www.elsevier.com/locate/meatsci

Meat Science 76 (2007) 321–326

MEATSCIENCE

Sonoelasticity to monitor mechanical changes during rigor and ageing

A. Ayadi b,*, J. Culioli a, S. Abouelkaram a

a Station de Recherches sur la Viande, INRA de Theix, 63122 Saint Genes Champanelle, Franceb Laboratoire des Systemes Electro-Mecaniques, ENIS, Route Soukra km 3.5, B.P. W, 3038 Sfax, Tunisia

Received 4 February 2006; received in revised form 12 October 2006; accepted 27 November 2006

Abstract

We propose the use of sonoelasticity as a non-destructive method to monitor changes in the resistance of muscle fibres, unaffected byconnective tissue. Vibrations were applied at low frequency to induce oscillations in soft tissues and an ultrasound transducer was used todetect the motions. The experiments were carried out on the M. biceps femoris muscles of three beef cattle. In addition to the sonoelas-ticity measurements, the changes in meat during rigor and ageing were followed by measurements of both the mechanical resistance ofmyofibres and pH. The variations of mechanical resistance and pH were compared to those of the sonoelastic variables (velocity andattenuation) at two frequencies. The relationships between pH and velocity or attenuation and between the velocity or attenuationand the stress at 20% deformation were highly correlated. We concluded that sonoelasticity is a non-destructive method that can be usedto monitor mechanical changes in muscle fibers during rigor-mortis and ageing.� 2006 Elsevier Ltd. All rights reserved.

Keywords: Sonoelasticity; Compression; pH; Biceps femoris; Rigor-mortis; Ageing

1. Introduction

The mechanical properties of muscle change markedlypostmortem and induce variation in the tenderness ofmeat. Although several methods were developed to mea-sure the mechanical properties of muscle fibres and connec-tive tissue as well as interactions between them, few canreflect the resistance of muscle fibres, uninfluenced by con-nective tissue (Bouton, Ford, Harris, & Ratcliff, 1975; Har-ris & Shorthose, 1988; Lepetit, & Culioli, 1994). However,a selective measurement is necessary to compare the ageingprocess of muscles with different collagen contents.

In the case of pre-rigor raw meat, Locker and Wild(1982a, 1982b) have shown that when a muscle is progres-sively tensioned, the stress-time curves of a tensile test showa plateau, representing the yield point of muscle fibres. Thisyield occurs at a deformation lower than that needed totension the collagen network. Sacks, Kronick, and Buech-

0309-1740/$ - see front matter � 2006 Elsevier Ltd. All rights reserved.

doi:10.1016/j.meatsci.2006.11.016

* Corresponding author. Tel.: +216 74 229 861; fax: +216 74 676 903.E-mail address: a.ayadi@ipein.rnu.tn (A. Ayadi).

ler (1988), also using a tensile test, have shown that thestress measured at a strain lower than 0.13 was supportedby muscle fibres only. Lepetit and Sale (1984, 1985) haveshown that in linear compression test, where a sample ofraw meat is compressed perpendicularly to muscle fibresand it can deform laterally only in the direction of musclefibres, then the mechanical properties of these muscle fibrescan be measured, without contributions from connectivetissue, at strain lower than a critical value, which can becalculated. Lepetit (1991) used a structural model for thedeformation of helical collagen fibres, assuming that thedeformation of muscle fibres was isovolumic, and calcu-lated the variation of this strain with sarcomere length.For rest length (2 lm) meat the model indicated that nocollagenous fibres were tensioned if the strain in the direc-tion of muscle fibres is less, or equal, to 0.25. He further,concluded that the strain of 0.2 is the largest deformationavailable in compression for the measurement of mechani-cal properties of muscle fibres, without contributions fromcollagen.

For dynamic perturbations, Sale, Noel, Lasteyras, andOleon (1984) used a sinusoidal compression test to measure

Sample

Probe

Fig. 1. Cell of linear compression with sample.

322 A. Ayadi et al. / Meat Science 76 (2007) 321–326

the mechanical properties of muscle fibres. They concludedthat in the post-mortem state the muscle fibres mechanicalproperties were more influenced by the strain rate ratherthan the connective tissue. Abouelkaram, Berge, and Culi-oli (1995) and Abouelkaram et al. (2000) used an ultrasonicmethod to characterise and classify bovine muscles. Themeasurement of speed and attenuation of ultrasound wavescan provide useful information on the composition and themechanical properties of the soft tissue (Miles, Fisher, Fur-sey, & Page, 1987; Miles, Fursey, Page, & Fisher, 1990;Park, Whittaker, Miller, & Hale, 1994).

Recently, a number of authors used low frequencyvibration to evaluate the characteristics of biological tis-sues (Catheline, Thomas, Wu, & Fink, 1999; Catheline,Wu, & Fink, 1999; Fujii et al., 1995; Lerner, Huang, & Par-ker, 1990; Parker, Huang, Musulin, & Lerner, 1990;Yamakoshi, Sato, & Sato, 1990). In this method (sonoelas-ticity) low frequency vibration is applied to the tissues andtheir wave propagation characteristics are measured bysimultaneously transmitted ultrasonic waves. Some authorsemployed this method to produce hardness maps fromamplitude and phase data of low frequency propagation(Lerner et al., 1990; Yamakoshi et al., 1990). The othersused it to show variations in mechanical proprieties of dif-ferent regions of tissue (Catheline, Thomas et al., 1999;Catheline, Wu et al., 1999; Fujii et al., 1995; Parkeret al., 1990). However, it is known that compressionalmechanical vibration at low frequency generates, in soft tis-sue, a wave dominated by shear, where the values of veloc-ity and attenuation are of the same order. Indeed, it wasbeen demonstrated that, when the sample has inhomoge-neous regions, the use of a non negligible sized vibratorcan add to the reflected waves at boundaries, and the com-pressional vibrations at low frequency generate a shearwave slightly affected by a longitudinal one. In our work,we used the same technique of signal processing as Cathe-line et al. (1999), not to value the local characteristics or touse a hardness map, but to extract the global velocity andattenuation coefficient of the low frequency wave propaga-tion in the sample. We consider that the sonoelasticitymethod can be used as a non-destructive one to measuremyofibres resistance without collagen contribution.

It is important to point out that there are several workson the relationship between ultimate muscle pH and ten-derness within the pH range normally encountered inpost-rigor muscle. Bouton, Howard, and Lawrie (1957)have indicated that curvilinear relationship exists betweentenderness and pH.

The aim of the current work was to study the changes inmuscles during rigor onset and ageing. The changes havebeen followed by measurements of the sonoelastic parame-ters such as velocity of propagation and attenuation coeffi-cient at 80 and 100 Hz. The variations in these parameterswere compared to those of both pH and mechanical resis-tance obtained by linear compression. The experimentswere carried out on the M. biceps femoris muscles fromthree different animals.

2. Materials and methods

2.1. Materials

The experiments were carried out on three M. biceps

femoris (BF) muscles excised, one hour postmortem, fromthe carcasses of three different animals: a cull cow (10 yearsold, Holstein-Frisonne, carcass weight 294 kg) and twoyoung bulls (18 months old, Charolais, 384 kg and440 kg). The samples for the sonoelasticity measurementswere cut from the core of the muscles. The height, length,and width were respectively for the first animal (An1) 6,11, 11 cm, for the second (An2) 7, 11, 8 cm and for thethird (An3) 5.5, 11, 9 cm. Meat samples were wrapped ina plastic film to avoid evaporation and were kept at a tem-perature of 20 �C throughout the experiment. Sonoelasticmeasurements were performed perpendicular to the myofi-bril direction. The mechanical measurements were carriedout on meat slices cut off from another part of the samemuscle. The pH determinations (every hour until the ulti-mate pH value was reached) were performed both on acontrol and on the samples used for the sonoelasticity mea-surements. The first measurements were performed 2.5 hafter slaughter. However, to assure the measurement ofmaximum value of stress postmortem, the first mechanicaldetermination was carried out when the pH of the controlreached a value below 5.9.

2.2. Linear compression measurements

Linear compression measurements were carried out withan INSTRON 4501 Traction-compression machine. Thedata were obtained and processed with the INSTRON soft-ware. The 1 cm2 square-shaped probe provided a linearcompression motion. Samples were held in a two-sided cellduring compression. The muscle fibres of the sample layparallel to the sides of the cell and strain was applied atright angles to the fibres (Fig. 1). Samples were 10 mmthick and much care was taken to ensure this varied mini-mally. During the test, the probe that was initially located20 mm from bottom of the cell was moved down 18 mm.

6.1

6.4(N

/cm

2 ) 30

40pHStress

pH

A. Ayadi et al. / Meat Science 76 (2007) 321–326 323

This displacement was enough to produce a 0.2 compres-sion strain even in a sample less than 10 mm high. The realheight of the sample was determined as the height when ayield stress equal to 0.7 N/cm2 was applied on the probe.The software takes into account the real height and mea-sures the exact stress for 0.2 compression strain. The probedisplacement speed was set to 50 mm/min. All measure-ments were carried out at room temperature. Mean valueswere obtained from 10 to 12 measurements.

2.3. pH and temperature measurements

The pH was measured with a Corning pH electrode(METTLER TOLEDO) and a SCHOTT portable pHmeter (type CG 837). The temperature was measured witha thermocouple to correct pH values. The glass electrodeand thermocouple point were inserted 2–3 cm into the sam-ples. Two measurements of pH were used for each animal,one in the sonoelasticity sample and the second in the con-trol sample. Each pH value was the average of four mea-surements taken at four different points of bothsonoelasticity sample and control sample.

2.4. Sonoelasticity measurements

The sonoelasticity measurements were based on the useof two waves of low (perturbation) and high frequency(ultrasound), simultaneously emitted in the sample(Fig. 2). Each wave had a specific role. The first one wasused at two frequencies 80 and 100 Hz. It induced the softtissue to vibrate with 0.1 mm of amplitude with the 30 mmdiameter circular probe. The vibration propagates in thesoft tissue, so that all the planes at different depths vibratewith different amplitudes and with different phase delays.The second one (ultrasound transducer, frequency5 MHz) allows the analysis of the displacements of the dif-ferent planes by measuring the displacement of connectivetissue as local diffusers. A-scan lines were obtained bytransducer and recorded in a digitizer operating at

Transducer

Coupling media

Sample

Probe

Fig. 2. Sonoelasticity device.

40 MHz. Each line was divided into 40 segments, corre-sponding to slices within the soft tissue. The deformationsof slices were calculated with a correlation techniquebetween successive A-lines. Both velocity and attenuationcoefficient were determined by measuring, respectively,the global phase and amplitude degradation of the scatter-ing planes oscillations. Those parameters depend on themechanical properties of the soft tissue, i.e. mainly the elas-ticity and the viscosity. The determination of mechanicalproperties was not the aim of the work, because, the char-acteristics of propagation of imposed vibrations are notthose of a shear wave.

3. Results analysis

Before discussing the results obtained by sonoelasticity,linear compression and pH, we present typical curvesobtained by these three methods of evaluation. In Fig. 3athe pH and stress parameters changes are shown. pHdecreased during rigor-mortis and reached a constantvalue. In the linear compression test, the rigor point wascharacterised by the higher stress at 20% deformationcurve. At rigor-mortis, the stress increased quickly to reachthe maximum value. In Fig. 3b the evolution of parametersof the sonoelastic wave, is shown. The velocity curve had,qualitatively, the same evolution as the stress curve. In con-trast, the attenuation curve had the inverse shape.

The relationships between pH and velocity, or attenua-tion, were only determined for pH values that were higherthan 5.4. In Fig. 4a–d the variation with pH in the velocityand attenuation coefficients, at two frequencies, 80 and100 Hz, is shown. While pH decreases from value less than6.5–5.4, the velocity of the mechanical waves increased by a

5.2

5.5

5.8

0 20 40 60 80

Time (hour)0

10Str

ess

Atte

ntio

n (d

B/c

m)

Vel

ocity

(m

/s)

20

0

1

2

3

4

5

0 20 40 60 80

Time (hour) 0

10

20

30AttenuationVelocity

Fig. 3. Typical evolution of different parameters on biceps femoris for thethird animal (An3). (a) Evolution of both stress at 20% of strain and pH;(b) evolution of both velocity and attenuation of the low frequency wavepropagation at 100 Hz.

0

2

4

6

8

5.3 5.6 5.9 6.2 6.5

pH

An1An2An3Regression

0

10

20

30

5.3 5.6 5.9 6.2 6.5

pH

An1An2An3Regression

0

2

4

6

8

5.3 5.6 5.9 6.2 6.5

pH

An1An2An3Regression

0

10

20

30

5.3 5.6 5.9 6.2 6.5

pH

An1An2An3Regression

Atte

ntio

n (d

B/c

m)

Atte

ntio

n (d

B/c

m)

Vel

ocity

(m

/s)

Vel

ocity

(m

/s)

Fig. 4. Correlation between the evolution of low frequency wave parameters and the evolution of pH of muscle for each animal and regression betweenthese variables for all animals. (a) Attenuation of the wave at 80 Hz; (b) attenuation of the wave at 100 Hz; (c) velocity of the wave at 80 Hz; (d) velocity ofthe wave at 100 Hz.

324 A. Ayadi et al. / Meat Science 76 (2007) 321–326

factor of 2 (from 12.5 to 25 m/s). In contrast, the attenua-tion coefficient decreased steeply from 8 dB/cm to less than2 dB/cm. These results were in agreement with the fact thatrigor onset induced an increase in the rigidity of meat and,as a consequence, a lower contribution of the viscous com-ponent to the mechanical properties.

The correlation coefficients between pH and velocity orattenuation are indicated in Table 1 below, separately foreach animal and also for all animals. Except in one case(animal 3, pH-velocity at 100 Hz), the absolute linearregression values ‘‘r’’ were statistically significant(P < 0.05) and are higher than 0.8, even reaching 0.97.The combination of the results of the three animalsdecreased the r-values that, however, remained significant(n = 39, P < 0.05). This decrease can be explained by thefact that, the sonoelastic variables are not much linked tothe pH of the whole of animals, and can be influenced, todifferent extents for each animal, by other factors.

Fig. 5a–d show the relationships between the velocity orattenuation, at 80 and 100 Hz, and the mechanical resis-tance of the meat. These relationships were establishedfor the ageing period, i.e. after pH had reached the con-stant value. During ageing the stress decreased from values

Table 1Correlation coefficients obtained between velocity or attenuation and stress or

Wave parameters Animal 1 Animal 2

Stress, n = 6 pH, n = 13 Stress, n = 5 pH

Velocity 80 Hz 0.95 �0.81 0.82 �0.

Attenuation 80 Hz �0.87 0.89 �0.80 0.

Velocity 100 Hz 0.99 �0.88 0.86 �0.

Attenuation 100 Hz �0.97 0.96 �0.89 0.

Values in bold indicate significant correlation (n – number of values of rigor o

of 35 N/cm2 to a final value of around 4 N/cm2. Thisdecrease in the mechanical resistance of the meat induceda decrease in the velocity of the mechanical wave and,inversely, an increase in the attenuation, which indicateda larger contribution of the viscous component of the softtissue.

The stress values were initially different for all three ani-mals (from 10 to 25 N/cm2), due either to different intensityof rigor-mortis or to fast ageing of each muscle. The initialvelocities were also different for each animal, at both fre-quencies 80 and 100 Hz (from 16 to 25 m/s), although thefinal velocity values were similar, between 12 and 15 m/s.The correlation coefficients (Table 1) between the velocityand stress were above 0.86, except for animal 2 at 80 Hz(r = 0.82, P < 0.05). However, the correlation coefficientsare smaller when the three animals are considered together(r = 0.79 at 100 Hz and r = 0.85 at 80 Hz; n = 13, P < 0.05).

The initial attenuation coefficient was dependent on theanimal, at both frequencies (from 1.2 to 3.5 dB/cm). Itincreased markedly when meat aged, to reach values upto 6 dB/cm. The correlation coefficient between attenuationand stress was higher than 0.85 (P < 0.05), except for theanimal 2 at 80 Hz (r = 0.80, P < 0.05). These coefficients

pH parameters

Animal 3 All animals

, n = 17 Stress, n = 6 pH, n = 9 Stress, n = 17 pH, n = 39

88 0.99 �0.94 0.85 �0.74

97 �0.85 0.96 �0.81 0.68

79 0.97 �0.62 0.79 �0.54

93 �0.95 0.81 �0.79 0.59

r ageing development, P < 0.05) according to the FISHER test.

0

2

4

6

8

0 10 20 30 40

Stress (N/cm2)

An1An2An3Regression

0

10

20

30

0 10 20 30 40

An1An2An3Regression

0

2

4

6

8

0 10 20 30 40

Stress (N/cm2)

Stress (N/cm2)

Stress (N/cm2)

An1An2An3Regression

0

10

20

30

0 10 20 30 40

An1An2An3Regression

a

b

c

d

Fig. 5. Correlation between the evolution of low frequency wave parameters and the evolution of stress at 20% of strain of muscle for each animal andregression between these variables for all animals. (a) Attenuation of the wave at 80 Hz; (b) attenuation of the wave at 100 Hz; (c) velocity of the wave at80 Hz; (d) velocity of the wave at 100 Hz.

A. Ayadi et al. / Meat Science 76 (2007) 321–326 325

were smaller for velocity when results from the three ani-mals were combined. Moreover, although they remainedsignificant (r = �0.81 at 80 Hz and r = �0.79 at 100 Hz,n = 13, P < 0.05), they were also smaller than thoseobtained with velocity.

We believe from these results,that muscles studied in thiswork at a temperature of 20 �C, can be considered aged asindicated: on the one hand by mechanical measurementswhen stress is lower 5.5 N/cm2, and on the other handfor sonoelastic variables, when the velocity is lower than17 m/s at 80 Hz, and 13 m/s at 100 Hz, or when the atten-uation is higher than 2.8 dB/cm at 80 Hz and 3.8 dB/cm at100 Hz.

4. Conclusion

The relationships between pH and velocity or attenua-tion were only determined for pH values higher than 5.4.The results showed that during rigor development, thevelocity of the mechanical waves increased with decreasingpH. In contrast to the velocity, the attenuation coefficientdecreased steeply. The relationships between velocity, orattenuation, and stress at 20% deformation have beenestablished exceptionally for the ageing period (after pHhad reached the constant value 5.4) and showed significantcorrelations with high coefficients according to FISHERtest (P < 0.05). These results of decreasing attenuationand increasing velocity were in agreement with the fact thatduring rigor-mortis onset the rigidity of meat increases.During ageing, the mechanical resistance of meat decreasesand largely affects the viscous component of the soft tissue,and induces attenuation increases and velocity decreases ofthe mechanical wave. In short, one can remark that themeat can be considered aged, at a temperature of 20 �C,

when the stress is lower than 5.5 N/cm2, and when thevelocity is lower than 17 m/s at 80 Hz, and 13 m/s at100 Hz, or when the attenuation is higher than 2.8 dB/cmat 80 Hz and 3.8 dB/cm at 100 Hz. From these results itcan be concluded that the sonoelastic method can be usedto follow both rigor-mortis and ageing.

Acknowledgement

This work was accomplished at Station de Recherchessur la Viande, INRA.

References

Abouelkaram, S., Berge, P., & Culioli, J. (1995). Caracterisation ultras-onore des tissus musculaires bovins. Recontres autour des Recherches

sur les Ruminants, 2, 231–234.Abouelkaram, S., Suchorski, K., Buquet, B., Berge, P., Culioli, J.,

Delachartre, P., & Basset, O. (2000). Effects of muscle texture onultrasonic measurements. Food Chemistry, 69, 447–455.

Bouton, P. E., Howard, A., & Lawrie, R. A. (1957). Studies on beefquality. VI. Effects on weight losses and eating quality of other pre-slaughter treatments. CSIRO Division of Food Preservation and

Technologys (Paper No. 6).Bouton, P. E., Ford, A. L., Harris, P. V., & Ratcliff, D. (1975). Objective–

subjective assessment of meat tenderness. Journal of Texture Studies, 6,315–328.

Catheline, S., Thomas, J. L., Wu, F., & Fink, M. (1999). Diffraction fieldof a low frequency vibrator in soft tissues using transient elastography.IEEE Transaction on Ultrasonics, Ferroelectrics, and Frequency Con-

trol, 46(4), 1013–1019.Catheline, S., Wu, F., & Fink, M. (1999). A solution to diffraction biases

in sonoelasticity: The acoustic impulsion technique. Journal of the

Acoustical Society of America, 105(5), 2941–2950.Fujii, K., Sato, T., Kameyama, K., Inoue, T., Yokoyama, K., &

Kobayashi, K. (1995). Imaging system of precise hardness distributionin vivo using forced vibration and ultrasonic detection. Acoustical

Imaging (Vol. 21). New York: Plenum.

326 A. Ayadi et al. / Meat Science 76 (2007) 321–326

Harris, P. V., & Shorthose, W. R. (1988). Meat texture. In R. Lawrie(Ed.). Developments in meat science – 4 (pp. 245–296). London:Elsevier Applied Science.

Lepetit, J., & Sale, P. (1984). Proprietes rheologique de la viande encompression. Proceedings of the 30th European meeting of meat

research workers (4, pp. 201–203). .Lepetit, J., & Sale, P. (1985). Analyse du comportement rheologique de la

viande par une methode de compression sinusoıdale. Science des

Aliments, 5, 521–540.Lepetit, J. (1991). Theoretical strain ranges in raw meat. Meat Science, 29,

271–283.Lepetit, J., & Culioli, J. (1994). Mechanical properties of meat. Meat

Science, 36, 203–237.Lerner, R. M., Huang, S. R., & Parker, K. J. (1990). Sonoelasticity images

derived from ultrasound signals in mechanically vibrated tissues.Ultrasound in Medicine and Biology, 16(3), 231–239, Printed in USA.

Locker, R. H., & Wild, D. J. C. (1982a). A machine for measuring yieldpoint in raw meat. Journal of Texture Studies, 13, 71–82.

Locker, R. H., & Wild, D. J. C. (1982b). Yield point in raw beef muscle.The effects of ageing, rigor temperature and stretch. Meat Science, 7,93–107.

Miles, C. A., Fisher, A. V., Fursey, G. A. J., & Page, S. J. (1987).Estimating beef carcass composition using the speed of ultrasound.Meat Science, 21, 175–188.

Miles, C. A., Fursey, G. A. J., Page, S. J., & Fisher, A. V. (1990). Progresstowards using the speed of ultrasound for beef leanness classification.Meat Science, 28, 119–130.

Park, B., Whittaker, A. D., Miller, R. K., & Hale, D. S. (1994). Ultrasonicspectral analysis for beef sensory attributes. Journal of Food Science,

59, 697–701.Parker, K. J., Huang, S. R., Musulin, R. A., & Lerner, R. M. (1990).

Tissue response to mechanical vibrations for ‘‘sonoelasticity imaging’’.Ultrasound in Medicine and Biology, 16(3), 241–246, Printed in USA.

Sacks, M. S., Kronick, P. L., & Buechler, P. R. (1988). Contribution ofintramuscular connective tissue to the viscoelastic properties of post-rigor bovine muscle. Journal of Food Science, 53, 19–24.

Sale, P., Noel, Y., Lasteyras, A., & Oleon, C. (1984). Sinusoidalcompression system to study rheological properties of foods intransient state. Journal of Texture Studies, 15, 103–114.

Yamakoshi, Y., Sato, J., & Sato, K. (1990). Ultrasonic imaging of internalvibration of soft tissue under forced vibration. IEEE Transaction on

Ultrasonics, Ferroelectrics, and Frequency Control, 37(2), 45–53.