ReviewMeat quality assessment using biophysical methods related to meat structureJean-Louis Damez*, Sylvie ClerjonINRA, UR370 QuaPA, F-63122 Saint Gens Champanelle, Francearti cle i nfoArticle history:Received 29 February 2008Received in revised form 21 May 2008Accepted 26 May 2008Keywords:Meat qualityMeat structuresensorSpectroscopyMicroscopyBiophysical methodsOptical propertiesNMRMRIUltrasoundsX-raysMechanical propertiesImpedanceMicrowaveFluorescencePolarizationAnisotropyabstractThis paper overviews the biophysical methods developed to gain access to meat structure information.The meat industry needs reliable meat quality information throughout the production process in ordertoguaranteehigh-qualitymeatproducts forconsumers. Fast and non-invasivesensorswill shortlybedeployed, based on the development of biophysical methods for assessing meat structure. Reliable meatquality information (tenderness, avour, juiciness, colour) can be provided by a number of different meatstructure assessment either by means of mechanical (i.e., WarnerBratzler shear force), optical (colourmeasurements, uorescence) electrical probing or using ultrasonic measurements, electromagneticwaves, NMR, NIR, and so on. These measurements are often used to construct meat structure images thatare fusioned and then processed via multi-image analysis, which needs appropriate processing methods.Qualitytraitsrelatedtomechanical propertiesareoftenbetter assessedbymethodsthat takeintoaccountthenatural anisotropyofmeatduetoitsrelativelylinearmyobrillarstructure. Biophysicalmethods of assessment can either measure meat component properties directly, or calculate them indi-rectly by using obvious correlations between one or several biophysical measurements and meat compo-nent properties. Takingthese calculations andmodelling the mainrelevant biophysical propertiesinvolved can help to improve our understanding of meat properties and thus of eating quality. 2008 Elsevier Ltd. All rights reserved.Contents1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1332. Mechanical methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1352.1. Instrument measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1352.2. Ultrasound methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1353. Optical methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1353.1. Spectroscopic methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1353.1.1. Infrared spectroscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1363.1.2. Raman spectroscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1363.1.3. Visible spectroscopy and colorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1373.1.4. Fluorescence spectroscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1373.2. Imaging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1383.2.1. Microscopic imaging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1383.2.2. Optical microscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1383.2.3. Histology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1383.2.4. Confocal laser scanning microscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1393.2.5. Electron microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1393.2.6. Scanning electron microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1390309-1740/$ - see front matter 2008 Elsevier Ltd. All rights reserved.doi:10.1016/j.meatsci.2008.05.039*Corresponding author. Tel.: +33 4 73 62 41 87; fax: +33 4 73 62 40 89.E-mail address: damez@clermont.inra.fr (J.-L. Damez).Meat Science 80 (2008) 132149ContentslistsavailableatScienceDirectMeat Sciencej our nal homepage: www. el sevi er. com/ l ocat e/ meat sci3.2.7. Transmission electron microscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1403.2.8. Macroscopic imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1404. Dielectric methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1414.1. Impedance measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1414.1.1. Detection of frozen meats. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1414.1.2. pH. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1414.1.3. Fat content. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1414.1.4. Tenderness. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1414.1.5. Ageing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1414.2. Microwave characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1415. X-ray measurements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1426. Nuclear magnetic resonance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1426.1. NMR spectroscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1426.2. Magnetic resonance imaging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1436.3. Magnetic resonance elastography. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1437. Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1441. IntroductionOnechallengefacingthemeat industryistoobtainreliableinformationonmeatqualitythroughouttheproductionprocess,whichwouldultimatelyprovide a guaranteedqualityof meatproducts for consumers. To meet this challenge requires fast, accu-rate and non-invasive techniques for predicting technological andsensoryqualities. Overthelastfewyears, anumberofmethodshavebeendevelopedtoobjectivelymeasuremeatqualitytraits.The majority of these methods are invasive, meaning that a samplehas to be taken or that they are difcult to implement on-line. Inmusclefood, thepivotal qualitativecharacteristicsthat needtobe determined are texture, nutritional value, and appearance. Sev-eral very promising measurement techniques are currently beingstudiedandusedinlaboratories, someof whichwill shortlybeready for industrial deployment.The great variability in raw meat leads to highly variable prod-uctsbeingmarketedwithout acontrolledlevel of quality. Thisproblem is aggravated when the industry is unable to satisfactorilycharacterize this level of quality and cannot therefore market prod-ucts with a certied quality level, which is an otherwise essentialcondition for the survival and development of any modern indus-try. Meat quality depends on the same criteria generally attachedtootherfood. Thebasictraitsrelatetonutritional contentsuchas proteins, fat, bers, vitamins and minerals, mainly iron. Anotherkeycriterionissafety. Thefoodmustbecleanintermsofagro-chemical residue, heavy metals, pathogenic micro-organisms, andanyothersubstancerepresentingapotential healthhazard. Theother aspect of quality deals with functional characteristics, i.e.,related to the sensory properties of taste and appearance (Grunert,1997). The variability in functional traits is in part related to thebiologicaldiversityoftheanimalsfromwhichmeatisobtained.Numerous studies have shown the important inuence of zootech-nical characteristicsonmeat tenderness, otherstudieshavefo-cusedoncollagenandmyobrillarstructure. Factorsinuencingmeat properties are partly related to breed, age and sex (Judge &Aberle, 1982; Huff-Lonergan, Parrish, &Robson, 1995; Horcada,Beriain, Purroy, Lizaso, &Chasco, 1998). Thesefactorsareeitherknown or can be contained and controlled. However, the variabilityin myobrillar and conjunctive components remains uncontrolled.In fact, meat toughness depends mainly on these two structures:the myobrillar structure, and conjunctive tissue. Myobrillarstructure is strongly inuenced by the animal rearing conditions.Forinstance, (Greenwood, Harden, &Hopkins, 2007), foundthatsingle- or multiple-reared lambs present signicant differences inmyobertypesandsoinmyobrils, whereas(Gondret, Combes,Lefaucheur, &Lebret, 2005), foundchanges inmyober typesaccording to indoor or outdoor rearing systems. On the other hand,conjunctive tissue is directly related to the zootechnical character-istics of the animal at slaughter. These components need to be as-sessed not only in terms of quantity but also in terms ofintramuscular distribution. Thespatial organizationof thecon-junctivenetworkof fat andmeat bersbundles, whichdenesthe meat grain and marbling, is one of the meat structure traitsstrongly connected to meat tenderness. The assessment of this traitis of prime interest, not only for the development of a diagnosticsystemmaking itpossibletodeterminethemuscularoriginofameat sample and therefore optimize production processes, but alsoas a non-invasive method of sorting muscle meat in terms of po-tential tenderness, sinceconsumerdemandisforconsistencyinmeat tenderness.Beyond tenderness, meat structure, as considered in this paper,groupstogethersensorypropertiesassociatedwithmeateatingliketexture, pastiness, crusting, palatability, chewiness, juicinessand of course tenderness. These sensory properties are associatedwith several objective physical properties of the product: fat con-tent, fat spatial organization, collagencontent, collagenspatialorganization, myobers spatial organization, myobers type, size,shape and density, sarcomere length, Z lines and I bands integrity,membranesintegrityandsarcolemmaattachmenttomyobrils.Water content also takes part in these physical properties becauseitisconnectedwithjuicinessandwithpale, softandexsudative(PSE) and dark rm dry (DFD) defects. These defects are more pre-cisely related to water holding capacity (WHC), a water propertycorrelated with water activity (Kuo & Chu, 2003). PSE and DFD de-fects can also be accessed thanks to the measurement of myobersmetabolism. Another indirect marker of structure is salt diffusionwhichdependsonstructureintegrity. Thisreviewgivesanon-exhaustiveoverviewofbiophysicalmethodswhichcanmeasureone or another of all these sensory and direct or indirect physicalparameters. ThesemethodsaresummarizedinTable1withthetypeof informationtheygiveandwiththeir mainadvantagesand drawbacks.Biophysical methodsof assessmentcaneithermeasuremeatcomponent properties directly or calculate them indirectly (Monin,1998)byusingobviouscorrelationsbetweenoneorseveralbio-physical measurements and meat component properties (Brunton,Lyng, Zhang, & Jacquier, 2006; Swatland, 1997b). This paper pin-points ve groups of biophysical assessment: mechanical methods,optical methods, dielectrical methods, X-raymeasurements, andJ.-L. Damez, S. Clerjon/ Meat Science 80 (2008) 132149 133nuclear magnetic resonance (NMR) measurements. Other reviewshavedealtwiththeon-lineevaluationofmeat(Swatland, 2003)or food (Scotter, 1997) quality, but we focused here on structureassessment.Table 1Summary of biophysical methods with the type of information they give in meat science and their main advantages and drawbacksMethods Type of information Advantages DrawbacksInstrumental mechanical methodsWarnerBratzlershear force testTenderness Destructive, anisotropicSlice shear forcetestTenderness More strongly correlated with tenderness than wbsf Destructive, anisotropic20% and 80% ofmeat sampledeformationTenderness Assess myobrillar proteins (20%) and intramuscularconnective tissues (80%) separatelyDestructive, anisotropicArmortenderometerTenderness Portable, non-destructive InvasiveTorquetenderometerTenderness Portable, non-destructive InvasiveTendertecpenetrometerTenderness Portable, non-destructive InvasiveUltra soundUltrasonic spectralanalysesTexture, fat content, collagen content 3D, on live animals, on whole carcassesUltrasonicelastography(or transientelastography)Local viscoelastic properties 3D, non-invasiveOptical spectroscopyInfraredspectroscopyStructure of molecules Non-contacting, rapid Need complex data analysisNear infraredspectroscopyStructure of molecules, instrumental texture, sensorytenderness, pastiness, crusting, juiciness, discriminationbetween fresh and frozen-thawed products, waterholding capacityaNon-contacting, rapid Need complex data analysisRamanspectroscopyStructure of molecules, interaction of molecules,structure of proteins, water activity, fat organization,water holding capacity, instrumental texture, tendernessNon-contacting, rapid, can be performed in vivo usingoptical ber, small sample portionNeed complex data analysisVisiblespectroscopy,colorimetryColour, sarcomere length, myolaments organization,PSE, tenderness, water holding capacity, drip loss,collagen content, sh freshnessOften performed with polarized lightFluorescencespectroscopyTryptophan microenvironment, connective tissuecontent, palatability, chewiness, sh freshness,tenderness, myober organization, collagendestructuration with heating, aging, sarcomere lentghNon-contacting, rapid, can be performed in vivo usingoptical ber, often performed with polarized lightOften need extrinsicuorophore probe, hightemperature sensitivityMicroscopyOpticalmicroscopyFat organization, collagen organization, myober typing,discrimination between fresh and frozen-thawedproducts, myober spacing, Z line degradation,sarcomere length, endomysium structure, myoberdiameter, myober density, myober organization, PSE,specic proteins detection, collagen typing,myolaments organizationSelective analyse, 3D reconstruction with confocalmicroscopySample preparation: thin cuts(except for confocalmicroscopy) and oftenstainingElectronmicroscopyMyolaments structure changes, structure of proteins,connective tissue organization, endomysium andperimysium structure, Z lines degradation, I band breaks,sarcolemma attachment to myobrilsGreater resolution and magnication, 3Dreconstruction for the SEM, observation of sampleswithout dehydration or freezing with environmentscanning electron microscopySample preparation:cryoxation, dehydration,embedding, or staining withheavy metalMacroscopicimagingCollagen organization, lipids organization, tenderness, fatcontent, collagen content, juicinessNon-contacting, rapid, easy to useImpedancemeasurementMembranes integrity, aging, discrimination betweenfresh and frozen-thawed products, pH, fat content,tendernessNon-destructive, inexpensive InvasiveMicrowavemeasurementWater activity, aging, sh freshness Potentially non-contactingX-raymeasurementFat content, myolaments structure changes Use of ionising radiationMagnetic resonanceNMR spectroscopyand imagingWater activity, water content, salt content,discrimination of Na ions population, water holdingcapacity, denaturation of connective tissue, pH, cookinglosses, fat content, fat organization, PSE, DFD, collagencontent, collagen organization, myober typingAccuracy, 3D reconstruction ExpensiveMagneticresonanceelastographyLocal viscoelastic properties 3D reconstruction of mechanical properties ExpensiveaDebated.134 J.-L. Damez, S. Clerjon/ Meat Science 80 (2008) 1321492. Mechanical methodsMechanical methodsforassessingtextural sensoryattributeshavebeenwidelyusedsincethethirties. Theyincludeinvasivemethods, suchascompression, tractionandshearing, whichre-quire sampling, and non-invasive methods, such as direct or reso-nance tests that can be performed on intact muscles.2.1. Instrument measurementsTheWarnerBratzlershearforce(WBSF)testusesaWarnerBratzler apparatus tomeasuremaximumshear force. ClassicalWBSF measurements are widely used, but the results usually showtenderness discrepancies when estimated by a trained sensory pa-nel (TSP) or other objective measurements (Lepetit & Culioli, 1994;Shackelford, Wheeler, &Koohmaraie, 1995; Timmet al., 2003).These discrepancies stem from the orientation of the probe, withmeasurements taken in the parallel orientation (along the lengthof the muscle) being more consistent in predicting tenderness thanmeasurementsin theperpendicular orientation. Moreover, WBSFmeasurementsdiffer betweenrawandcookedmeat (Tornberg,1996). Despite (i) the precautions that have to be taken orientingthe measurement probe due to variations in muscle ber direction(Stephens et al., 2004) and (ii) the destructive and time-intensivenature of the method, WBSF remains the most widely used instru-ment technique for assessing meat toughness. Attempts have beenmadetoimproveor streamlineinstrument methods usingatblades (slice shear force, SSF) instead of the V-shaped blade usedwith WBSF. (Shackelford, Wheeler, & Koohmaraie, 1999), reportedthat SSFmeasurements aremorestronglycorrelated(r = .82)with TSP tenderness rating than WBSF (r = .77), and it has beenreportedthatSSFcanaccuratelyidentifytenderbeef(Wheeleretal., 2002). Ashighlightedby(Bouton, Ford, Harris, &Ratcliff,1975;Carroll, Thiessen, Rollins, &Powers, 1978;Sacks, Kronick,& Buechler, 1988), in compression tests on raw meat, as deforma-tionincreases, the three structural components myobrillarproteins, intramuscular connective tissue, and perimysiumsuccessivelyplayaroleinmechanicalresistance. Thispromptedthe development of complementary mechanical methods to assessmyobrillar proteins and intramuscular connective tissues by,respectively, measuringstrainat 20%and80%of meat sampledeformation(Lepetit&Culioli, 1994). Portableapparatuseshavebeendeveloped: the modiedArmor Tenderometer withsixsharpneedles, asdescribedbyTimmet al. (2003), whichgivespretty good results in predicting TSP toughness on raw meat (Ste-phens et al., 2004), the TorqueTenderometer from MIRINZ, NewZealand, andtheTendertec mechanical penetrometer fromtheAustralian Meat Research Corporation (Ferguson, 1993) whichgives a good correlation with meat toughness (Belk et al., 2001).2.2. Ultrasound methodsAnalyzingtheacousticparametersofwavespropagatinginamedium makes it possible to assess the characteristics of the prop-agation medium and to characterize it. Two methods using ultra-soundcanbeusedinfunctional qualityassessmentsof musclefood:ultrasonicspectral analysis(Abouekaram, Berge, &Culioli,1997) and ultrasonic elastography or transient elastography(Ophir, Miller, Ponnekanti, Cespedes, &Whittaker, 1994). Theacousticparameters takenintoaccount includethevelocityofthe propagating waves, and spectral parameters such as attenua-tionandbackscattercoefcientinthemedium. Ultrasonicwavepropagationinmeatdependsnotonlyonthecomposition(e.g.,waterandlipidcontent)butalsothestructure(e.g., orientationof muscle bers, organization of connective tissue). Some studiesdiscriminatingmusclesamplesintermsoffatandcollagencon-tents reported better results than those obtained by the mere ana-lyse of thechemical andmechanical properties (Abouelkaram,Laugier, Fink,& Culioli,1992; Abouelkaram etal.,2000; Morlein,Rosner, Brand, Jenderka, & Wicke, 2005). Fat content has been re-portedtobecorrelatedwithultrasoundpropagationspeed, withfat and lean showing reverse temperature dependencies on soundvelocity(Abouelkarametal., 2000;Benedito, Carcel, Rossello, &Mulet, 2001). AsreportedbyMonin(1998), ultrasonicmeasure-ments give a good prediction of meat texture on live animals andwholecarcass, while at the same time beinginexpensive andnon-invasive.Biological tissues behave as viscoelastic materials, i.e., theypresent both uid viscosity properties and solid elasticity proper-ties. Sinceacousticwavepropagationisdirectlylinkedtothesemechanical properties, following the tissue propagation of acousticwaves could be a solution for measuring local viscoelastic proper-ties. Thiscanbedonebymeansofanechographicsystem, viaatechnique called transient elastography.Complementarywithultrasoundanalysis, transientelastogra-phy is a novel and non-invasive technique for evaluating the localmechanical properties of biological tissues. It consists in an ultra-sonic transducer, which is applied at the surface of a biological tis-sue, coupled with an ultrasonic pulse echo system. The basic idea isthat the low frequency vibrations generated by the transducer it-self inducealowfrequencymotionof thescatterersinsidethemedia. This motion can be detected and measured with the con-ventional echographic system, and via an inverse problem resolu-tionthismotioncanyieldthelocal viscoelasticproperties. Thesame technique is applied in the evaluation of brosis in chronicliver disease.Recent reportshavedescribedthat transient elastographyisable to work in anisotropic media like muscle (Gennisson, Cathe-line, Chaffai, & Fink, 2003; Gennisson, Cornu, Catheline, Fink, & Por-tero, 2005; McAleavey, Nightingale, Stutz, Hsu, &Trahey, 2003)thanks to the polarization of the low frequency shear strain waves.Sabra, Conti, Roux, andKuperman(2007)workedonalow-costtransient elastographytechnique for monitoringbiomechanicalmusclepropertiesinvivo, anapproachthat couldbeuseful forindustrial prototyping.In the eld of meat science, the same team has transposed thistechnique to beef meat structure evaluation(Catheline et al.,2004). Apostrigorbicepsfemorisbeefmusclewassubmittedtoan acoustic wave from 50 to 350 Hz. Local displacements resultingfromthis mechanical excitationwereaccessedwithaclassicalmedical transducerarray. Theselocal displacements(speedandattenuation)arethenusedtocomputeviscoelasticpropertiesinaccordance with the mechanical Voigts model. (Berg et al., 1999)assessed pork quality with this same technique.A similar approach is used in combining NMR and mechanicallowfrequencyshearwavesinmagneticresonanceelastography(see paragraph 6.3.). Dutt et al. (2000) have published a compari-son of ultrasound elastography and magnetic resonanceelastography.3. Optical methods3.1. Spectroscopic methodsWavespropagateandthestudyof radiation, absorptionandmore generally of any interactions between electromagnetic radia-tionandmatteriscalledspectroscopy. Fromlowfrequenciestohigh frequencies, optical spectroscopy covers near-infrared (NIR),infrared (IR), visible, and ultra-violet (UV) (including uorescence).SpectroscopicmethodsarewidelyusedformusclefoodqualityJ.-L. Damez, S. Clerjon/ Meat Science 80 (2008) 132149 135assessment andcontrol, inbothlaboratoryandmeat industrysettings (Hildrum, Wold, Vegard, Renou, & Dufour, 2006). Opticalspectroscopy offers a panel of useful techniques for on-linecharacterization because of its non-contacting possibilities and be-cause of the bre-optical components which make it easy to designportabledevices. Ithasbeenwidelyinvestigatedintheeldofmeat science as a means of gaining structural information.Polarized light gives additional organizational data and are there-fore oftenusedfor these applications. Belowis anoutline ofthe main methods developed and used in meat science and some-times in the biomedical eld, where tissue organization is also ofinterest.3.1.1. Infrared spectroscopyInfrared (IR) spectroscopy is a spectroscopic method that dealswiththeinfraredregionof theelectromagneticspectrum(fromabout 800 to 2500 nm). It is typically employed in pharmaceuticalapplications, medical diagnostics, foodandagrochemical qualitycontrol, andcombustionresearch. Infraredspectroscopyisbasedon the principle that the chemical bonds in organic molecules ab-sorb or emit infrared light when their vibrational state changes. Inthe near infrared area spectrum, there are major changes in vibra-tional state. Amajorchallengeinmeat scienceapplicationsfornearinfraredspectroscopyissamplepresentation. Transmissionis the most powerful method well-suited to liquids and gases butis inappropriate for undiluted solids. Reection spectroscopy offersthealternativethat isalmost alwaysusedinmeat andmusclestudies, andithasbeenwidelyinvestigatedasameansofindi-rectlymeasuringmeatstructure. Indeed, althoughinfraredspec-troscopy gives direct molecular-level information, research showsthatitcanbesuccessfullyusedtodeterminemacroscopicstruc-tural changes associated with meat or muscle structure.New developments in IR spectroscopy will expand its applica-tions further. These include hand-held bre-optic-enabled instru-mentsabletomakeinstantaneousmeasurementsinalmostanypart of a product. IR spectroscopy, with its speed, ease of use andversatility, may well become one of the most powerful analyticaltechniquesavailabletonalmeatproductevaluation(vanKem-pen, 2001).Fourier transform infrared (FT-IR) spectroscopy is a fairly newtechnique for collecting infrared spectra. Instead of recording theamountofenergyabsorbedwhentheinfraredlightfrequencyisscanned (monochromator), the IR light is guided through an inter-ferometer. After passing through the sample, the measured signalistheinterferogram, atime-domainsignal. Performingamathe-matical Fourier transform on this signal results in a spectrum iden-tical to that from conventional (dispersive) infrared spectroscopy,and measuring a single spectrum is faster. Due to the advantagesoffered, virtually all modern infrared spectrometers are FT-IRinstruments.Near infrared spectroscopy, often extended to the visible region,is under investigation in several laboratories with the aim of eval-uating itspotential use for meat structure control. Most of theselaboratories are working on eating meat quality, focusing oninstrumentaltexture, sensorytenderness, pastiness, crustingandjuiciness. For instance, Andres et al. (2007) are working sorting ex-tremesamplesoflambmeatintoahigh-qualityclassintermoftendernessandjuiciness. Thisresultmayhavepracticalimplica-tionsforsortingmeatintoahighqualityclass, whichcouldbebranded and sold at a higher price. Ellekjaer, Isaksson, and Solheim(1994) have published results on sausage quality. Fiber-opticprobes have been used to predict pastiness or crusting of dry-curedham (Garcia-Rey, Garcia-Olmo, De Pedro, Quiles-Zafra, & de Castro,2005; Ortiz, Sarabia, Garcia-Rey, & de Castro, 2006), and beef ten-derness has also been investigated: Liu et al. (2003) classied ten-der and tough samples correctly to 83%, Shackelford, Wheeler, andKoohmaraie(2005) performedon-linetendernessclassicationsonUSbeef carcasses, whileByrne, Downey, Troy, andBuckley(1998) Park, Chen, Hruschka, Shackelford, and Koohmaraie(1998) worked on the prediction of WBSF and sensory tenderness.Meullenet, Jonville, Grezes, andOwens(2004) showedthat NIRspectroscopy could be used to predict the instrumental texture ofcooked poultry meat and to classify muscles according to tender-ness levels.IRspectroscopyhasalsomadeitpossibletodiscriminatebe-tween fresh and frozen-thawed products from broiler breast meat(Lyon, Windham, Lyon, & Barton, 2001) or sh (Uddin et al., 2005)which is of high interest for the control of fraudulent frozing-thaw-ing cycle.The evaluation of thepork water holding capacity (WHC) isamorecontroversial topic. While PedersenandEngelsen(2001)present encouraging results, Hoving-Bolink et al. (2005) are moreprudent about the ability of IR spectroscopy to predict this default.Ithasbeendemonstratedthatmicrostructurechangesinsaltedpork can be detected by FT-IR microspectroscopy (Bocker, Ofstad,Bertram, Egelandsdal, &Kohler, 2006). Uddin, Okazaki, Ahmad,Fukuda, and Tanaka (2006) also investigated protein denaturationand changes in water state in shmeat gels while heating with IRspectroscopy. To our knowledge, Swatland is the only author whousedpolarizedIRlighttoaccessstructuralinformationonmeat.The polarization gives additional information on sample organiza-tion that can be exploited to detect cold shortening in pork (Swat-land, 1995) and beef (Swatland, 1996). Swatland and Barbut (1995)also showed that myobrillar near-IR birefringence in turkey meatiscorrelatedwithWHCof rawmeatandwithuidlossduringcooking.3.1.2. Raman spectroscopyRamanspectroscopyisalsoavibrational spectroscopictech-nique used in condensed matter physics, biomedical applicationsand chemistry to study vibrational, rotational, and other low-fre-quency modes in a system. It relies on inelastic scattering of mono-chromatic light, usually from a laser in the visible, IR, or near-UVspectra. It gives similar but complementary informationto IRspectroscopy.Raman spectroscopy has great potential for biochemical tissueanalysisat boththemacroscopicandmicroscopicscale. Oneofthe great advantages of this technique is its ability to provide infor-mation on the concentration, structure and interaction of biochem-ical molecules in their microenvironments within intact cells andtissues (i.e., insitu), non-destructively, andwithout homogeniza-tion, extraction, or the use of dyes, labels, or other contrast-enhancingagents. Furthermore, Ramanspectroscopycanbeper-formed in vivo using optical ber technology.In a recent review, Herrero (2008) compared Raman spectros-copy to various conventional methodologies such as protein solu-bility, apparent viscosity, WHC, instrumental texture methods,dimethylamine content,peroxide values, and fatty acid composi-tion, all commonlyusedtodeterminequalityinshandmeatmuscletreatedunderdifferenthandling, processingandstorageconditionsthroughthechangesintheproteins, waterandlipidsof muscle food. It has been shown that Raman spectroscopy dataare related to the results obtained with these conventional qualitymethodsandcouldbeusedtoevaluatemusclefoodquality. Inaddition, it hasbeenshownthat Ramanspectroscopyprovidesstructural information on the changes of proteins (Li-Chan, Nakai,& Hirotsuka, 1994), water (Herrero, Carmona, Garcia, Solas, & Car-eche, 2005) andlipids (Beattie, Bell, Borgaard, Fearon, &Moss,2006) of muscle food that occur during the deterioration. Further-more, thisspectroscopytechniquehasseveral advantagescom-pared to traditional methods, since it is a direct and non-invasivetechnique that requires only small sample portions.136 J.-L. Damez, S. Clerjon/ Meat Science 80 (2008) 132149Before being investigated in food science, Raman spectroscopyhad been widely used in biomedical applications where it is wellknown for its ability to determine the degree of saturation in fattyacids, averysignicantnutritionalaspect. Areviewinthiseld(Manoharan, Wang, & Feld, 1996) outlined the main applicationsfor analysis of biological tissue, describing the advantages and dis-advantages of visible, N-IR and UV excitations and addressing theproblems and prospects of using these methodologies for diseasediagnosis. Still in the biomedical eld, Buschman et al. (2001) ex-plored cellular and extracellular morphologic structures, and Bren-nan, Wang, Dasari, and Feld (1997) developed a N-IRRamanspectrometer for in situ clinical investigation. This portable devicewas equipped with an optical bre for in vivo measurements.Marquardt (2001) has also reported on the development of anon-line probe designed and optimized for performing Raman mea-surements in both laboratory and industrial environments. Devel-oping this kind of Raman spectrometer in the biomedical eld haspaved the way to development in food science. Moreover, the com-ponents contributing to Raman scattering in muscle include colla-gen and elastin, both well known for their role in meat structure.At theproteinlevel, Beattie, Bell, Farmer, Moss, andDesmond(2004) demonstrate the ability of Raman spectroscopy to measurechanges in secondary structure and then to be useful for determin-ing textural properties such as beef meat tenderness. N-IR FT-Ra-manspectroscopywasalsousedtoinvestigateproteinstructurechangesinfoodduringheating(Beattieetal., 2004;Ozaki, Cho,Ikegaya, Muraishi, & Kawauchi, 1992).In a recent study, porcine muscle tissue was subjected to differ-ent processingfactors, includingageing, saltingandheat treat-ment, in order toinduce structural changes. These changeswerethen investigated withboth FT-IRand Raman microspectroscopyin order to compare the techniques. Bocker et al. (2007) concludedthatRamanisasstrongasIRspectroscopyindetectingmusclestructure changes. Still in the eld of pork processing, it has beenshown that Raman scattering is able to predict WHC early in freshmeat (Pedersen, Morel, Andersen, & Engelsen, 2003). Ellis, Broad-hurst, Clarke, andGoodacre(2005) proposedanapproachthatwouldaidfoodregulatorybodies torapidlyidentifymeat andpoultry products, highlighting the potential application of Ramanspectroscopyforrapidassessmentof foodadulterationanddis-criminationbetween both species and distinct muscle groupswithin these species.Again, polarizingtheexcitinglight canimprovetheperfor-mance of Raman spectroscopy in investigating structured materi-als. Smith and Berger (2005) performed polarized Ramanspectroscopyonatwo-layerdiffusingbiologicaltissue. Polarizedlight directly backscattering off from the supercial layer partiallyretains its sense of polarization, whereas deeper-probing light willbeincreasinglydepolarizedbydiffusion. Thistechniqueleadstodepth-dependent selectiveinformation, whichcanbeuseful incomplex foods studies.3.1.3. Visible spectroscopy and colorimetryThis eld of biophysical methods covers the visible spectra (of-tenextendedtonear-UVand/orN-IRregions)andtheCIEL*a*b*colour space as objective and non-destructive tools for tissuecharacterization.There has been a great deal of biomedical research into tissuecharacterizing using these techniques. As this research is often fo-cussed on gaining structural information, it merits coverage heredue to the potential for use in muscle and therefore meat - struc-ture evaluation.The use of visible light to access muscle structural informationis not a new concept. Rome (1967) utilized light scattering on iso-lated rabbit muscle to measure sarcomere length. Haskell, Carlson,and Blank (1989) worked on muscle birefringence to measure sar-comere length and characterize orientation disorders in the myol-amentarray. Birefringencemeasurementswereperformedusingpolarized light. Thebirefringenceofrabbit muscle wasalsousedto study and better understand the muscle cell biology of the rigortorelaxationphase(Taylor, 1976). Morerecently, Xia, Weaver,Gerrard, and Yao (2006) reported using light scattering to evaluatesarcomere structure changes in whole muscle. Visible spectroscopyisoftenperformedwithpolarizedlight toprovidequantitativemorphologicaldataonstructuralchanges(Sokolov, Drezek, Gos-sage, &Richards-Kortum, 1999). Forinstance, malignanttissuesare less organizedthanothers, andanisotropic scattering andabsorptionparameterscanbeexploitedtodetectthem(Ghosh,Mohanty, Majumder, & Gupta, 2001; Kim et al., 2003). Refocusingon muscle studies, Binzoni et al. (2006) worked on anisotropic pho-ton migration in human skeletal muscle as a tool to access infor-mation on bre organization.Intheeldof meatscience, earlydetectionof pale, softandexsudative(PSE)meatisamajorpotentialapplicationofvisiblespectroscopyandcolorimetryforbothpork(Chizzolini, Novelli,Badiani, Rosa, & Delbono, 1993; Swatland, 1997a; Swatland & Irie,1992; Xing, Ngadi, Gunenc, Prasher, & Gariepy, 2007) and poultrymeat(Barbut, 1993;Marquez, Wang, Lin, Schwartz, &Thomsen,1998; Sante, Lebert, Le Pottier, & Ouali, 1996).Valkova, Salakova, Buchtova, and Tremlova (2007) showed theabilityof theCIEL*a*b*systemtopredictsensorytendernessincooked pork ham. Liu and Chen (2001) studied changes in visiblespectra versus post mortem chicken meat degradation. Visible spec-troscopywithpolarizedlightcanbeusedtopredicttoughness,WHCanddriploss insaltedcomminutedchickenbreast meat(Swatland & Barbut, 1999).There have also been studies on beef meat, such as to measuresarcomere length and collagen content (Xia, Berg, Lee, & Yao, 2007)ortoselectanimalswithlow-temperaturegelatinizationofcon-nective tissue (Swatland, 2006), based on the fact that connectivetissue is closely related to beef meat tenderness. Lastly, the Euro-peanFAIRproject(MUSTEC)investigatedthesemethodsforon-line predicting of sh freshness (Olafsdottir et al., 2004).3.1.4. Fluorescence spectroscopyFluorescencespectroscopyisatypeof electromagneticspec-troscopywhichanalyzesuorescencefromasample. Itinvolvesusing a beam of light, usually UV light, that excites the electronsin molecules of certain compounds and causes them to emit a low-er-energy light. In uorescence spectroscopy, the species is rst ex-cited, byabsorbingaphotonoflight, fromitsgroundelectronicstatetooneofthevariousvibrationalstatesintheexcitedelec-tronic state. Collisions with other molecules cause the excited mol-ecule to lose vibrational energy until it reaches the lowestvibrational state of the excited electronic state. The molecule thendropsbackdowntooneof thevariousvibrational levelsof itsground electronic state, emitting a photon in the process. As mol-ecules can drop down into any of several vibrational levels in thegroundstate, theemittedphotonswill havedifferent energies,andthusfrequencies. Therefore, analyzingthedifferentfrequen-cies of light emitted in uorescent spectroscopy, along with theirrelativeintensities, makesitpossibletodeterminethestructureof the different vibrational levels.Tryptophan is an important intrinsic uorescent probe that canbe used to assess the nature of the tryptophan microenvironment.Proteins that lack tryptophan can be attached to an extrinsic uo-rophore probe.For opaque samples such as meat and meat products, front faceuorescenceisused, andbecausetheseproductscontaintrypto-phan, thistechniquehasbeenusedinmeatandmusclescienceto investigate sample structure without using extrinsic uorophoreprobes.J.-L. Damez, S. Clerjon/ Meat Science 80 (2008) 132149 137It is well known that the amount of connective tissue in meat isdirectly linked to meat tenderness (Light, Champion, Voyle, & Bai-ley, 1985). Since connective tissue is the most intrinsic uorophorein meat, its uorescence intensity should be a good marker of meattenderness.Since as far back as 1987, Swatland has been published researchonthepotentialityofautouorescenceforgaugingmeatquality.He has correlated connective tissue uorescence with meat qualityparameters like palatability (Swatland, Gullett, Hore, & Buttenham,1995) or chewiness (Swatland, Nielsen, & Andersen, 1995) and hasalsodevelopedon-lineprobescombininguorescencemeasure-ments with other optical measurements like reectance (Swatland,2000). In 1993, Swatland (1993) highlighted the necessity of con-trolling temperature when connective tissue uorescence is usedtodetect toughmeat. Front faceuorescencespectroscopywasalsoinvestigatedtomeasuretextureof meat emulsions(Allais,Viaud, Pierre, &Dufour, 2004), shfreshness(Andersen&Wold,2003; Dufour, Francia, & Kane, 2003) and meat tenderness(Egelandsdal, Wold, Sponnich, Neegard, &Hildrum, 2002), while(Christensen, Norgaard, Bro, & Engelsen, 2006) reviewed autouo-rescence in food and particularly in meat and sh products.All previousarticlestreatnonpolarizedlightapproaches. Asstructure is often connected with preferential alignments, the useof polarized light in uorescence spectroscopy is an improvementforstructurestudy. Polarizationallowspreferentialexcitationofuorophorewithtransmissionmomentsparalleltothedirectionof polarization. Several authors have used this property. The stud-iesonpolarizationof tryptophanuorescencebeganinthe60swith (Aronson & Morales, 1969) and was used in biological tissuesstudies in the biomedical eld (Borejdo et al., 2004), in particular todetectmalignanttumours(Ghoshetal., 2001;Mohanty, Ghosh,Majumder, & Gupta, 2001). Moreover, proteins can be labeled withuorescentdyesformusclestudyinpolarizedlight(Daleetal.,1999; Van Der Heide, Orbons, Gerritsen, & Levine, 1992).In the eld of food science, Marangoni (1992) used polarizationuorescence spectroscopy to determine microviscosity and struc-tural order in complex lipid systems. More recently, Yao, Liu, andHsieh(2004)investigateduorescencepolarizationspectroscopyfor characterizing ber formation in meat analogues.The INRA has investigated the uorescence of tryptophan andconnective tissues in meat with polarized excitation and emissionbeams to evaluate the potential use of this tool for meat structurecontrol (Luc, 2007; Luc, Clerjon, Peyrin, & Lepetit, 2008). For a well-ordered biological tissue, uorescence is anisotropic and thisanisotropytendstodisappearwithstructuraldegradation. Fig. 1presents thedecreaseof uorescenceanisotropywithcollagenheating. Destructuring processes like ageing, grounding or heatinghavebeensuccessfullyinvestigated. Thetechniquealsoappearsusefulfordetectingcoldshorteningbovinemuscle(Luc, Clerjon,Peyrin, Lepetit, & Culioli, 2008).3.2. Imaging3.2.1. Microscopic imagingMicroscopyhasbeenwidelyusedtocontrol meat andmeatproduct structure. It can be split into two elds: optical micros-copy and electron microscopy.3.2.2. Optical microscopyOptical microscopy offers the simplest way to obtain magniedimages of biological tissues. This eld covers a large range of tech-niquesthat havebeenusedforyearstocharacterizemeat andmeat product structures. Techniques can be classed simplydepending on whether samples must be prepared in thin cuts ornot. Non-thin-cuts samples were used for the very early phase con-trast measurement (Ranvier, 1889) allowing the detection of A andI bands in muscle and for the new confocal laser scanning micros-copy, which will be discussed later.3.2.3. HistologyHistology is a widely used method for observing biological tis-sues at the microscopic level, particularly as a tool for controllingmeattextureinfoodscience. Thetechniquemayormaynotre-quire tissue staining with specic dyes. However, histology alwaysneeds very thin sample cuts. In biological microscopy, it is almostalwaysnecessarytoenhancecontrast byusingspecicdyestomake certain biological components more visible during histolog-ical observation.This techniquecanbeusedtoenhancelipids: for instance,Thakur, Morioka, Itoh, andObatake(2003)workedontheeffectof compositionanddepositionof lipidsstainedwithSudandyeonshmeattexture. Collagenarchitecturecanalsobeaccessed,either for ber typing inpork withmyosinATPase or NADHstaining(Oshimaetal., 2007), orfordetectingchangesinstruc-ture withfrozenprocess incarpmyobrillar proteins stainedwith SDH enzyme or myofobrillar ATPase (Jasra, Jasra, & Talesara,2001).Severalstudieshaveconductedtocorrelatepostmortempro-cesseswithhistologicalmusclebertraitsinmeat. Forinstance,Ichinoseki, Nishiumi, and Suzuki (2006) studied the effect of highpressure on intramuscular collagen brils in bovine connective tis-sue, while Rusman, Gerelt, Yamamoto, Nishiumi, and Suzuki(2007) reported on the effects of high pressure and heat on histo-logical characteristics of bovinemuscle, suchas inter-myoberspace, Zlinedegradation, sarcomerelengthdecreaseandendo-mysium structure. The cold shortening and structural characteris-ticsofsingleporkmusclebers(Willems&Purslow, 1996)andstructural characterization of mechanically recovered meat(Tremlova, Sarha, Pospiech, Buchtova, &Randulova, 2006) havealsobeeninvestigated. Vonlengerken, Maak, Wicke, Fiedler, andintact20'60'090180270Fig. 1. Angular plot of bovine collagen intrinsic uorescence anisotropy (arbitraryunits). Angular coordinates are relative to the angle between collagen bres maindirection and exciting light polarisation direction. Collagen is heated at 60 C during0 min (intact sample), 20 and 60 min. Destructuration with heating is pointed outbythedisappearanceof angular dependenceof anisotropy: after 60of heating,anisotropy tends to a constant value whatever the exiting light polarisationdirection (Luc, 2007).138 J.-L. Damez, S. Clerjon/ Meat Science 80 (2008) 132149Ender (1994) have worked on myobers structural and functionaltraits (type, diameter and number of bers) in muscle for geneticimprovementof pork. Fiberdisorganization, bermisalignment,and increase in bers spacing (Fig. 2) have been analyzed to char-acterize PSE zones in pigs with hematoxylineosinsafran staining(Laville et al., 2005).Karlsson, Klont, and Fernandez (1999) reviewed skeletal musclebres as factors for pork quality. The histochemical properties of amuscle, suchasbretypecomposition, brearea, oxidativeandglycolytic capacities,andglycogen andlipid contents,arefactorsthathavebeenfoundtoinuencemeatquality. Similarly, histo-chemistry, combinedwithseveral otherstainingprotocolssuchas Sudan black-B, myosin ATPase or NADH-tetrazolium reductase,has been used to classify bres.Immunohistochemical labelling is more selective because anti-bodiesareusedtovisualizespecicproteins andlipids:thesec-ondary antibody is labelled withan enzyme or a uorescentcomponent. Severalapplicationsofimmunoenzymologyhavefo-cused on collagen architecture and collagen typing in pork(Nakamuraet al., 2003) andpoultry(Oshimaet al., 2007; Royet al., 2006). Astruc, Marinova, Labas, Gatellier, and Sante-Lhoutel-lier(2007)localizedoxidizedproteinsinmuscletopinpointtheroleof membraneproteinsinoxidation. Theoccurrenceof fastandslowmyosinisoforms inbretypes was alsodetectedbyimmunohistochemistryinpigs(Fiedler, Dietl, Rehfeldt, Wegner,& Ender, 2004; Fiedler et al., 1999).Whenantibodies arelabelledwithuorescent dyes, weareworking with immunouorescence, a common technique for visu-alizing sub-cellular distribution of biomolecules of interest. Immu-nouorescent-labelled tissue sections are observed using auorescence microscope orby confocal microscopy,as explainedfurtherdown. ImmunouorescencehasbeenusedtostudyPSEin turkey breast muscle (Pietrzak, Greaser, & Sosnicki, 1997)andmeasure thin muscle lament lengths of beef, rabbit, and chickenmyobrils (Ringkob, Swartz, & Greaser, 2004). However, histologi-cal techniques are often used in combination to obtain a maximumof information from a given tissue.3.2.4. Confocal laser scanning microscopyConfocal laser scanning microscopy is a uorescence techniquefor obtaining high-longitudinal resolution optical images. Itis anevolution of the more traditional uorescence microscopy, its keyfeature being the ability to produce point-by-point in-focus imagesof thickspecimens, allowing3Dreconstructionsof complextis-sues. Becausethis techniquedepends onuorescence, samplesusually need to be treated with uorescent dyes to make objectsvisible, but contrary to the histological techniques, there is no needfor thin cuts.Inmeat science, Straadt, Rasmussen, Andersen, andBertram(2007) recently appliedconfocal laser scanning microscopy tomonitor changes in fresh and cooked pork muscle during ageing.Two different magnications, i.e., at 10 and 200, give spectacu-lar views of myobers and myolaments, respectively.Nakamura et al. (2007) successfully studied changes in bovineconnective tissue according to animals feeding (concentrate- androughage-fedgroups) usingconfocal laser scanningmicroscopycoupled with immunohistochemical typing of collagen and proteinstructure evaluation in perimysium and endomysium. Immunohis-tochemical/confocal laser-scanning microscopy is a useful tool forstudying structural relationships among connective tissue compo-nents in skeletal muscle.3.2.5. Electron microscopyThe use of electrons beam to illuminate a specimen and createanenlargedimageleadstoimageobservationthathasamuchgreater resolving power than with optical microscopes. The greaterresolutionandmagnicationofelectronmicroscopystemsfromthe electron wavelength which is much smaller than light photonwavelength.Biological investigationsusescanning(reectionmethod) ortransmissionelectronmicroscopyaccordingtothe application.We can also cite here the coupling of an X-ray probe to the elec-tronic microscope to perform X-microanalysis, i.e., the local mea-surement, at microscopic level, of the X-ray spectrum.3.2.6. Scanning electron microscopyScanningelectronmicroscopy(SEM)givesimageswithgreatdepth-of-eldyieldingacharacteristic3Ddisplaythat providesgreaterinsightintothesurfacestructureof abiological sample.For SEM, samples require preparation, such as cryoxation, dehy-dration, embedding (in resin. . .), or staining (with heavy metal).SEMisahigh-performancetool for investigatingprocess-re-latedchangesinmeatultrastructure. Incombinationwithhisto-logical analyses, Tornberg(2005) reviews effects of heatingonchanges in secondary, tertiary and quaternary structure of proteinsand then on cooked meat quality. This paper shows how SEM is apowerful tool tobetter understandrelations betweenproteinsstructure andmeat quality. Palka andDaun(1999) have alsoworkedonstructural changesduringheatinginbovinemuscle.Structural changes in intramuscular connective tissue during ten-derization of bovine meat by marinating in a solution containingproteolytic enzyme (Chen, He, Jiao, & Ni, 2006) is also of interestin meat process control and can be accessed with SEM.Larrea et al. (2007) outlined the comparisons and complemen-taritiesof SEM, cryo-SEMandtransmissionelectronmicroscopyfor themeasurement of process-relatedultrastructural changesin ham. Lastly, Yang and Froning (1992) reported structural differ-ences in washed or unwashed mechanically deboned chickenmeat, whileNishimura, Hattori, andTakahashi (1999) focusingon the ante-mortem stage, studied structural changes in intramus-cular connective tissue during fattening.Cryo-scanning electron microscopy consists in SEM observationafter cryoxation, and is well adapted to biological tissues that arerelatively unaffected by this specic sample preparation. More spe-cically, cryoxationisasolutionforvisualizingtheelectrolytesthataresuppressedinaclassical dehydrationpreparation. Gar-cia-Segovia, Andres-Bello, and Martinez-Monzo (2007) studiedthe effects of cooking temperature and cooking time on losses, col-our and texture of beef steaks using cryo-SEM to assess the endo-mysium and perimysium microstructure. Pig muscle cellultrastructure versus freezing rate and storage time has also beenFig. 2. Observation of increase in ber spacing to characterize PSE zones in pig withhematoxyline eosine saffran staining. (Laville et al., 2005).J.-L. Damez, S. Clerjon/ Meat Science 80 (2008) 132149 139investigated with this technique (Ngapo, Babare, Reynolds, & Maw-son, 1999). Cryoxationis alsousedintransmissionelectronicmicroscopy.Environmental scanningelectron microscopy (ESEM)isanewdevelopmentintheeldofelectron microscopy. Itopens upthepossibilityof observingsamples at almost normal atmosphericpressures(unlikeclassical SEM)withouthavingtodehydrateorfreeze them. YarmandandBaumgartner (2000) usedESEMtostudy the structure of semimembranosus veal muscle. Even thoughESEMofferpromisingpossibilitiesforobservingintact samples,contrast is often less good than in traditional SEM when the sampleis stained with a heavy metal.3.2.7. Transmission electron microscopyIn transmission electron microscopy (TEM), electrons arepassed through the sample. Resolution is higher than in SEM andthesamplecanbestainedwithheavymetalstoimproveimagequality. Unfortunately, sampleshavetobepreparedinverythinslices and put on a grid for the observation, making the techniquedifcult to implement.This technique has been used to observe Z line removal duringculledcowmeat tenderizationbyproteolytic enzymes (Gerelt,Ikeuchi, & Suzuki, 2000) and by calcium chloride (Gerelt, Ikeuchi,Nishiumi, & Suzuki, 2002) following osmotic dehydration.Ho, Stromer, and Robson (1996) used TEMto study the effects ofelectrical stimulation of bovine carcasses on post mortem change inskeletal muscles. Nakamura, Ando, Seoka, Kawasaki, andTsuka-masa(2006) usedthesametechniquetostudychangesintheultrastructural propertiesof tunamuscleduringchilledstorage.Sen and Sharma (2004) also used TEM to show that freezing/thaw-ing cycles do not signicantly change the ultrastructural propertiesof buffalo muscle if freezing is applied in good conditions.To round off, TEM has also been used to quantify ageing-relatedultrastructural changes, I band breaks and sarcolemma attachmenttomyobrils, ingamemeat (Taylor, Fjaera, &Skjervold, 2002).Even when aged for a long time (14 days), game meat presents veryfew I band breaks. These results are not in agreement with gamemeathightendernesswhichseemstobeattributedtomyobersmall size better thanto ultrastructural post-mortemchanges.Myober attachment in salmon llets (Taylor, Labas, Smulders, &Wiklund, 2002) was alsoinvestigatedwithTEM. Authors havedemonstratedthatthestructural changeassociatedwithlossofmusclehardnessisbreaksinmyober-to-myoberattachments,and that loss of rigor stiffness is associated with breaks in myo-ber-to-myocommataattachments. Theseresultsimplythat ulti-mate llet texture occurs by a combination of these distinctstructural changes.As for histological methods, immunoelectronmicroscopyal-lows the detection of specic proteins in ultra-thin tissue sections.This technique is actually a TEM technique conjugated with spe-cicantibodieslabelledwithheavymetal particles(oftengold).An original application studying ageing and structural weakeningof myobrils and collagen is given in Takahashi (1996).3.2.8. Macroscopic imagingVisual inspection is used extensively for quality assessments onmeatproductsappliedtoprocessesrunningfrominitialgradingthrough to consumer purchases. Numerous laboratories haveinvestigatedthe possibilityof usingimage-basedmeat qualityevaluation, and a number of reviews have listed the main applica-tions for food and more specically meat science (see Brosnan &Sun, 2004; Du & Sun, 2004). In meat science, image analysis con-sistsinanalyzingthetextureof imagesproducedfrommuscleandmeat sectionsat oneor morewavelengths. Thetechniquemakes it possible to clearly highlight the collagen and lipid struc-tures of muscular tissue.Kirschner, Ofstad, Skarpeid, Host, and Kohler (2004) used FT-IRimaging(associatedwithFT-IRmicroscopy) tomonitorthermaldenaturation processes in aged beef loin. Recently, the same teamdevelopedanalgorithmfor analyzingsets of FT-IRmicroscopyimages of tissue sections, which was applied to FT-IR microscopyimagesofbeefloinmusclescontainingmyoberandconnectivetissueregions. Theimageswereinvestigatedfor variationsdueto ageing duration and due to homogeneity in the connective tis-sue regions. Keeping with beef quality, it has been shown (Shac-kelford, Wheeler, &Koohmaraie, 1998)thatearlyimageanalysisof a steak from the 12th rib region can predict tenderness classi-cation. Shiranita, Miyajima, andTakiyama(1998) describedamethod for determining meat quality using the concepts of mar-blingscore andtextureanalysis. Marblingscore, whichwasameasurement of fat distributiondensityintheribeyeregion,was consideredas atexturepattern. Textureanalysis has alsobeenusedtoclassifyphotographic images of beef meat slices(Basset, Buquet, Abouelkaram, Delachartre, & Culioli, 2000).Among the multiple muscular tissue characteristics that inuencemeatquality, connectivetissuecontentandspatial distribution,whichdenemeat grain, areparticularlyimportant sincetheyare directly related to tenderness. Connective tissue containstwo key components, namely fat and collagen, which vary accord-ing tomuscle, breed,andage. Thesecomponents are clearly vis-ible on photographic images. Finally, image processing techniqueshave been developed to predict cooked-beef tenderness fromfresh-beef image characteristics (Li, Tan, Martz, & Heymann,1999). Anotherrecentpaper(Chandraratne, Samarasinghe, Kula-siri, & Bickerstaffe, 2006) investigated how the surface character-istics (geometric and texture) or rawmeat could be used topredictthetendernessofthecookedlambmeat. Thepredictionshowedencouragingresults indicatingthat there is signicantrelation between the raw meat surface and the cooked meat ten-dernessusingnon-linearandarticial neural networksanalyses(R2= 0.602and0.746, respectively). Ithasalsobeenpossibletoanalyzeporkqualityandmarblinglevel usinga hyperspectral(4001000 nm) imagingsystem(Qiao, Ngadi, Wang, Gariepy, &Prasher, 2007). Hyperspectral analyses provide supplementaryinformationthanclassicalvisibleimagery, andaneuralnetworkmodel basedononly10principal components extractedfromthehyperspectrumof rawmeat canclassifymeat accordingtoitsrmnessandexudation. Inspiteof thesepromisingresults,authorsdonotsucceedinprovingthat spectrumwideningim-provesclassicationresults.Meat analogues produced from vegetable proteins have gener-ated the need for a new imaging system and reliable non-destruc-tivetechniquesfordeterminingthetextural propertiesof thesenew products (Ranasinghesagara, Hsieh, & Yao, 2005). Image anal-ysiscouldalsobeusedtodeterminebreorientationinmeatproducts.The fact that meat structure is closely linked with muscle breparameters promptedthe INRAtodevelopdedicatedsoftware(Buche & Mauron, 1997) to determine bre structural parameters(area, shape factors, etc.) in pork, chicken, trout, beef, rabbit, mut-tonandturkeymuscles. Imaginganalysisofmagneticresonanceandhistologicalimagescangeneratepreciseandobjectivemea-surements of the two main structural component of muscle tissue(myobersandintramuscular connectivetissue) (Sifre-Maunier,Taylor, Berge, Culioli, & Bonny, 2006).Multispectral image analysis (MIA) on images acquired at dif-ferent wavelengths in the UVVisible range was used to correlatemeatcomponentswithsensorial andphysical properties, givinggood results in terms of predicting composition (collagen(R2= .88), lipids (R2= .87)) and sensory attributes (tenderness(R2= .75), juiciness (R2= .84)) (Abouelkaram, Chauvet, Strydom,Bertrand, & Damez, 2006).140 J.-L. Damez, S. Clerjon/ Meat Science 80 (2008) 1321494. Dielectric methods4.1. Impedance measurementThe rst research involving impedance measurements on meatwaspublishedbackinthe1930s(Callow, 1936). Electricimped-ance is the property of a material to oppose the ow of electric cur-rent. When this property is not dependent on the frequency of thecurrent, it is qualied as resistance; otherwise, as it is the case withbiological tissues, the impedance has a resistive component and acapacitivecomponent. Inschematicterms, biological tissuesarecomposed of cells that are surrounded by an extracellular liquid.The cell membrane acts as an insulator at low frequencies, behav-ing like a capacitor. Biological tissue, particularly meat, has aniso-tropic impedance, i.e., impedance varies according to whether thecurrent runs parallel or perpendicular tomuscleber (Damez,Clerjon, Abouelkaram, & Lepetit, 2007; Swatland, 1980). Theimpedance of the meat decreases quickly with rigor and continuestodecrease, albeit muchmoreslowly, duringstorage(Pliquett,Pliquett, Schoberlein, & Freywald, 1995).Electricimpedanceisusedfor abroadrangeof purposesinmeat technology:4.1.1. Detection of frozen meatsIn the 1970s, it was shown that frozen meat samples have veryweakimpedance(Sale, 1974). However, wenowknowthatthismeasurement cannot certify that a meat has been frozen, becausemeat aged for very long can present similarly low impedance (Da-mez, Clerjon, Abouelkaram, & Lepetit, 2008).4.1.2. pHSince the 1980s, the vast majority of research on impedance hasinvolved using this measurement for controlling the drop in pH orfor evaluating ultimate pH. It mainly concerns pig meat (Swatland,1985) although beef has also been studied (Byrne, Troy, & Buckely,2000). In the case of pork, one of the major problems is the evalu-ation of water holding capacity (Schafer, Rosenvold, Purslow,Andersen, & Henckel, 2002) and the detection of pale soft exuda-tive (PSE) meat, which has a low pH and is highly exudative, andthus unsuitable for processing. In the case of beef, the problem isthedarkrmdry(DFD) meatwithhighpH. Thesetwodefectsare associated with modications of membranes and extracellularuids, which therefore affect the meats electrical properties. Elec-trical measurements have been used to compensate the inaccuracyof pH measurements. The majority of the studies in this eld havefocusedondetectingthedefectsearly, i.e., within45 minto1 hpost-slaughter. However, recent results show that electrical mea-surements do not permit the early detectionof DFD(Forrestet al., 2000; Guerrero et al., 2004). The difculty in detecting PSEmeats during rigor set in is due to the fact that meat evolves rapidly(pH, temperature) during this period whereas the associated met-abolicchangeswill onlyaffectitsstructureandthusitselectricpropertieslateron. However, impedance(conductivity) ismorecapableofdetectingPSEmeatsoncetheultimatepHisreached(Guerrero et al., 2004).4.1.3. Fat contentMany studies conducted since the 1980s, have attempted to useelectrical properties to estimate fat content in animal carcasses ormeat. Fat isanelectrical insulator andthereforeinuencestheimpedance of tissues. Electric impedance methods can obtainremarkableresults. Asimpleelectricconductivitymeasurementonacarcassimmediatelyafterslaughtercanbeassociatedwithanatomical data togive fat content withremarkable accuracy(R2= .95). Thiscouldbeexplainedbythefactthattherearenomembraneorextracellularcompartmentmodicationsoccurringimmediatelyafterslaughter, andthemeasurementsaremadeata stable temperature. Apatented systemhas been developed(Madsen, Rasmussen, Boggaard, &Nielsen, 1999) for measuringfat content in muscle. This portable apparatus uses electrodes in-sertedinthemuscle, andfatcontentisestimatedviameasure-ments made atseveral frequencies. Measurementsoffatcontentafter rigor are not consistent, because impedance in this case is alsoinuenced by membrane state.4.1.4. TendernessA study (Byrne et al., 2000) related the electrical properties ofmuscleaftercookingtotendernessasassessedbyWBSFandat-tempted to establish a link between the electrical properties andthe mechanical resistance of meat. The results showed there wasnodirectrelationshipbetweenmeattendernessandstraightfor-ward electrical measurements. This is due to the fact that conjunc-tivetissue, whichplaysacrucial roleintenderness, hassimilarimpedance to muscle ber and thus cannot be detected by electri-cal measurements.4.1.5. AgeingAgeing involves meat-tenderizing biochemical and physico-chemical processes. These processes include the action of endoge-nous proteases on muscle bre structure, a progressive increase inmembranewaterpermeability, andtheweakeningofconnectivetissues. Faure et al. (1972) set out to evaluate state of maturationbyquantifyingtheseeffects. Theyproposedanapproachbasedon the ratio of low-frequency impedance to high-frequency imped-ance, which decreases during refrigerated storage. However, Lepe-tit, Sale, Favier, andDalle(2002) showedthat between-animalvariations of this ratio stemmed from variations in ion or fat con-tents. Furthermore, thisimpedanceratiocannotreliablyindicatethe state of meat maturation or destructuring. A similar study re-portedtheratioof capacity(thedielectricparameter reectingtheinsulatingstateof themembranes) toelectrical resistance(Kleibel, Pfzner, &Krause, 1983), buttheparametersmeasuredwere also affected by tissue adiposity.Muscle is electrically anisotropic, meaning that muscle and thusmeat exhibit changes inelectrical properties according to thedirection of the electrical elds in the sample. After rigor mortis,the electrical impedance of meat decreases linearly with themechanicalresistanceofmusclebres, andelectrical anisotropyis a better predictor of muscle bre strength than impedance alone(Lepetit et al., 2002). The rate of ageing in beef varies tremendouslyfromoneanimal toanother. Thestrengthof musclebrescanreach its minimum value within a few days, whereas for the samemuscle in another animal it can take more than two weeks. It hasbeenshown(Lepetit&Hamel, 1998)thatitispossibletoselectmeats which age rapidly if the state of ageing is known at 48-h postmortem. Thiswill avoidstoring, already-agedmeatsduringlongperiods. The expected benets include a 50% cut in storage costs.Thestudymeasuredageingstateusingadestructivemechanicalmethod, but thissameinformationcanbeobtainedfromnon-destructive sensors. One such sensor made by Damez et al.(2008) useselectrical impedanceanisotropy, andhasbeenpat-ented (Lepetit et al., 2007).4.2. Microwave characterizationThe interactionof microwaves andfoodproducts has beenexploitedforheatinginmanyapplicationsforthawing, cookinganddisinfectionpurposes. Recently, however, therehasbeenanemergenceof sensor systems basedontheinteractionof low-power (microwave sources no more powerful than in mobilephone devices) electromagnetic microwaves with biological mat-J.-L. Damez, S. Clerjon/ Meat Science 80 (2008) 132149 141ter. In the microwave frequency range (almost 0.3300 GHz), thedielectric properties of biological tissues are closelycorrelatedwithwatercontentandstate(Kent&Jason, 1974). Inparticular,dielectricrelaxationspectroscopydeterminesthemolecularmo-tion response of sample polar molecules (mainly water) to a weakexternal alternativeelectriceld. Aselectriceldfrequencyin-creases, it reaches a frequency called relaxation frequency wherethepolar moleculecannolongerrotatewiththeelectriceld.Dielectric properties change markedly around this relaxation fre-quency. Thetechniquehas beenexploredfor measuringwateractivityinproteicgels(Clerjon, Daudin, &Damez, 2003). Wateractivity is a parameter closely connected to water binding whichisrelatedtowaterholdingcapacityinmeatandsotoPSEandDFDphenomena. Moreover, fraudulentlyaddedwater inmeatproducts can be detected thanks to its higher value of water activ-ity(Kent &Anderson, 1996; Kent, Knochel, Daschner, &Berger,2001).Dielectric properties not only depend on water binding in foodmaterial butalsoonfoodcomposition. Foranygivenmolecularcomposition, thedielectricspectrumwillchangewithmolecularbinding. In real material, the complex interplay between molecularcomposition, presence of ions, electrical charges on proteins, andpHvariationsleadstoacomplexdielectricspectrumregulatedby several phenomena.Microwave approaches applied for macroscopic structural mea-surementsarebasedonthedielectricanisotropyofmuscle. Thestructural organizationandcompositionof muscle makes it ahighlyanisotropicdielectricmaterial. This dielectricanisotropywas modelled by Felbacq, Clerjon, Damez, and Zolla (2002) to pro-videinsightintomicrowave-muscleinteractions. Ittendstode-crease during ageing- or process-related cellular degradation.Polarimetric measurements,i.e., with a linearly polarized electriceld, makeit possibletoevaluateanisotropy. Thismethodhasbeen applied to assessmeat ageing and sh freshness (Clerjon &Damez, 2005; Clerjon & Damez,2007),and Brunton et al. (2006)linked dielectric beef muscle anisotropy to proteic changes duringcooking.Tejada, De las Heras, and Kent (2007) reported the results of theTorrymeter Distell Freshness Meter, a commercial microwave sen-sor dedicated to sh freshness evaluation. At a more macroscopiclevel, the premature eld of microwave tomography (Christensen,2004; Semenov et al., 2007) shows promise for giving informationon tissues organization within a piece of meat.5. X-ray measurementsX-rays have long been used in medicine and others areas. Theprinciple is to obtain a measurement of the attenuation of the pe-netratingenergy. Different materials havedifferent attenuationproperties, andsodependingonthelevelofpenetratingenergy,it should be possible to obtain quantitative measurements, in par-ticular for bone, lean meat and fat. Multiple technology tools usingX-ray beams at different energy levels have been developed, mak-ing it possible to discriminate fat, bone and lean meat according tothe energy attenuation measured. Over the last 30 years, the meatindustry has been using low-energy X-ray systems like the Anyl-Ray system (The Kartridg Pak Co., Iowa) (Gordon, 1973).Dual-energy X-ray absorption (DXA) is a useful technology formeat fat assessment. Absorption at lowX-ray energies (e.g.,62 keV) is dependent onbothfat content andsampledensity,while absorption at higher energies (e.g., 120 keV) mainly dependsonthedensity. Couplingthetwomeasurementandsubtractingone from another gives the fat content (Brienne, Denoyelle, Bauss-art, & Daudin, 2001; Hansen et al., 2003) with very good accuracycomparedtochemical analysis(R2valuesfrom.7to.97). Otherresearchers have attempted to use DXA to predict the tendernessof raw beef and cooked lamb meat, but the method gave moderateand poor results in comparison with WBSF (R2= .69 andR2= .12,respectively) (Kroger, Bartle, West, Purchas, & Devine, 2006). As re-ported by (Mercier et al., 2006) although DXA is too slow for com-mercial use, it may be used as reference method in carcasscomposition studies. It should also be noted that Earlier, Diesbourg,Swatland, andMillman(1988)postedencouragingresultsforX-ray diffraction measurements of post mortem changes in the porkmyolament lattice.6. Nuclear magnetic resonanceNMRcontributes to the characterizationof many products,includingmusclefood. Thehighcostsinvolveddomakeitcur-rently difcult to consider installing NMR systems on productionlines. The tool nevertheless has a wide range of research applica-tions, particularly for product assessment, and can be seen as a ref-erence method given the richness of measurements obtained: thediffusion coefcient and the relaxation time being the most useful.NMRisbasedontheabsorptionandemissionof energyintheradiofrequency range of the electromagnetic spectrum. All nucleithat contain odd numbers of protons or neutrons have an intrinsicmagnetic moment andangularmomentum. Themostcommonlymeasured nuclei are hydrogen-1 (the most sensitive isotope at nat-ural abundance) and carbon-13, although nuclei from isotopes ofmanyother elementscanalsobeobserved(23Na,31P. . .). NMRstudies magnetic moments by aligning them with an applied con-stant magnetic eld and perturbing this alignment using anorthogonal alternating radiofrequency magnetic eld. This pertur-bation induces a resonant phenomenon which is exploited in NMRspectroscopy and magnetic resonance imaging (MRI).6.1. NMR spectroscopyA review (Bertram & Andersen, 2004) and several papers (Ber-tram, Purslow, &Andersen, 2002;Bertrametal., 2001)describethe status of NMR applications in meat science and explain the po-tential and relevance of spectroscopic and relaxation-based meth-odologiestodifferenttopicsofimportanceformeatscience. Themost widely explored area of NMR in meat science is proton relax-ometry. The use of relaxometry has been highly successful due toits ability to characterize water and structural features in heteroge-neous systems like meat. Venturi et al. (2007) showed how NMRspectroscopycanmeasurewateractivityinfreeze-driedchickenbreast meat by studying of the shape of the T2 relaxogram. Low-eld waterproton NMRT2 relaxometry has been widely used todetermine WHC, which is closely linked to myobrillar structure,for beef (Tornberg & Larsson, 1986) and processed pork and meatquality by the INRA (Renou, Kopp, Gatellier, Monin, & Kozakreiss,1989)andtheDanishInstituteforAgriculturalScience(Bertram& Aaslyng, 2007; Straadt et al., 2007; Bertram, Kristensen, & Ander-sen, 2004; Mortensen, Andersen, Engelsen, & Bertram, 2006). Mick-lander, Peshlov, Purslow, and Engelsen (2002) underlined theability of NMR to track structural changes in pork during cooking.A recent study (Wu, Bertram, Bocker, Ofstad, & Kohler, 2007) dem-onstrated that the changes in water proton T2 relaxation times af-fected by heating rate and raw pork quality (DFD, PSE or normal)arecloselyrelatedtoproteinsecondarystructurechanges. Thesame method was exploited by Ahmad, Tashiro, Matsukawa, andOgawa(2005)tostudythegelationcharacteristicsofshsurimigel byobservingitsmolecular dynamics. Thermal denaturationof bovine connective tissue has also been studied using water pro-ton NMR T2 relaxometry (Rochdi, Foucat, & Renou, 1999; Rochdi,Foucat, & Renou, 2000).142 J.-L. Damez, S. Clerjon/ Meat Science 80 (2008) 132149Relaxation parameters are also a good indicator of water hold-ingcapacity(WHC)inmeat. WHCisanimportantmeatqualitytraitforconsumeracceptance. MeatWHCdependsprimarilyonthe lateral tensing of the myobrils during the rise in rigor mortischanges associated with uid ow between water compartmentsin muscular tissue (Offer & Knight, 1988). NMR measurements ofwaterprotonrelaxationtimesgiveinformationonthedynamicsofwater. Signicantcorrelationshavebeenhighlightedbetweenmeasuredvalueandindicatorrelaxationtimesformeatqualityparameters such as pH, WHC or losses to cooking (Fjelkner-Modig,Persson, & Tornberg, 1986; Renou, Kopp, & Valin, 1985).Sincefatcontentcontributestomeattextural properties, wecanalsocite studies byRenou,Monin, and Sellier (1985)Foucat,Donnat, Martin, Humbert, and Renou (1997) on meat fat contentmeasurement using NMR spectroscopy. Furthermore, post mortemmetabolism has been studied in pig muscle using31P NMR to pre-dict PSE and DFD defaults early, either post mortem (Miri, Talmant,Renou, & Monin, 1992) or ante mortem (Lahucky et al., 1993).Very recently, a new technique was described for the absolutequanticationof double-quantumlteredspin-3/2nuclei23Naspectra(Mouaddab, Foucat, Donnat, Renou, &Bonny, 2007). Thismethod paves theway for absolute quanticationof both boundandfreefractionsof Na+, whicharedeterminingfactorsinthecharacterization of salted/brined/dried meat products. Still in theeldof saltedmeat, Foucat, Donnat, andRenou(2003) studiedthe interactions of sodium and chloride ionswith meat productsbymeansof23Naand35ClNMRspectroscopy. Thesamekindofinvestigationwas done onfreshandfrozen-thawedcodllets(Erikson, Veliyulin, Singstad, & Aursand, 2004).Thepostmortemevolutionof energy-richcompoundscanbefollowed using31PNMR (Renou,Canioni,Gatelier, Valin, &Cozz-one, 1986). This technique makes it possible to measure the con-centrations of ATP, creatinephosphates, sugar phosphates andinorganic phosphate (Pi). The pH calculation starts from the chem-ical shift of the peak associated to Pi. NMR has so far proven veryuseful tool forstudyingmetabolicchangesandpHevolutioninthe muscle in relation to the technological quality of the meats.6.2. Magnetic resonance imagingThemajor interestof thistechnique is that it cansolvemanyproduct control problems during production. NMR sequences runon bovine samples can produce images where the water and lipidsignals are selected, making it possible to identify the various com-ponents of the conjunctive network. This technique is further en-hancedbysusceptibility imagingwhichmakes it possibletolocatetheconjunctivenetworkberswhosethicknessismuchlowerthanthedimensionsof theNMRimagevoxels. NMRmi-cro-imaging on meat samples can be used to quantitatively charac-terizelipids. Fat content analysis byNMRrequires choosingasequenceof impulses(Inversion-Recovery, SpinEcho, orothers)tobe appliedtothe product. Theresults highlight theversatilityand practicability of the technique, since the equipment involvedis compact and the method can be equally well deployed for con-trolling food composition as for checking food quality.NMR imaging (MRI) generates a morphological image of a sam-ple distinguishing bone, fat and lean meat. Sample elements can bedifferentiated by differences in water content and in water mobil-ity in various biological elements. Water content and water mobil-ity are variables that can be studied by measuring particular NMRparameters (protondensity, relaxationtime, T1, T2, T2, diffusioncoefcient, etc.). Fat was quantied in ground beef by NMR (Foucatetal., 1997). Theresultsshowedexcellentcorrelation(R2= .992)with the Soxhlet method. For the test range of 515% fat content,actual fatcontent was determined with verygood accuracy. It isthuspossibletocharacterizetheconnectivetissuestructureofthe perimysium (Fig. 3) (Bonny et al., 2001; Laurent, Bonny, & Re-nou, 2000).23NaNMRspectroscopypresentedin6.1. isalsousedinMRIstudies. Saltingressinmuscleproductsisconnectedwithstruc-tural integrity as membranes act as barriers for ions diffusion. Gui-heneuf, Gibbs, and Hall (1997) showed with 23Na MRI that sodiumions ingress into post rigor porcine muscle during brining follow aFicks second law. Bertram, Holdsworth, Whittaker, and Andersen(2005)introducedtheuseofcombined23NaMRIand23NaNMRspectroscopyforthestudyof thediffusionof sodiumionsintothemeatduringcuring. Resultsrevealedadecreaseindiffusioncoefcient, suggesting that changes occur in the microscopic struc-ture of the meat during curing.23Na NMR spectroscopy gives herecomplementary information to MRI by the identication of two so-dium populations.23Na MRI quantication of sodium mobility inporkduringbrinewasalsoinvestigatedatdifferentpHandpostmortem ages (Vestergaard, Risum, & Adler-Nissen, 2005) suggest-ing the diffusion coefcient to be affected by changes in NaCl con-centration, swelling and degree of dehydration.In the past, post mortem muscle studies have characterized bertypes (type I: slow-twitch oxidative; type IIa: fast-twitch oxidativeglycolytic;typeIIb:fast-twitchglycolytic)withT1andT2values(Adzamli, Jolesz, Bleier, Mulkern, & Sandor, 1989; Lerumeur,Decertaines, Toulouse, & Rochcongar, 1987), where T2 can differen-tiate type I bers from others with slightly higher values.In vivostudies on rabbits highlighted a higher T2 for muscles characterizedby type I bers (Bonny et al., 1998), which the authors suggestedwasduetowaterstructuralizationinmusclebers, fatcontent,and myoglobin state. Fig. 4 illustrates these researches.6.3. Magnetic resonance elastographyLiketransientelastography, magneticresonanceelastography(MRE) measures local viscoelastic properties by following anacoustic wave in a biological tissue. The technique takes advantageof MRI image quality, and microscopic MRE, which is made possi-blewithhigh-resolutionMRI, (Othman, Xu, Royston, &Magin,Fig. 3. Magneticresonanceimaging(MRI). Characterizationof perymisium(P);myobers (M). Image of beef gluteo biceps 24Hpost mortem, MRI 4.7 Tesla,resolution 300 lm (Bonny et al., 2001).J.-L. Damez, S. Clerjon/ Meat Science 80 (2008) 132149 1432005), is a promising method for carrying out mechanical investi-gations at the microscopic scale.MRE has not been yet used in the eld of meat science, but theapproach is under development at the INRA in France to measurelocal mechanical parameters on whole muscle to gain insight intothe structural changes occurring in meat products during process-ing. MRE has already been applied in the biomedical eld to studybreast and liver diseases (Sinkus et al., 2000; Manduca et al., 2001).Research has been published on muscles, which are an anisotropicmaterial, and the conclusions will drive MRE development in themeat sciences (Basford et al., 2002; Dresner et al., 2001; Papazog-lou, Rump, Braun, & Sack, 2006; Uffmann et al., 2004; Sinkus et al.,2005).7. DiscussionThis overview has highlighted how many biophysical methodsareabletomeasureparametersdirectlyorindirectlyconnectedwithmeatstructure. Thewiderangeof physical principlesem-ployed means that individual methods give different key informa-tionformeatstructureassessment. Table1givesasummaryofthesemethodswiththetypeofinformationtheygiveandtheirmainadvantagesanddrawbacks. Whilecomparativeanalysisisuseful (Brondum et al., 2000) the way forward is to combine thetechnique and thus improve sensor performances. Several labora-tories have followed this direction.Sifre et al. (2005) investigated spatial organization of the peri-mysiuminbeef meat usingacombinedhistologyandMRI ap-proach. Thetechniqueprovidedcomplementarymicroscopic andmacroscopicdatasetsonintramuscularconjunctivetissuestruc-tures, both of which are necessary for predicting sensory tender-ness. Thesamemethodsareexploitedinassociationwithrigorindextopredictsofteshproblemsinfreshwaterrainbowtrout(Foucat, Taylor, Labas, & Renou, 2004) or to detect frozen-thawedsh (Foucat, Taylor, Labas, & Renou, 2001).Fourier transforminfrared microspectroscopy and low-eldproton NMR transverse relaxation measurements were combinedtostudyageing-relatedchanges inproteinsecondarystructureand water distribution in pork following salting and cooking (Wuet al., 2006).Swatlandsteamshaveinvestigatedanumberof biophysicalmethodsforanalyzingmeatquality, sometimesincombination.Forinstance, Swatland(2001)dealswithcombiningspectropho-tometry and uorometry inthesameprobe todiscriminate con-nectiveandadiposetissues. Theyhavealsocoupledoptical andmechanical measurements to evaluate meat tenderness (Swatland,Brooks, & Miller, 1998), pairing these two biophysical parameterswith electrical properties to develop a three-system probe for meattexture measurement (Swatland, 1999).Another eld under development that is closely linked to bio-physical methods in meat science is data processing. Spectral andimage analysis or combined biophysical methods are very prolicdata generators, and it is a real challenge to nd solutions optimiz-ing the data processing steps. Multivariate data analysis (chemo-metrics) canmoreefcientlytacklefoodscienceproblemsand,moreover, solveproblemsthatcouldnotbehandledbefore(Broet al., 2002). For instance Kohler et al. (2007) developed a very use-ful tool for FT-IR microscopy data processing.The linear myobrillar structure of muscle food gives of its bio-physicalpropertiesstronganisotropythatcanbeassessedusingappropriate techniques. Such techniques, using polarized electro-magneticwaves, eithervisible, IR, UVopticalbands, microwavesas well as electrical conductivity and US, have been highlighted.All the results presented here point out the wealth of potentialfor using biophysical methods in meat quality investigations.Although a high number of studies have already been conducted,the eld of research is vast, and meat scientists still have years ofexciting work ahead before offering to meat industry a cheap, ro-bust, reliable, portable, rapid, universal . . . magnicent meat qual-ity sensor!AcknowledgementsAuthors would like to thank their colleagues Sa d Abouelkaram,ThierryAstruc, LoicFoucat, RolandLabasandFredericPeyrinforuseful discussions.ReferencesAbouekaram, S., Berge, P., &Culioli, J. (1997). Applicationof ultrasonicdatatoclassify bovine muscles. Proceedings of IEEE Ultrasonics Symposium, 2,11971200.Abouelkaram, S., Chauvet, S., Strydom, P., Bertrand, D., & Damez, J. L. (2006). Musclestudy with multispectral image analysis. In Proceedings of the 52nd internationalcongress of meat science and technology (pp. 669670). Declan Troy Edt:Wageningen Academic Publishers.Abouelkaram, S., Laugier, P., Fink, M., & Culioli, J. (1992). Ultrasonic measurement ofmeat heterogeneity due to connective tissue. In: Clermont-Ferrand (Ed.),Proceedings of the 38th ICoMST (Vol. 5, pp. 855858), 2328 aot.Abouelkaram, S., Suchorski, K., Buquet, B., Berge, P., Culioli, J., Delachartre, P., et al.(2000). Effects of muscle texture on ultrasonic measurements. Food Chemistry,69, 447455.Adzamli, I. K., Jolesz, F. A., Bleier, A. R., Mulkern, R. V., & Sandor, T. (1989). The effectof GadoliniumDtpaontissuewatercompartmentsinslow-twitchandfast-twitch rabbit muscles. Magnetic Resonance in Medicine, 11(2), 172181.Ahmad, M. U., Tashiro, Y., Matsukawa, S., & Ogawa, M. (2005). Comparison of horsemackerel and tillapia surimi gel based on rheological and H-1 NMR relaxationproperties. Fisheries Science, 71(3), 655661.Fig. 4. Characteristic MRT2parametric image of rabbit thightransverse slicedistinguishing Type I ber muscles from Type II ber muscles. BF, biceps femoris; RF,rectus femoris; SM, semimembranosus; SP, semimembranosus proprius; VI, vastusinternalis (Bonny et al., 1998).144 J.-L. Damez, S. Clerjon/ Meat Science 80 (2008) 132149Allais, I., Viaud, C., Pierre, A., & Dufour, E. (2004). A rapid method based on front-faceuorescence spectroscopy for the monitoring of the texture of meat emulsionsand frankfurters. Meat Science, 67, 219229.Andersen, C. M., & Wold, J. P. (2003). Fluorescence of muscle and connective tissuefrom cod and salmon. Journal of Agricultural and Food Chemistry, 51(2), 470476.Andres, S., Murray, I., Navajas, E. A., Fisher, A. V., Lambe, N. R., & Bunger, L. (2007).Predictionof sensorycharacteristicsof lambmeatsamplesbynearinfraredreectance spectroscopy. Meat Science, 76(3), 509516.Aronson, J. F.,&Morales, M. F. (1969). Polarizationoftryptophan uorescenceinmuscle. Biochemistry, 8(11), 45174522.Astruc, T., Marinova, P., Labas, R., Gatellier, P., &Sante-Lhoutellier, V. (2007).Detection and localization of oxidized proteins in muscle cells by uorescencemicroscopy. Journal of Agricultural and Food Chemistry, 55(23), 95549558.Barbut, S. (1993). Color measurments for evaluating the pale soft exudative (PSE)occurrence in turkey meat. Food Research International, 26(1), 3943.Basford, J. R., Jenkyn, T. R., An, K. N., Ehman, R. L., Heers, G., & Kaufman, K. R. (2002).Evaluation of healthy and diseased muscle with magnetic resonanceelastography. Archives of Physical Medicine and Rehabilitation, 83(11),15301536.Basset, O., Buquet, B., Abouelkaram, S., Delachartre, P., & Culioli, J. (2000).Applicationof textureimageanalysis for theclassicationof bovinemeat.Food Chemistry, 69(4), 437445.Beattie, R. J., Bell, S. J., Farmer, L. J., Moss, B. W., & Desmond, P. D. (2004). Preliminaryinvestigation of the application of Raman spectroscopy to the prediction of thesensory quality of beef silverside. Meat Science, 66(4), 903913.Beattie, J. R., Bell, S. E. J., Borgaard, C., Fearon, A., & Moss, B. W. (2006). Prediction ofadipose tissue composition using Raman spectroscopy: Average properties andindividual fatty acids. Lipids, 41, 287294.Belk, K. E., George, M. H., Tatum, J. D., Hilton, G. G., Miller,R. K., Koohmaraie, M.,et al. (2001). Evaluation of the Tendertec beef grading instrument to predict thetenderness of steaks frombeef carcasses. Journal of Animal Science, 79(3),688697.Benedito, J., Carcel, J. A., Rossello, C., & Mulet, A. (2001). Composition assessment ofraw meat mixtures using ultrasonics. Meat Science, 57(4), 365370.Berg, E. P., Kallel, F., Hus