original and decomposed MMG. One-third of the 12 subjects demonstrated a curvilinear relationship between the original MMG and force throughout the range of force. In the other two-thirds, the MMG saturated or reduced beyond 70% of the maximum voluntary contraction (MVC). After decomposition, the MMG increased progressively with force up to 70% MVC, beyond which it decreased in all subjects. The spectral decomposition method described here is considered to be a useful tool with which to examine in more detail the MMG/force relationship of different pennate muscles.

Frequency response model of skeletal muscle and its association with contractile properties of skeletal muscle

The aims of the present study were to develop a mathematical model of the skeletal muscle based on the frequency transfer function, referred to as frequency response model, and to presume the relationship between the model elements and skeletal muscle contractile properties. Twitch force in elbow flexion was elicited by applying a single electrical stimulation to the motor point of biceps brachii muscles, and then analyzed visually by the Bode gain and phase diagram of the force signal. The frequency response model was represented by a frequency transfer function consisting of five basic control elements (proportional element, dead time element, and three first-order lag elements). The model element constants were estimated by best-fitting to the Bode gain and phase diagram of the twitch force signal. The proportional constant and the dead time in the frequency response model correlated significantly with the peak torque and the latency in the actual twitch force, respectively. In addition, the time constants of the three first-order lag elements in the model correlated strongly with the contraction time and the half relaxation time in the actual twitch force. The results suggested a possibility that the individual elements in the frequency response model would reflect the biochemical and biomechanical properties in the excitation-contraction coupling process of skeletal muscle.

Spectrum analysis of the mechanomyogram: elimination of the longitudinal shortening component of muscle fibers

The mechanomyogram (MMG) is a record of very small vibrations appearing on the body surface above an active muscle. It is attributed to an increase in the radius of the muscle fiber in the lateral direction. In pinnate muscle, however, the muscle fibers run in the radial direction, and it is expected that vibration due to the shortening of the muscle fiber in the longitudinal direction will be included in the MMG. The purpose of this study is to verify that a longitudinal shortening component exists in the MMG derived from the tetraceps femoris, a pinnate muscle, and then to propose a method of separating the components. In this study, the existence of the longitudinal shortening component in the MMG is verified by cross-spectrum analysis between the vibration signal FT derived from the rough surface of the tibia and the MMG. Then, based on the coherence between the MMG and FT, a method of separating the lateral expansion component and the longitudinal shortening component from the MMG is devised. The method is applied to measured MMGs and its validity is verified; the separated longitudinal shortening component exhibits the same tendency as the FT in both power and average frequency. It is also shown that the MMG after elimination of the longitudinal shortening component more clearly represents the activity of the muscle fiber. Thus, the effectiveness of the proposed method is demonstrated.