A fully implantable telemetry system for the long-term measurement of habitual bone strain

A fully implantable telemetry system for the long-term measurement of habitual bone strain
of 5
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Short communication A fully implantable telemetry system for the long-term measurement of habitual bone strain W.C. de Jong  , J.H. Koolstra, L.J. van Ruijven, J.A.M. Korfage, G.E.J. Langenbach Department of Functional Anatomy, Academic Centre for Dentistry Amsterdam (ACTA), Research Institute MOVE, University of Amsterdam and VU University Amsterdam,Meibergdreef 15, 1105 AZ Amsterdam, The Netherlands a r t i c l e i n f o  Article history: Accepted 27 September 2009 Keywords: Bone strainTelemetry In vivo Strain eventStrain rate a b s t r a c t Long-term  in-vivo  recordings of habitual bone strain in freely moving animals are needed to betterunderstand the everyday mechanical loading environment responsible for bone-tissue maintenance.However, wireless methods to make such recordings are scarce. We report on the successfulcustomisation of a commercially available voltage transmitter hooked-up to a strain-gauge rosette, itssubcutaneous implantation in rabbits, and the quality of the implant’s strain-gauge recordings.Continuous wireless recordings of a completely operational strain-gauge rosette glued to themandibular surface of a freely moving rabbit could be made up to 33h. The resolution of the systemwas 1.5 microstrains/bit. The noise in the signal was 4.5 microstrains. To facilitate the automaticcounting of bone-strain events in the retrieved data, and to calculate their peak amplitude, a novelapproach is presented. The described technique enables the quantification of the daily bone-strainhistory defining the architecture and composition of bone tissue, and can help to further elucidate thestrain parameters which influence bone tissue. &  2009 Elsevier Ltd. All rights reserved. 1. Introduction In-vivo  measurements of bone strain are crucial to ourunderstanding of the loading environment on which bone basesits mechanisms of modelling and remodelling (Burr et al., 1996;Main and Biewener, 2006; Milgrom et al., 2007). The majority of  these measurements, however, are made during specific beha-viours only and do not provide much information on the dailybone-strain history. As small postural loads, and the length anddistribution of periods of relative loading ‘silence’, might influenceadaptive responses of bone just as well (Fritton et al., 2000;Gardner et al., 2008; Robling et al., 2002; Turner et al., 1995), the feasibility of long-term  in-vivo  recordings of habitual daily bone-strain histories is of great importance.For  in-vivo  measurements of bone deformation strain gaugeshavebeen used successfullyever since 1944 (Gurdjian and Lissner,1944; Main and Biewener, 2006). As they are not bone-damaging mechanically, strain gauges maybe regarded as the bestoption formeasuring bone strain  in vivo  despite their need for constantpowering. Central to long-term  in-vivo  measurements is tele-metry, which avoids hindering the laboratorysubject in its normalbehaviours. The combination of telemetryand strain gauges needsfurther attention to allow for long-term measurements of   in-vivo bone strain.We report on the successful customisation of a commerciallyavailable voltage transmitter into a bio-implantable telemetrypackage able to read out a bone-bonded strain-gauge rosette in vivo . The experiments were conducted using freely moving,but caged, rabbits. Daily habitual mandibular bone strainswere collected. Furthermore, a new approach is presented todetect individual bone-strain events and to calculate their peakamplitudes. 2. Materials and methods  2.1. Implantable transmitter  A MicroStrain V-Link Voltage Node transmitter (Williston, Vermont, USA) wasused for the data transmission. After removal of its housing, the device wasconnected to two parallel-linked 3.7-Volts lithium-ion rechargeable batteries(Ultralife Batteries, Newark, New York, USA) and fitted with a bipolar Hall latch(US3881, Melexis, Concord, New Hampshire, USA) to enable the transmitter to beswitched on and off magnetically (Fig. 1). Three of the V-Link’s voltage channelswere connected to a strain-gauge rosette using insulated stranded copper wire.The V-Link featured on-board Wheatstone-bridge completion, shunt-calibrationresistors, and a programmable hardware gain. The modified transmitter wascovered with medical-grade silicones (Sylgard 186 silicone elastomer, DowCorning, Midland, Michigan, USA). The transmitter’s dimensions wereapproximately 9.0  9.0  1.8cm, weighing about 100g.The strain-gauge voltages were sampled at a hardware-determined frequencyof 617Hz using a 12-bit analog-to-digital converter. The resolution was 1.5 ARTICLE IN PRESS Contents lists available at ScienceDirectjournal homepage:  Journal of Biomechanics 0021-9290/$-see front matter  &  2009 Elsevier Ltd. All rights reserved.doi:10.1016/j.jbiomech.2009.09.036  Corresponding author. Tel.: +31205665355. E-mail address: (W.C. de Jong). Journal of Biomechanics 43 (2010) 587–591  ARTICLE IN PRESS microstrains/bit ( me /bit). Data transmission took place at a carrier frequency of 2.4GHz. A MicroStrain USB Base Station, placed within 2m of the laboratoryanimal’s cage, received the data and stored them unprocessed on a desktopcomputer. The MicroStrain software program Agile-Link was used for real-timemonitoring of the data and wireless calibration of the strain-gauge voltagechannels.  2.2. Bone-strain sensor  Stacked, 350- O  strain-gauge rosettes (L2A-06-031WW-350, Vishay, Malvern,Pennsylvania, USA) were used to measure bone-surface deformation of the leftlateral side of the rabbit mandibular corpus. The gauges were positioned withinthe rosette in a 0  45  90 1  configuration, each featuring a sensor grid of 0.79  1.79mm. The strain-gauge factor  G  (=resistance increase  srcinal re-sistance  1  strain  1 ) was 2.11. The soldered gauge rosette was insulated withPlastik 70 (CRC Industries Europe N.V., Zele, Belgium), coated with nail polish, andcovered with Colt  ene President polyvinyl siloxane (Cuyahoga Falls, Ohio, USA) formechanical protection.  2.3. Laboratory animals and surgical procedure Eight adult male New Zealand white rabbits ( Oryctolagus cuniculus ) withweights between 3.5 and 4.0kg were used. The animals were kept in 73  73  46cm cages, provided with food pellets and water  ad libitum , and hay and cageenrichment. Lights were dimmed between 18.00 and 6.00h. The animals weregiven at least two weeks of acclimatisation time before the surgical placement of the telemetry package. The transmitter was positioned on the rabbit’s left flankand its battery compartment on the right flank, both subcutaneously. The wiredstrain-gauge rosette was led subcutaneously to a second incision in themandibular region. There, the rosette was glued (Histoacryl, B. Braun, Tuttlingen,Germany) to the mandibular surface after spreading of the overlying soft tissues,careful removal of the periosteum, degreasing of the bone surface with 80%ethanol, and drying of the bone surface to the air (Cochran,1972). Care was takento position the rosette the same way in each animal (Fig. 1). The behaviour of therabbits was monitored by animal caretakers and researchers at a daily basis beforeand during the experimental procedure. Rabbits were sacrificed after failure of allthree strain-gauge channels or after battery-power depletion. The Animal EthicsCommittee of the Academic Medical Centre of the University of Amsterdamauthorised the study, which was executed according to Dutch legislation.  2.4. Strain-event detection and amplitude calculation To facilitate the automatic counting of bone-strain events in the long-termrecordings, we defined a strain event to take place whenever the recorded bonestrain increased for an uninterrupted period. A low-pass filter was used to smooththe srcinal strain data (i.e. a moving average over 0.040s), after which the datawere transformed into their first derivative: the slope of the strain or strain rate. Aminimum threshold above the noise level was superimposed on the strain-ratesignal. All episodes exceeding the threshold were counted as strain events (Fig. 2).To calculate the amplitude of a bone-strain event, the area under theaccompanying, threshold-crossing slope-of-strain curve was measured. This areaequals the difference between the strain at the end and the strain at the beginningof a period of increasing strain (Fig. 2).Analysing procedures were performed using a custom-made script in Spike2,version 5.21 (Cambridge Electronic Design Limited, Cambridge, England). 3. Results No aberrant behaviour was observed in any of the rabbits afterplacement of the telemetry implant. Continuous telemetricrecordings of   in-vivo  bone strain with all three gauges operationalwere made up to 33h, starting four to five hours after surgery(Table 1). Battery depletion mostly determined the recording endpoint, but, incidentally, one or more strain gauges failed earlierdue to disconnected lead wires or moist short-circuiting thesolder dots. In all rabbits, post-mortem dissection revealed thatthe gauge rosettes were still firmly attached to the bone surface.The long-term recordings displayed drift of the zero-strainlevel (Fig. 3A). The transmitter delivered a high-quality strainsignal able to keep up with high amplitudes (Fig. 3B) and highfrequencies (Fig. 3C). The noise in the strain recordings, measuredfrom background recordings in the telemetry room, encompassedthree bits (4.5 me ) maximally.Manual comparisons between the number of strain eventscounted by the new detection method and those found in the rawdata revealed that the computed number corresponded very wellwith the amount of strain events in the original bone-strainrecordings (Fig. 4). The calculated amplitudes, however, almostalways underestimated the amplitudes read off directly from therecordings: the mean amplitude of the strain events in Fig. 4 is38 7 16 me  when read off versus 18 7 9 me  (not shown) whencalculated. 4. Discussion 4.1. Telemetry system We have developed a fully implantable telemetry system, ableto sample a bone-bonded strain-gauge rosette at a rate of 617Hzfor up to 33h continuously  in vivo . The implant is suitable formiddle-sized and larger animals, as it weighs   100g. The highsample frequency and resolution of this system will enable,amongst others, frequency analysis and measurement of the dailyhabitual bone strains featuring the high-frequency and low-amplitude characteristics associated with beneficial bone adapta-tions (Beck et al., 2006; Rubin et al., 2001). In-vivo  telemetry of skeletal loading has since long beenaccomplished using instrumented prosthetic implants, mostlyfemora (Bergmann et al., 1988; Davy et al., 1988; English and Kilvington,1979). This approach is of great importance to the fieldof orthopaedics and has seen many technical advances over theyears (Bassey et al., 1997; D’Lima et al., 2005; Graichen et al., 2007; Lu et al.,1997). Nevertheless, load-sensitive prosthetics are of limited use to the field of bone mechanobiology. To studyprocesses such as adaptive bone modelling and remodelling, trans-mitter US3881 + - 9 cm2 rostroventral vertical caudoventral 13 Fig.1.  Left: schematic representation of the electronic circuit connecting the Hall-effect sensor (US3881), the transmitter, and its battery pack.1=supply voltage, 2=opendrain output, and 3=ground. Middle: photograph of the implant. Right: Left lateral view of the rabbit skull, and some masticatory muscles, showing the location of thestrain-gauge rosette (superimposed rectangle) on the mandibular corpus and the directions (arrows) in which its three strain gauges are aligned. W.C. de Jong et al. / Journal of Biomechanics 43 (2010) 587–591 588  ARTICLE IN PRESS which are mostly tissue-deformation driven, it is needed tomeasure the strains of bone itself. To this day, the telemetricrecording of bone strain  in vivo  is still a great challenge.Young et al. (1977) succeeded in implanting a bone-straintransducer and its associated transmitter into a monkey hind leg.Their telemetry package collected tibial loads during variousactivities of the caged animal and transmitted these throughpulse-interval ratio modulation. However, their transducer wasbone-invasive, compromising the architecture of the bone andpossibly causing unphysiological dissipation of bone loads. In1980, Schatzker et al. wirelessly collected femoral strainsfrom a dog. Although they did make use of bone-sparingstrain gauges, their telemetry system was not fully implantable,as the transmitter was placed in a harness strapped to the dog.The same approach was practised in gibbons by Swartz et al.(1989). In 2002, Szivek et al. reported on a MicroStrain bone- strain telemetry unit, which was later fully implanted in a 17-year-old female to capture vertebrae strains (Szivek et al., 2005).Biweekly wireless measurements were made up to 219 days post-operatively. The transducer they used was a calcium-phosphate-ceramic-coated strain gauge; a gauge variant that is incorporatedinto the bone surface over time and evades the problem of  strain [ µε ] 0 = 25 µε strain rate [ µε /s]   = threshold A 2 sec Fig. 2.  Diagram of the strain-event detection approach. The upper graph is a 2-second piece of srcinal unprocessed strain recording. The lower graph is the accompanyingstrain-rate signal, which is the derivative of the strain signal. Episodes during which the strain-rate amplitude was higher than a beforehand-determined threshold(horizontal dashed line) were counted as strain events. Of each strain-rate episode above this threshold the area under the curve (  A  in  me /s  s= me ) was measured toapproximate the amplitude ( e  in  me ) of the related strain event. The dotted lines between the strain and the strain-rate signal illustrate for two peaks, the link between thehighest strain-rate value and the steepest part of the strain increase, as well as the link between a strain rate of zero and the strain-event peak value.  Table 1 Lengths and mode of failure of the retrieved  in-vivo  bone-strain recordings for all eight experiments.Rabbit no. Strain gauge Recording duration (h:min) Recording end point determined by1 Rostroventral 19:15 Battery-power depletionVertical 19:15 Battery-power depletionCaudoventral 19:15 Battery-power depletion2 Rostroventral 11:00 Battery-power depletionVertical 11:00 Battery-power depletionCaudoventral 11:00 Battery-power depletion3 Rostroventral 0 Moist intrusionVertical 17:00 Battery-power depletionCaudoventral 17:00 Battery-power depletion4 Rostroventral 6:15 Moist intrusionVertical 15:00 Battery-power depletionCaudoventral 5:15 Moist intrusion5 Rostroventral 27:00 Battery-power depletionVertical 27:00 Battery-power depletionCaudoventral 27:00 Battery-power depletion6 Rostroventral 38:00 Battery-power depletionVertical 34:00 Moist intrusionCaudoventral 33:00 Moist intrusion7 Rostroventral 23:00 Battery-power depletionVertical 23:00 Battery-power depletionCaudoventral 23:00 Battery-power depletion8 Rostroventral 22:30 Lead wires torn off Vertical 30:00 Battery-power depletionCaudoventral 24:30 Moist intrusion W.C. de Jong et al. / Journal of Biomechanics 43 (2010) 587–591  589  ARTICLE IN PRESS dissolving tissue adhesives. However, their strain recordingscontained more noise ( 7 15 me ) than ours ( 7 5 me ). In subsequentpapers by the same research group, telemetry was combined withstrain-gauged scaffold plugs placed within the dog femoralcondyle to measure joint forces up to 98 days post-operation(Geffre et al., 2008; Szivek et al., 2006). Although this approach yields highly reliable joint-load measurements (Bliss et al., 2007),the invasive character of the scaffolds does not allow its use forbone-strain studies.In our system, power supply is the biggest constraint. Despiteovernight battery charging, there was considerable variation inbattery-pack lifetime. Because stronger batteries will increase thevolume and weight of the implant, power transfer via induction(Budgett et al., 2007; Geffre et al., 2008) might be an alternative. Unfortunately, the distance between a power coil and a transmit-ter pick-up coil cannot be too great and will limit the animal’smovements and behaviours. Yet another solution is to not recordcontinuously, but intermittently, using a remote-controlled on/off switch such as the Hall latch in our setup.The second constraint of our approach is debonding of rosettesin longer recordings. Calcium-phosphate-ceramic-coated straingauges could address this problem, although they can take up tosix weeks to fully attach to the bone surface (Rabkin et al., 2001).Other means of attaching gauges include miniature gauge-bearingframes (Szivek and Magee, 1989) or staples (Buttermann et al., 1994) but these methods require screws to be drilled into thebone and may distort natural strain patterns and initiate bone-healing processes. 8 hrs8 sec8 secverticalcaudoventralrostroventral= 600 µε = 600 µε  = 200 µε Fig. 3.  (A) Example fragment of eight hours of raw bone-strain data. Note the drift in all three gauge channels. Positive deviations refer to tensile deformation; negativedeviations refer to compressive deformation. (B) Eight-second fragment showing cyclic strain events with amplitudes up to 600 me . (C) Eight-second fragment showingloose events and a short 12Hz cyclic strain episode. 350 µε 300 µε 2000 µε /s-2000 µ ε /s0= thresholdstrainstrain rate4.5 sec Fig. 4.  Example fragment of 4.5s of data from a rostroventrally oriented gauge. The upper graph is the strain and the lower graph is the strain rate. The dotted line in thestrain-rategraph indicates an exemplary threshold of 150 me /s. Note the matching between strain events in the upper graph and crossingsof the threshold bythe strainratein the lower graph. W.C. de Jong et al. / Journal of Biomechanics 43 (2010) 587–591 590  ARTICLE IN PRESS 4.2. Strain-event analysis Often, strain events cannot be counted from strain recordingsautomatically, as drift of the zero-strain level can impair theplacement of a minimum threshold. A high-pass filter can removedrift, but will also introduce artefacts, due to overcompensation.Therefore, we used the strain-rate recordings instead, defining astrain event as an uninterrupted period of positive strain rate. Thismethod proved very effective to detect strain events, both aboveand below the zero-strain level, the tensile-strain, and relaxationevents, respectively. However, a distinction between tensile-strainevents and relaxation-strain events cannot be made with thismethod.Despite the fact that the calculated strain-event amplitudeconsistently underestimated the amplitude as read off from thesrcinal recordings, this method does allow for analysis of theamplitude distribution of daily bone-strain events.In summary, the technique reported in this study enables theexamination of all habitually occurring bone strains, resultingfrom spontaneous daily behaviour, in freely moving animals. Thesubsequent quantification of the number of daily strain events,and theiramplitudes, can be used tocharacterise a bone’s habitualloading history which, on its turn, is linked to bone’s tissuearchitecture and composition. Conflict of interest None.  Acknowledgements We thank David Crick from Cambridge Electronic DesignLimited for his help with the Spike2 software and our colleaguesat the Animal Research Institute Amsterdam (ARIA, AcademicMedical Hospital of the University of Amsterdam) for theirassistance with the  in-vivo  experiments. References Bassey, E.J., Littlewood, J.J., Taylor, S.J.G.,1997. Relations between compressive axialforces in an instrumented massive femoral implant, ground reaction forces,and integrated electromyographs from vastus lateralis during various ‘osteo-genic’ exercises. Journal of Biomechanics 30, 213–223.Beck, B.R., Kent, K., Holloway, L., Marcus, R., 2006. Novel, high-frequency, low-strain mechanical loading for premenopausal women with low bone mass:early findings. Journal of Bone and Mineral Metabolism 24, 505–507.Bergmann, G., Graichen, F., Siraky, J., Jendrzynski, H., Rohlmann, A., 1988.Multichannel strain gauge telemetry for orthopaedic implants. Journal of Biomechanics 21, 169–176.Bliss, C.L., Szivek, J.A., Tellis, B.C., Margolis, D.S., Schnepp, A.B., Ruth, J.T., 2007.Sensate scaffolds can reliably detect joint loading. Journal of BiomedicalMaterials Research Part B: Applied Biomaterials 81B, 30–39.Budgett, D.M., Hu, A.P., Si, P., Pallas, W.T., Donnelly, M.G., Broad, J.W.T., Barrett, C.J.,Guild, S.J., Malpas, S.C., 2007. Novel technology for the provision of power toimplantable physiological devices. Journal of Applied Physiology 102,1658–1663.Burr, D.B., Milgrom, C., Fyhrie, D., Forwood, M., Nyska, M., Finestone, A., Hoshaw, S.,Saiag, E., Simkin, A.,1996. In vivo measurement of human tibial strains duringvigorous activity. Bone 18, 405–410.Buttermann, G.R., Janevic, J.T., Lewis, J.L., Lindquist, C.M., Wood, K.B., Schendel, M.J.,1994. Description and application of instrumented staples for measuring invivo bone strain. Journal of Biomechanics 27, 1087–1094.Cochran, G.V.B., 1972. Implantation of strain gages on bone  in vivo . Journal of Biomechanics 5, 119–123.Davy, D.T., Kotzar, G.M., Brown, R.H., Heiple Jr., K.G., Goldberg, V.M., Heiple, K.G.,Berilla, J., Burstein, A.H., 1988. Telemetric force measurements acrossthe hip after total arthroplasty. The Journal of Bone and Joint Surgery 70-A,45–50.D’Lima, D.D., Townsend, C.P., Arms, S.W., Morris, B.A., Colwell Jr., C.W., 2005. Animplantable telemetry device tomeasureintra-articular tibial forces. Journal of Biomechanics 38, 299–304.English, T.A., Kilvington, M., 1979. In vivo records of hip loads using a femoralimplant with telemetric output (A preliminary report). Journal of BiomedicalEngineering 1, 111–115.Fritton, S.P., McLeod, K.J., Rubin, C.T., 2000. Quantifying the strain history of bone:spatial uniformity and self-similarity of low-magnitude strains. Journal of Biomechanics 33, 317–325.Gardner, M.J., Ricciardi, B.F., Wright, T.M., Bostrom, M.P., van der Meulen, M.C.H.,2008. Pause insertions during cyclic in vivo loading affect bone healing.Clinical Orthopaedics and Related Research 466, 1232–1238.Geffre, C.P., Bliss, C.L., Szivek, J.A., DeYoung, D.W., Ruth, J.T., Margolis, D.S., 2008.Sensate scaffolds coupled totelemetrycan monitor in vivo loading fromwithinajoint over extended periods of time. Journal of Biomedical Materials ResearchPart B: Applied Biomaterials 84B, 263–270.Graichen, F., Arnold, R., Rohlmann, A., Bergmann, G., 2007. Implantable 9-channeltelemetry system for in vivo load measurements with orthopedic implants.IEEE Transactions on Biomedical Engineering 54, 253–261.Gurdjian, E.S., Lissner, H.R., 1944. Mechanism of head injury as studied bythe cathode ray oscilloscope — preliminary report. Journal of Neurosurgery 1,393–399.Lu, T.-W., Taylor, S.J.G., O’Connor, J.J., Walker, P.S., 1997. Influence of muscleactivity on the forces in the femur: an  in vivo  study. Journal of Biomechanics30, 1101–1106.Main, R.P., Biewener, A.A., 2006. In vivo bone strain and ontogenetic growthpatterns in relation to life-history strategies and performance in twovertebrate taxa: goats and emu. Physiological and Biochemical Zoology 79,57–72.Milgrom, C., Radeva-Petrova, D.R., Finestone, A., Nyska, M., Mendelson, S., Benjuya,N., Simkin, A., Burr, D., 2007. The effect of muscle fatigue on in vivo tibialstrains. Journal of Biomechanics 40, 845–850.Rabkin, B.A., Szivek, J.A., Schonfeld, J.E., Halloran, B.P., 2001. Long-term measure-ment of bone strain in vivo: the rat tibia. Journal of Biomedical MaterialsResearch Part B: Applied Biomaterials 58, 277–281.Robling, A.G., Hinant, F.M., Burr, D.B., Turner, C.H., 2002. Improved bone structureand strength after long-term mechanical loading is greatest if loadingis separated into short bouts. Journal of Bone and Mineral Research 17,1545–1554.Rubin, C.T., Sommerfeldt, D.W., Judex, S., Qin, Y.-X., 2001. Inhibition of osteopeniaby low magnitude, high-frequency mechanical stimuli. Drug Discovery Today6, 848–858.Schatzker, J., Sumner-Smith, G., Hoare, J., McBroom, R., 1980. A telemetric systemfor the strain gauge determination of strain in bone in vivo. Archives of Orthopaedic and Traumatic Surgery 96, 309–311.Swartz, S.M., Bertram, J.E.A., Biewener, A.A., 1989. Telemetered in vivo strainanalysis of locomotor mechanics of brachiating gibbons. Nature 342,270–272.Szivek, J.A., Magee, F.P.,1989. A long-term in vivo bone strain measurement device. Journal of Investigative Surgery 2, 195–206.Szivek, J.A., Roberto, R.F., Slack, J.M., Majeed, B.S., 2002. An implantable strainmeasurement system designed to detect spine fusion. Spine 27, 487–497.Szivek, J.A., Roberto, R.F., Margolis, D.S., 2005. In vivo measurements fromhardware and lamina during spine fusion. Journal of Biomedical MaterialsResearch Part B: Applied Biomaterials 75B, 243–250.Szivek, J.A., Bliss, C.L., Geffre, C.P., Margolis, D.S., DeYoung, D.W., Ruth, J.T., Schnepp,A.B., Tellis, B.C., Vaidyanathan, R.K., 2006. An instrumented scaffold canmonitor loading in the knee joint. Journal of Biomedical Materials ResearchPart B: Applied Biomaterials 79B, 218–228.Turner, C.H., Yoshikawa, T., Forwood, M.R., Sun, T.C., Burr, D.B., 1995. Highfrequency components of bone strain in dogs measured during variousactivities. Journal of Biomechanics 28, 39–44.Young, D.R., Howard, W.H., Orne, D., 1977. In-vivo bone strain telemetry inmonkeys ( M. nemestrina ). Journal of Biomechanical Engineering, 104–109. W.C. de Jong et al. / Journal of Biomechanics 43 (2010) 587–591  591
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks