“Design, Characterization and Improvement of

Sensors for indirect detection of muscle force by

means of estimates of transversal stiffness”

By

Nikhilesh KumarB.Tech(Hons.)

Indian Institute of Technology, Kharagpur

Under the guidance ofProf. Roberto Merletti

LiSiN, Politecnico di Torino

1

Acknowledgements

With great pleasure and deep sense of gratitude, I express my indebtedness to Prof. Roberto Merletti for his invaluable guidance and constant encouragement at each and every step of my project work. I also would like to acknowledge the encouragement that I received from all my colleagues and senior members of LiSiN through out my journey on this bold and ambitious project.

I express my gratitude to the LiSiN and COREP for providing a platform that allowed me to experiment without hesitation.

I would also like to express my gratitude to Mr. Luca Greco my Lab mate and my friend for working with me as a team to develop methods to convert the ideas for this project into a possibility.

Finally, I wish to acknowledge the support and the teachings of my parents who have taught me to work hard with integrity and faith to achieve ambitious goals in pursuit of constant improvement. I dedicate this thesis to them

Nikhilesh KumarLiSiN

2

CONTENTS

1. Abstract…………………………………………………………………. 42. Literature Survey

Soft tissue Mechanic Stiffness meter…………………………… 5-7 Ultrasound Palpation meter……………………………………… 8-9 Scanning laser Meter…………………………………………….. 10 Electrical Impedance and admittance as a measure of

Muscle force……………………………………………………….. 11-123. Component Analysis

Analysis of commercially available shaker/vibrators………….. 13-16 Analysis of available force sensors suitable for our

Stiffness meter………………………………………………………17-204. Relevant Circuits

LM1875 and Power amplifier circuit……………………………… 21 Driving circuit for the Flexiforce…………………………………… 21 Low pass filter………………………………………………………. 22

5. Methodology adopted…………………………………………………… 236. Calibration of the instruments

Calibration of the flexiforce………………………………………… 24 Calibration of the shaker…………………………………………... 25

7. Results and Data………………………………………………………… 26-308. References……………………………………………………………….. 31

3

ABSTRACT

Many people have tried to measure the muscular force as well the stiffness of muscle, but none have tried to measure the contractile muscular force by measuring the transversal mechanical impedance of the muscles.

When we try to relate the EMG and the contraction force by the muscle, we are unable to succeed as estimation of the force contributed by a particular muscle becomes very difficult. But by measuring the stiffness of the muscle and knowing the global force we can estimate the contraction force of a single muscle.

Here we intend to estimate the contraction force of the muscle by measuring the transversal stiffness of the muscle. To measure the transversal impedance of the muscle we can calculate the stiffness of the muscle which can be estimated through the simple formula:

Stiffness of the material = (force applied normal to the surface)/ (depth penetrated to the surface)

So we impose a constant displacement shaker on the surface of the muscle which penetrates the muscle to a depth and the muscle produces a feedback force proportional to the stiffness of the muscle.We know the amplitude of the displacement or we can say the depth to which the shaker penetrates the surface of the muscle and hence by measuring the feedback force we will be able to measure the stiffness of the muscle.

After a lot of survey of the commercially available components we choose the Electro-magnetic Shaker to impose a constant displacement on the surface of the muscle and Flexiforce (force transducer) to measure the feedback force.

After proper calibration of the shaker and the flexiforce we took many readings of Arm muscles. The conclusion for the collected data is that with more number of sample we can develop a data-base and with its statistics we can estimate the contribution of a single muscle while knowing the global force.

4

INTRODUCTION

Analysis of literature on mechanical measurement of transversal muscle stiffness as a measure of muscle

force

Used search engine: PubmedUsed keywords: Muscular stiffness

Muscular hardnessElasticity measurement

Soft tissue mechanic stiffness meter:

Ishan Barman, Sujoy K. Guha“Analysis of a new combined stretch and pressure sensor for internal nodule palpation”Sensor and Actuators A125 (2006) 210-216

Aim is to develop a palpation sensor based on stretch and pressure sensor.Metallic strain gauges are used for implementing the pressure sensors and a stretch element for the stretch sensing part.

Sensor is a 2-Dimensional array of the tactels.

Same principle can be adopted for sensing the force/pressure in finding the muscular impedance using vibrator.

DFSA sensor configuration with sensor pressed upon a deformable slab. Enlarged view of a tactile element (tactel) is shown on the top left corner. A strain gauge is attached to a spring leaf, which is joined to a piston–cylinder guide. The combination is fixed on the tactel base. Enlarged view of the stretch element conforming to the contour of the slab material, bulging upwards in between the tactel tips, is shown on the top right corner.

Jari P A Arokosky, Jarkko Surakka, Tuula Ojala

5

“Feasibility of the use of a novel soft tissue stiffness meter”Phyiol.means 26 (2005) 215-228

The aim of the present study was to evaluate the feasibility of using a novel hand-held computerized soft tissue stiffness meter (STSM).

LEFT. Schematic presentation of the system devised for the tissue stiffness measurement and the measurement

principle during forearm extensor loading. RIGHT. Schematic presentation of the soft tissue stiffness meter

Firstly, the STSM was used to test elastomer samples with known mechanical properties. In the in vivo assessment, 12 healthy, non-disabled adults (age range, 24–57 years) and 16 subjects with chronic myofascial neck pain syndrome (age range, 27–55 years) were studied.

Makoto Morisada, Kaoru Okada, kenji Kawakita“Quantitative analysis of muscle hardness in tetanic contractions induced by electrical stimulation in rats”Eur J Appl Physiol (2006) 97: 681-686

The purpose of this study was to investigate the in vivo relation between muscle hardness during an electrically induced contracting state and neuromuscular functions (M-wave and developed tension).The hardness measured during nerve stimulation was correlated with the amplitude of the M-wave and the developed tension .

a Aspect of the device. b Hardness is estimated automatically based on the distance moved by the main spindle

Annaliese D. Heiner, M. James Rudert, Todd O. Mc Kinley

6

“In vivo measurement of translation stiffness of rabbit knees”Journal of Biomechanics(Article in press, Accepted 26 October 2006)

This paper describes the design, evaluation, and preliminary results of a specialized testing device and surgical protocol to determine translational stiffness of a rabbit knee.A linear stepper motor draws the tibia upward then returns to the home position, and a load cell measures the resisting force; force–displacement knee stiffness is then calculated.

Rabbit knee testing device—(a) overview, (b) rabbit in place

7

Ultrasound palpation meter:

Philippe Pourcelot, Marielle Defontaine, Berangere Ravary“A non invasive method of tendon force measurement”Journal of Biomechanics 38 (2005) 2124-2129

This study shows that the forces acting on tendons can be measured, in a non-invasive way, from the analysis of the propagation of an acoustic wave. Using the equine superficial digital flexor tendon as a model, it is demonstrated that the velocity of an ultrasonic wave propagating along the main axis of a tendon increases with the force applied to this tendon.

This figure shows tendons and muscles that resist overextension of the metacarpophalangeal joint during limb compression. The probe sed to record the US velocity in the SDF tendon is made of one emitter and five receivers.

Y.P.Zheng, Z.M. Li, A.P.C. Choi“Ultrasound palpation sensor fo tissue thickness and elasticity measurement – Assessment of transverse carpal ligament”Ultrasonic 44 (2006) e313-e317

Tissue ultrasound palpation sensor (TUPS) provides a feasible solution that makes the palpation of soft tissues not subjective feeling. It is comprised of an ultrasound transducer together with a load cell to form the finger-sized probe. The probe is used to push against the soft tissue surface to measure the thickness and elasticity of the soft tissues.

Schematic of the TUPS system including the finger size probe together with other electronic parts as a control system of the load cell and ultrasound transducer.

8

Youjiro Tamura, Ichiro Hatta, Takashi matsuda“Changes in muscle stiffness during contraction recorded using ultrasonic waves”Reprinted from Nature, Vol. 299, No. 5884, pp 631-633, 1982

A technique has been developed with which the stiffness changes in frog skeletal muscle can be continuously recorded by measuring the propagation velocity of ultrasonic waves (3-7 MHz) with negligibly small perturbations to the contractile system.The longitudinal muscle stiffness increased during isometric contraction at a rate faster than the force development. On the other hand, the transverse muscle stiffness decreased during isometric contraction at a rate faster than the force development.

Faletti Carlo, Pozza Simona, De Marchi Armanda“Elastosonography of muscle tendons pathology: preliminary experience”Biomedical engineering in exercise and sports internal congress 2006March 23-25, Torino-Italy

I.Hatta, Y. Tamura, T. Matsuda“Muscle stiffness changes during isometric contraction in frog skeletal muscle as studied by the use of ultrasonic waves”Edited by Gerald H. Pollack, and Haruo SugiContractile mechanisms in muscle, plenum publishing 1984

I.Hatta, Y. Tamura, T. Matsuda“Stiffness changes in frog skeletal muscle during contraction recorded using ultrasonic waves”Journal of Physiology (1988) 403 pp 193-209

9

Scanning laser meter:

Claudio Orizio, Massimiliano Gobbo, Arsenio Veicsteinas“Transients of the force and surface mechanomyogram during cat gastrocnemius tetanic stimulation”Eur J Appl Physiol (2003) 88: 601-606

The aim of the study was to investigate the time relationship between force and muscle surface displacement, detected as the surface mechanomyogram (MMG) by a laser distance sensor. Force was detected by a transducer connected at the distal tendon while MMG was measured after pointing the laser beam at the muscle belly.

Schematic drawing of the experimental set-up. (F Force transducer connected to the tendon, LG lateral gastrocnemius, MG medial gastrocnemius.) The four black circles indicate the detection sites of the muscle displacement in the four investigated

cats

Teizo Tsuchiya, Hiroyuki Iwamoto, Yojiro Tamura“Measurement of transverse stiffness during contraction in frog skeletal muscle using scanning laser acoustic microscope”Japanese journal of Physiology 43, 649-657, 1993

The scanning laser acoustic microscope (SLAM) was utilized to measure the change in the propagation velocity in the transverse direction during contraction in living skeletal muscles of the frog. In all the measurements (n = 15), the velocities during contraction were clearly slower (-7.6 m/s) than at rest and this means that the transverse stiffness decreased during contraction (-2.4 x 10(7) N/m2).

Teizo Tsuchiya, Hiroyuki Iwamoto, Yojiro Tamura“Measurement of transverse stiffness change during contraction in frog skeletal muscle by scanning laser acoustic microscope”Edited by H.Sugi and G. H Pollack plenum press, New York, 1993Mechanism of Myofilament sliding in muscle contractio, page 715-724

10

Electrical impedance and Admittance as a measure of muscle force

Used search engine: PubmedUsed keywords: Electrical impedance

EIM

Carl A Shiffman, Ronald Aaron and Seward B Rutkove“Electrical impedance of muscle during isometric contraction”Physiol. Meas. 24 (2003) 213–234

Non-invasive measurements of the 50 kHz impedance of the anterior forearm show that the resistance and reactance increase under voluntary isometric contraction of the finger flexor muscles. The relationship between impedance and force is nonlinear. Supporting data are presented for six healthy men and women, withages ranging from 19 to 70 years.

Sketch of the physical arrangement and block diagram of the data acquisition system.

Seward B. Rutkove, Kyungmouk S. Lee, Carl A. Shiffman, Ronald Aaron“Test–retest reproducibility of 50 kHz linear-electrical impedance myography”Clinical Neurophysiology 117 (2006) 1244–1248

Fifty kilohertz linear-EIM was performed on the biceps, quadriceps, and tibialis anterior of 30 nor mal subjects ranging in age from 21 to 90 years, and the major outcome variable, the spatially averaged phase (qavg), measured. The measurements were repeated within 250 days and comparisons between the two data sets made. Fifty kilohertz linear-EIM demonstrates excellent test–retest reproducibility.

Seward B. Rutkove*, Ramon A. Partida, Gregory J. Esper, Ronald Aaron, Carl A. Shiffman

11

“Electrode position and size in electrical impedance myography”Clinical Neurophysiology 116 (2005) 290–299

High frequency alternating current is injected into the body via two surface electrodes, and the resulting voltage pattern over an elected muscle is measured using a second, larger set of electrodes. The precise location and size of the electrodes can be critical to the data obtained, and in this study the effects of variation in these factors were evaluated.Linear-EIM results depend systematically on current and voltage electrode positions, but with reasonable care variation can be minimized.

Voltage electrode array configuration in linear-EIM for the study of biceps.

Rui Nie, N. Abimbola Sunmonu, Anne B. Chin, Kyungmouk S. Lee, Seward B. Rutkove

“Electrical impedance myography: Transitioning from human to animal studies”Clinical Neurophysiology 117 (2006) 1844–1849

Ronald Aaron, Gregory J Esper, Carl A Shiffman1, Kaca Bradonjic, Kyungmouk S Lee and Seward B Rutkove

“Effects of age on muscle as measured by electrical impedance myography”Physiol. Meas. 27 (2006) 953–959

Andrew W Tarulli, Anne B Chin, Ramon A Partida and Seward B Rutkove“Electrical impedance in bovine skeletal muscle as a model for the study of neuromuscular disease”Physiol. Meas. 27 (2006) 1269–1279

12

Analysis of vibrators/shakers commercially available and suitable for the stiffness measurements

(Esposito research)

Linear solenoid 6EPM 196659-02312 V DC

€ 230,00 delivery in 10 days

Giorgio ViganòVibrazioni indutriali S.N.C.Via Sostegno 7510146 Torinoe-mail: vibind@tin.ittel: 011-4532027

www.vibrazioni-industriali.it

Electric shaver engine

13

4 mm

18 mm

15 mm

26 mm 18 mm

Stepper motor1.5 – 4.00 V DCRange of frequency 700Hz ±100 Hz

Eccentric vibrational engine

3 V DC

Electro-magnetic Shaker

Shaker Specifications:

27 mm

13 mm

19 mm

6 mm

14

18 mm

Signal Force – Data physics Corporation (CE).Type: V4, Serial No. 06|A6Q|21318Power: 100 Watts Weight: 1.7KG.Max Sine Force: 17.8N, Max Random Force: 5.9N, Max Shock Force: 53.4NMax Acceleration: 91g, Max Velocity: 1.49mps, Peak to Peak Displacement: 5mm Armature Diameter: spigot, Armature mass: 20gm, Armature thread: M4 Armature Resonance: 12000Hz, Freq Range: 0-14000, Static Payload Support – Axial Stiffness: 0.45 Kgf/mm,

Shaker Dimensions:

15

SHAKER SPECIFICATION

Load cell:

S-Type Load cell

SN-9404

16

Shaker V2 is small in dimensions and can be suitable for developing a compact stiffness meter for the Muscle.

Analysis of force sensor commercially available and suitable for the stiffness measurements

Piezo-electric transducer model-219A05, price 1417 € Charge converter model either 422E12 (10 mV/pc) or 422E13 (1 mV/pc) price 282 €

each Cable for connecting force transducer to charge converter. Connectors 5-44,

transducer side, 10-32, charge converter side, length of about 1 meter, 3 feet. Model 003G03, price 68 €

PCB Piezotronics srlCentro Direzionale Rondo' di Curnasco Via F.lli Bandiera 2Treviolo/BG, 24048ITALY  Paolo Sermisoniinfo@pcbpiezotronics.it  Phone:+39 035 201421

17

Flexiforce sensorThickness 0.008" (.208mm)Length 8" (203mm) 6" (152mm) 4" (102mm) 2" (51mm)Width 0.55" (14mm)Sensing Area 0.375" diameter (9.53mm)Connector 3-pin male square pinLinearity Error <+/-5%Repeatability <+/-2.5% of full scaleHysteresis <4.5% of full scaleResponse Time <5 microsecondsForce Ranges 0-1 lb. (4.4 N) 0-25 lbs. (110 N) 0-100 lbs. (440 N)

Strain gages system

Vimel – sistemi metrologiciDott. Ing Ugo TurinVia Saluzzo 11610064 Pinerolo TOE-mail: uturin@vimel.itTel: 0121-375500Avaiable models:

1. ZF(BA)300 1AA (XX)Grid size: 1.1 x 1.3 mmBacking dimension 3.6 x 2.2 mm

2. ZF350-1AA(11)-W-XGrid size: 1.0 X 2.0 mmBacking dimension 3.0 x 2.0 mm

Load cell BursterPresent at LISiN:

1. Load cell Burster Typ: 8416-6001 1 KN

18

Schematic of the LAOD cell

Micro load cell AIFP (Arthroscopically Implantable Force Probe) 

19

10 mm

3 mm

6 mm

Via Pietro Mazza, 48 27057 Varzi - Pavia - Italia tel.: 0383.54.51.51 - fax: 0383.54.50.51

E-Mail: sales@spare.it

http://www.spare.it/ds_aifp.htm

- max force: 16 N - excitation: 14 mA cc - sensitivity: 30 mV/mA/N - non linearity:  ± 1,5%  fino a ½FS; ± 2,5%  fino a FS - hysterisis: 0,3% FS - non repeatability: 0.3% FS

Subminiature Load Cell tension compressionMod: 8417■ Measuring ranges from 0 ... 500 N■ Small dimensions■ Made of stainless steel■ Rugged construction■ Easy mounting by male threadsburster Tel. +49-07224-645-0 . Fax 645-88www.burster.com info@burster.de

Piezo-Materials

PVDF film 110 micron, .75" X 2.5"

PVDF film, 28 micron, .75" X 2.5"

Sensor #2910 with attached leads.

Coaxial PVDF - 5"

PZT wafer .2" X .8" X .02"

20

A complete sample package for the above films cost $30Unshielded PVDF Piezo Film Sensors

These 28 micron PVDF film sensors come with unshielded leads attached. The silver ink electrodes are covered with a clear polyester laminate on one side and urethane coated on the other.

TCL #2910   .640" X 1.63"  $ 7.50 TCL #3745   .640" X 2.86"  $13.50

TCL # 2405   .860" X 6.72"  $22.50

Shielded PVDF

TCL # 0288     (folded 28 micron film) .520" X 1.18" $29.95 The 0288 is fully shielded, with a shielded cable attached. The film is folded over, with the ground on the outside.

21

Relevant CircuitsTo drive the Shaker a 20W power amplifier along with a sinusoidal signal is used.The 20W power amplifier is being implemented through LM1875

The Pin Configuration for LM1875 and Circuit diagram:

Pin Diagram of LM1875 Circuit Diagram for Power Amplifier

Driving Circuit for the Flexiforce

Noise Elimination

22

Since the output of the Flexiforce driving circuit has high frequency noise component.

To Eliminate this noise I use a low pass filter with Fc(cut-off frequency)= 200Hz.

The values of the components are:

R1=R2=7.95KC1=C2= 0.1 microF

Fig. A Low Pass Filter with Fc=200Hz

Fig 6. Showing the Final driving circuit for the flexiforce with an isolated battery power supply inside it

23

Methodology

Our task is to estimate the contraction force of the muscle using the fact that it is related with the transversal impedance of the muscle.

To measure the transversal impedance of the muscle we can calculate the stiffness of the muscle which can be estimated through the simple formula: Stiffness of the material = (force applied normal to the surface)/ (depth penetrated to the surface).

In our case the surface is skin but we want to measure is stiffness of muscle which lies beneath the skin and adds mechanical impedance to system being in series with the muscle mechanical impedance. But the impedance of the skin is very less compared to the impedance of the muscle and can be neglected in our case.

To get an idea of the Stiffness of the muscle we have to calculate the force and the depth penetrated. For that we impose a constant displacement and measure the feedback force and then we calculate the stiffness.

To impose the constant the displacement we choose and Electro-Magnetic shaker which generates vibration of constant amplitude for a constant frequency an input power to the shaker. For our purpose we fixed the input power (at 4.2Vpp) and took readings at different frequencies so that for a particular set of reading we have constant amplitude of displacement.

Flexiforce is used to measure the feedback force. The sensor is coupled with an inverting amplifier and a low pass filter which produces a voltage proportional to the force applied to the sensor.

Following things should be kept in mind while setting up the device: The used sensor, which is coupled mechanically to the tip of the shaker , should be

very light in order to avoid any addition of the mechanical impedance of the sensor’s to the system, our system is analogous to the electrical system and if we consider ‘force’ equivalent to the ‘voltage’ then mass becomes inductance and stiffness is capacitance.

While sand witching the flexiforce it should be kept in mind that the force is completely transferred to the sensor’s sensing spot and is not shared by some other component which otherwise will decrease the sensitivity of the sensor.

24

Calibration of the Components

1. Calibration of the Flexiforce :

Fig 6.1: Calibration graph for the flexiforce i.e. Force applied Vs voltage output in mV.

To calibrate the flexiforce we sand witched it between to disc and then put it on the

digital weighing pan such the one surface of the disc touches the surface of the pan

perfectly, at the same time flexiforce is connected to its driving circuit and the voltage

output of the circuit is monitored by a DVM(digital volt meter).

We put sufficient weights to cover up our desired range and then noted down the values

of DVM for each weight.

25

Calibration of the Shaker:

Fig 6.2: The calibration plot for the shaker is Amplitude of the displacement of the shaker Vs Stiffness exposed (here calibrated springs) at different frequencies (Magenta=30Hz, Yellow=40Hz, Green=60Hz, Red=80Hz, Green=100Hz, Cyan=120Hz) and at constant voltage (4.2Vpp) supplied to the shaker.

The aim of this calibration graph is to understand the behavior of the shaker’s displacement amplitude when it is subjected to varying stiffness. In our case we measured the Displacement of the shaker’s tip by laser displacement meter, which is a very high precision instrument, while exposing it to different springs of varying stiffness from K=0.61 to K=4.4 N/mm and thus providing good range in terms of stiffness required in our case. Laser meter was kindly made available by Prof. Fasana of Physics department of Politecnico di Torino.

Inference from this calibration graph can be drawn that, since we want to have a steady displacement shaker or the amplitude of the displacement should remain constant through out the experiment, So for frequencies like 120Hz, 100Hz and 80Hz do provide us such condition with this shaker but at the same time the amplitude decreases and hence the sensitivity of the measurement will decrease.

26

Results

Biceps

Fig 7.1: The above plot shows the relation between the force produced by the Muscles acting on the elbow and the Transversal stiffness of the biceps measured at different frequencies (yellow is 80Hz, Green is 100Hz, red is 120 Hz).

To collect this data we set-up a small mechanism to fix the arm and the shaker in such a fashion

that the shaker’s tip (containing the force transducer ‘flexiforce’) hits the surface of the skin and

the muscle vertically, to insure that we measure transversal stiffness and not a component of

transversal and axial.

Once we fix the arm and frequency (80Hz), we take measurement of the skin without any load

that is with no contraction of the muscle and then we apply and increase the load (by holding

some weights) in steps of 2 Kg i.e. 2, 4, 6 and 8 Kgs.

We repeat the same thing at frequencies 100 and 120 Hz.

27

Fig 7.2: Picture demonstrating the process of holding the weight in the hand and acquiring the flexiforce output.

For all the samples of data we took, we maintain a constant geometry for holding the weight so

that we minimize the error introduced because of geometric artifacts.

To get a better Idea of the behavior of transversal stiffness and the global force (here in our case),

we took more number of samples of the same subject at different frequencies and at different

positions. Figure 7.3 explains different positions of the flexiforce, we sampled at 3 main

positions, 1st and 3rd one are directly above the short and long headed muscle and the 2nd one is

somewhere in between these 2.

28

Fig 7.3: Picture showing different positions of flexiforce, right-most is the Pos1 middle is Pos2 and the left-most is the Pos3.

Fig 7.4: The plot Between Transversal stiffness and Global force at 80 Hz (yellow is Pos1 green

is Pos2 and red is Pos3).

29

Fig. 7.5: The plot Between Transversal stiffness and Global force at 100 Hz (yellow is Pos1 green is Pos2 and red is Pos3)

Fig. 7.6: The plot Between Transversal stiffness and Global force at 120 Hz (yellow is Pos1 green is Pos2 and red is Pos3)

30

Fig 7.7: Plot between stiffness and global force for all the data plotted in Fig 7.6, 7.5 &7.4

(Data1, 2 & 3 represent the 120Hz with yellow, green and red representing Pos1, Pos2 and Pos3

respectively.

Data4, 5 & 6 represent the 80Hz with yellow, green and red representing Pos1, Pos2 and Pos3

respectively.

Data7, 8 & 9 represent the 100Hz with yellow, green and red representing Pos1, Pos2 and Pos3

respectively.

31