Table 9 MMG and muscle torque

[79]

Torque: strain gauge (model LURA-100NSA1, Kyowa Electronic Instruments, Japan); MMG: microphone sensor (QTEC, Japan; diameter = 8 mm, mass = 5 g)

BB

9 healthy male subjects, age (mean ± SE) 24.5 ± 1.1 years, height 171.3 ± 2.8 cm, weight 68.1 ± 2.6 kg

A 10 Hz, 5-s train at 30 Hz was applied at each angle to familiarize the subject with the pain induced by the tetanic stimulation. The elicited torque at 75, 90, 120, 135 and 150° was measured and was analyzed based on recorded MMG signals. A 100-µs pulse duration was provided to ensure an increase in the muscle contraction with torque

The MMG amplitude was correlated with the CT and the ½T_r and reflected the changes in torque relaxation observed with increases in the muscle length

Remark: the correlation of the MMG amplitude with the half-relaxation and contraction times indicates that MMG is a valid tool for monitoring changes in the contraction features of skeletal muscle

Study 2: analysis of the effect of age and muscle mechanics

[87]

MMG: 9-mm square with a thickness

of 4.5 mm and a mass of 0.75 g (MP110-10–101; MEDiSENS, Tokyo, Japan; sensitivity = 500 mV/g); EMG: (1.5—1 cm; Kendall-LTP, Chicopee, MA, USA)

TA

10 young male subjects, age 27.1 ± 3.8 years (range 21–33 years), height 174.2 ± 7.7 cm, weight 78.7 ± 7.8 kg; 10 old male subjects, age 79.0 ± 2.5 years (range 75–83 years), height 171.4 ± 5.7 cm, weight 79.2 ± 10.2 kg

The M-wave was detected after delivering a 400-V pulse with a width of 50 µs and a current intensity of 70–150 mA. The pT was assessed after a supramaximal twitch and a 30–60-s rest

The electrical and mechanical features of muscles that are either electrically or voluntary induced are affected in different manners by ageing

Remark: the MMG and EMG patterns presented similar shapes in young subjects and were altered in the old population

Study 3: analysis of muscle features during long and short stimulation protocols

[80]

MMG: laser-distance sensor (M5 L/20, MEL Mikroelektronik, Germany, range =  ± 10 mm, sensitivity = 1 V/mm, linearity = 0.6%, resolution < 6 µm, bandwidth = 0–10 kHz); EMG: two bar electrodes (1 cm × 1 mm × 1 mm); force: load cell (Interface, model SM-100 N, operating range = 0–100 N)

TA

14 healthy male subjects, age 20–35 years

12.5-s short duration pulses (0.4, 6.0, 1.0, 4.5, 1.8, 3.0, and 2.5 Hz) and 6-s pulses (0.4, 0.6, 0.8, 1.0, 1.2, 1.4, 1.6, 1.8, 2.0, 2.5, 3.0, 4.0, 5.0, and 6.0 Hz) with a 300-s rest between pulses to avoid fatigue

A similar transfer function from the torque and laser-detected MMG signal resulted in a decline in the sinusoidal amplitude and a phase shift in the torque and MMG

Remark: the torque and laser MMG recordings yielded similar transfer functions, which validates the use of MMG for screening the mechanical properties of muscle tendons units

Future work: elicitation of the whole motor nerve can be used to investigate the shortening of non-active muscle fibers

Study 4: model of human muscle using a triangular frequency and an amplitude train

[81]

MMG: laser distance sensor (M5 L/20, MEL Mikroelektronik, Germany, range =  ± 10 mm, bandwidth = 0–10 kHz, linearity = 0.6%); EMG: pair of silver bar electrodes; force: load cell (model SM-100 N, Interface Inc, Scottsdale, AZ, USA; operation range = 0–100 N)

TA

10 healthy male subjects, age 23–35 years

Increase from 2 to 35 Hz over 7.5 s or decrease from 35 to 2 Hz over 7.5 s; the amplitude was increased from V_min to V_max and decreased from V_max to V_min. The mean torque, MMG and EMG for each 5% in the ∆ frequency range of 2 Hz (0%) to 35 (100%) Hz were used to determine the stimulation frequency and additional muscle torque

Under DGR, the amplitude and frequency triangle are in line with the additional torque and muscle displacement

Remark: bypassing the CNS, the additional torque and MMG are induced by the intrinsic muscle property

Future work: the number of subjects needs to be increased to validate the EMG behaviors in DGR and UGR induced by a frequency triangle

Study 5: analysis of contractile parameters from the TMG and the torque twitch response in human VL muscle

[82]

TMG: displacement sensor (G40 digital-optical comparator, TMG-BM Ltd., Slovenia); force: transducer (TSD121C, BIOPAC Systems Inc., USA)

VL

19 healthy male subjects, age 46.1 ± 17.8, body mass 78.4 ± 12.7 kg, height 1.74 ± 0.05 m, BMI 26.0 ± 4.2 kg/m²

1-ms pulse current separated by 10 s from the motor threshold to the maximum stimulation current at incremental steps of 5 mA, with a maximum stimulation amplitude of 10–100 mA

Based on TMG, the force inter-method correlation was significant with Ts and Tr but not significant with other parameters

Remark: the TMG-measured contractile parameters are shorter during the contraction phase due to their dependence on intrinsic muscle properties

Study 6: prediction of NMES-evoked knee torque using SVR

[83]

ES: 9 × 15-cm² self-adhesive electrodes (Hasomed GmbH, D 39114, Magdeburg, Germany); MMG: accelerometer (Sonostics BPS-II VMG transducer, sensitivity = 30 V/g)

RF

8 healthy male subjects, age 23.4 ± 1.3 years, body mass 70.4 ± 5.8 kg, height 1.72 ± 0.05 m

Square-wave 30-Hz pulses with a 400-µs duration were applied; the current amplitude was increased from 20 to 80 mA over a duration of 48 h, and a recovery time of 10 min was established after each trial

A high prediction accuracy with a coefficient of determination of \({R}^{2}=94\%-89\%\) and a low root mean square error of 9.48–12.95 was obtained

Remark: the study was limited to healthy volunteers

Future work: the model should be examined with disabled subjects, and normal and fatigued knee extensors during standing or itinerant tasks should be classified

Study 7: analysis of twitch torque and reliability of recruitment curves

[84]

ES: self-adhering stimulating electrodes (1 × 1 cm); MMG: accelerometer (EGAS-FS-10-/VO5, Measurement Specialties Inc., Hampton, VA, USA)

Soleus

16 subjects, age 23.5 ± 1.9 years, body mass 71.1 ± 10.6 kg, height 173.4 ± 8.0 cm

A pulse (1 ms, 100–400 V, and 2–100 mA) was administered with increments of 0.5–2 mA to detect the maximum H-reflex and decreases in the H-reflex. Thereafter, an increment of 2–5 mA was applied to localize the plateau in the M-wave amplitude; a rest period of 3 to 5 s was applied between stimuli, and four to seven stimuli were included in each individual pulse

A strong test–retest reliability with p > 0.05 was found over days. A low coefficient of variation for MMG_max with the maximal stimulation intensity and torque was found

Remark: a new parameter, M + H, exhibited a strong correlation with the twitch torque and MMG

Future works:

1. The gross lateral movement of muscles under voluntary contractions should be investigated

2. Comparisons of M-wave and H-reflex should be performed, and features related to the soleus and gastrocnemius muscles after treatment should be investigated

Study 8: system identification of muscle with parallel fibers

[85]

ES: Ag–AgCl surface electrodes (F-150S, Nihon Kohden, Tokyo); DMMG: capacitor microphone (MX-E4758, Primo, Tokyo); AMMG: MP-110–10-101 (MediSens, Saitama); Force: FlexForce A201-1 (Nitta, Osaka)

Abductor pollicis brevis

6 healthy male subjects, age 21–25 years

The median nerve was electrically stimulated with a 500-µs rectangular pulse at an interval of 1 s and a bandwidth of -3Db; the pulse was repeated 20 times

Muscles with parallel fibers can be modeled by a 6^th-order model using system identification of DMMG and AMMG

Remark: the abductor pollicis brevis presented a higher frequency than the tibialis anterior, which might be due to the anatomical structure

Future work: the effect of the anatomical properties of the tibialis anterior and abductor pollicis brevis on the natural frequency should be verified

Study 9: prediction of the muscle torque using ANN

[86]

Torque: dynamometer (Biodex Medical system Shirley, NY, USA); ES: 9 × 15-mm² self-adhesive electrodes; MMG: transducer (Sonostics BPS-II VMG, sensitivity = 30 V/g)

Right and left GM

8 SCI volunteers with ISNCSCI classes A and B, 3 for training and 5 for testing

30 Hz, 200-µs duration and 100-mA amplitude; the MMGrms, MMG ZC and torque were recorded with a knee angle of 30°

The ANN-based prediction using MMG and dynamometer recordings were close with a p-value of 0.33

Remark: MMG-RMS-ZC showed higher prediction than RMS alone

Future work: other MMG features and other ANN models should be investigated