Journal of Biomechanics 33 (2000) 765}770

Technical note
A continuous pure moment loading apparatus for biomechanical
testing of multi-segment spine specimens
John T. Lysack *, James P. Dickey
, Genevieve A. Dumas, David Yen
Department of Surgery, Division of Orthopaedics and Clinical Mechanics Group, Queen's University, Kingston, Ont., Canada

Department of Human Biology and Nutritional Sciences, University of Guelph, Guelph, Ont., Canada
Department of Mechanical Engineering and Clinical Mechanics Group, Queen's University, Kingston, Ont., Canada
Accepted 15 December 1999

Abstract

An apparatus is described that enables the application of continuous pure moment loads to multi-segment spine specimens. This
loading apparatus allows continuous cycling of the spine between speci"ed #exion and extension (or right and left lateral bending)
maximum load endpoints. Using a six-degree-of-freedom load cell and three-dimensional optoelectronic stereophotogrammetry,
characteristic displacement versus load hysteresis curves can be generated and analyzed for di!erent spinal constructs of interest.
Unlike quasi-static loading, the use of continuous loading permits the analysis of the spine's behaviour within the neutral zone. This
information is of particular clinical signi"cance given that the instability of a spinal segment is related to its #exibility within the
neutral zone. Representative curves for the porcine lumbar spine in #exion}extension and lateral bending are presented to illustrate
the capabilities of this system.  2000 Elsevier Science Ltd. All rights reserved.

Keywords: Spine; Flexibility testing; Loading apparatus; Pure moments; Continuous loads

1. Introduction properties of di!erent spinal constructs (e.g. intact, injured,

The rationale for using the #exibility method with
non-constraining pure moment loads for the in vitro
biomechanical testing of multi-segment spine specimens
was "rst presented more than a decade ago (Panjabi,
1988). The #exibility method of testing involves the con-
trolled application of known loads to a specimen while
measuring the resulting motion behaviour of the spine.
The use of non-constraining pure moments ensures that
the load experienced by a specimen remains constant
along its length independent of its geometry, motion, or
sti!ness (Asazuma et al., 1990; Ashman et al., 1989; Craw-
ford et al., 1995; Panjabi, 1988). This means that, through-
out the loading cycle, the loading conditions at any two
cross-sections in the spinal column are identical. A major
advantage of pure moment loading, therefore, is that it
allows for the unbiased comparison of the biomechanical
* Correspondence address: Clinical Mechanics Group, Apps Medical
Research Centre, Kingston General Hospital, Queen's University,
Kingston, Ont., Canada, K7L 2V7. Tel.: 001-613-548-2430; fax: 001-
613-549-2529.
E-mail address: 3jtl@qlink.queensu.ca (J.T. Lysack).
instrumented, or fused specimens) (Adams, 1995).
A number of devices designed to apply non-constrain-
ing pure moment loads to spine specimens have been
described in the literature. The original concept of using
deadweights and cables acting about pulleys to produce
force couples (Goel et al., 1985) has been re"ned with the
implementation of pneumatic actuators (Panjabi, 1994;
Yamamoto et al., 1989) or a materials testing machine
(Crawford et al., 1995) for load generation. More com-
plicated systems based on the sophisticated control of
torque motors (Kunz et al., 1994; Wilke et al., 1994) and
stepper motors (Eggli et al., 1991) have also been de-
scribed. Many of these designs apply incremental quasi-
static loads to the spine. While the quasi-static approach
yields useful information about the high-sti!ness elastic
zone, it does not provide data regarding the behaviour of
the spine within the clinically important neutral zone
(Panjabi, 1988).
The purpose of this technical note is to describe a rela-
tively simple, inexpensive apparatus which can produce
continuous angle-moment data for the entire range of
motion of the spine * including the neutral zone. The
apparatus applies pure moment loads to a multi-segment

0021-9290/00/$ - see front matter  2000 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 2 1 - 9 2 9 0 ( 0 0 ) 0 0 0 2 1 - X
766 J.T. Lysack et al. / Journal of Biomechanics 33 (2000) 765}770

spine specimen and continuously cycles it between speci-
"ed #exion and extension (or right and left lateral be-
nding) maximum load endpoints. The main advantage of
this approach is that it permits the analysis of the behav-
iour of the spine within the neutral zone. This is of
particular interest given that the clinical instability of
a spinal segment has been related to its behaviour within
this zone (Panjabi et al., 1994).
2. Materials and methods
The loading apparatus described here is similar in
principle to the one described by Crawford et al. (1995)
with modi"cations to allow for continuous cyclic loading
between speci"ed #exion and extension (or right and left
lateral bending) maximum load endpoints (Fig. 1). The
mechanism is based on a modular stainless steel frame
600 mm wide;600 mm deep;1000 mm high (Item
Products Inc., Houston TX). Mounted to the top of the
frame is a load generating electromechanical linear ac-
tuator (Model B8.5-T2-23S1, Dynact Inc., San Jose CA),
which is driven by a miniature high-performance micro-
stepper driver-indexer (Model Panther LE2-DE, Intelli-
gent Motor Systems Inc., Marlborough CT). A multi-
segment spine specimen can be mounted in the apparatus
using two stainless-steel pots and dental plaster. The
bottom pot rigidly "xes the caudal end of the spinal
segment to the base of the frame through a six-degree-of-
freedom load cell (Model MC3-6-250, Advanced Mech-

Fig. 1. Loading apparatus for the application of continuous pure moment loads to multi-segment spine specimens. This apparatus allows continuous
cycling of the spine between speci"ed #exion and extension (or right and left lateral bending) maximum load endpoints: (a) full view, (b) close-up of the
apparatus set for #exion}extension. The specimen is rotated 903 within the apparatus for lateral bending.
 J.T. Lysack et al. / Journal of Biomechanics 33 (2000) 765}770
Fig. 1. Continued.
anical Technology Inc., Watertown MA) while the top
pot holds the cephalad end of the spine. A constant
tension spring and cable system unloads the weight of the
top pot allowing the specimen to &#oat' unconstrained in
space. Two 200 mm diameter stainless-steel discs are
attached to the top pot. A series of thin, plastic coated
stainless-steel cables connects these discs to a #oating
crosshead and to four hanging counterweights via a sys-
tem of low friction pulleys and linear bearings. By mov-
ing the #oating crosshead upwards and downwards, the
actuator imparts a pure moment load (e.g. pure
#exion}extension moment) to the top pot and thus to the
spine. Control of the actuator is achieved using special-
ized software (Quickstep 2, Psi Software Inc., Mar-
lborough CT). The apparatus in this con"guration can
produce a maximum moment of 28 Nm at a loading
frequency of 0.5 Hz.
Biomechanical testing of "ve multi-segment (L2}L5)
adult porcine osseoligamentous lumbar spines was per-
formed using a non-destructive, #exibility method of
testing under non-constraining pure moment loads as
proposed by Panjabi (1988). Using the loading apparatus
described above, pure bending moments were applied to
the specimens to a maximum of 5 Nm in #exion}exten-
sion, and right}left lateral bending. These maximum
loads were chosen based on pilot data which showed a
linear angle}moment response beyond 5 Nm. The speci-
mens were cycled at a frequency of 0.1 Hz for a total of
767
"ve cycles * the "rst four cycles were considered precon-
ditioning and the "fth cycle was used for analysis. No
compressive preloads were applied, and approximately
5 min were allowed between tests for viscoelastic recovery.
Load data from the six-degree-of-freedom load cell
were ampli"ed using a strain gauge ampli"er system
(Model MCA6, Advanced Mechanical Technology Inc.,
Watertown, MA), and fed-back to the controller pro-
gram via a 24-line, parallel, digital I/O interface board
(PIO-12, Keithley Instruments Inc., Taunton, MA). This
feedback loop allowed for the automation of the
load}unload cycling by comparing the preset load cuto!s
to the real-time loads. The load data were collected and
stored using an Optotrak Analog Data Acquisition Unit
and associated software (Northern Digital Inc., Water-
loo, ON). Comparison of the real-time loads with the
preset load cuto!s, and the real-time display of the load
cell data were achieved using a custom Viewdac 2.20
program (Keithley Instruments Inc., Taunton, MA).
Vertebral displacement data were simultaneously col-
lected using a commercial infrared light-emitting diode
(IRED) optoelectronic stereophotogrammetry system
(Optotrak 3020 Position Sensor, Northern Digital Inc.,
Waterloo ON). Position data were collected for the six
IREDs which were rigidly attached to the two vertebral
bodies of interest (three IREDs on each of L3 and L4).
Prior to testing, vertebral landmarks were digitized to
de"ne the two local body coordinate systems relative to
768 J.T. Lysack et al. / Journal of Biomechanics 33 (2000) 765}770

the IREDs. The local body coordinate system, with its

origin at the centre of the vertebral body, was de"ned as
per Panjabi (1988, 1977).
The raw load cell data and the corresponding raw
IRED position data were processed using custom soft-
ware to yield, in six degrees of freedom, the loads at L3
and the relative displacements of L3 with respect to L4.
To account for the e!ect of o!-axis loads, the raw load
cell data, which were transduced at the caudal end of the
specimen, were transformed into the local body coordi-
nate system of L3 as performed by others (Cholewicki et
al., 1996; Dickey, 1998). Relative angles were calculated
as Cardan angles with the sequence X (#exion), Z (right
lateral bending), Y (left axial twist) (Tupling and Pier-
rynowski, 1987). Noise was reduced by passing the data
through a digital second-order Butterworth low-pass "l-
ter with a cuto! frequency of 5 Hz (Winter, 1979). Using
commercial software (Excel, Microsoft Corporation,
Redmond WA), these transformed load and displacement
data were processed to yield #exibility curves (angle
versus moment) for the #exion}extension and lateral
bending tests.

3. Results

The apparatus described in this paper was capable of

applying continuous and bi-directional pure moments to
multi-segment spine specimens in a highly reproducible
fashion (Fig. 2a). The purity of the applied moments was Fig. 2. Curves depicting the: (a) moments, and (b) forces transduced
evidenced by the small o!-axis loads * the maximum by the load cell at the caudal end of the porcine lumbar spine specimen
during #exion}extension testing. The maximum o!-axis moment is 7%
o!-axis moments remained less than 7% of the peak of the peak primary #exion}extension moment (a), and the o!-axis
primary moment (Fig. 2a), and the o!-axis forces were forces remain less than $8 N (b). These are the "nal three of the "ve
always less than $8 N (Fig. 2b). Although small, these cycles reported in the text.
o!-axis loads could be seen to have a noticeable e!ect on
the actual loading conditions at the spinal level of inter-
est (L3) (Fig. 3). Transformation of the primary trans-
duced moment data into the local body coordinate
system of interest was necessary to determine the true
local body loading condition.
The apparatus described in this paper was able to
produce continuous data for the spine throughout its
range of motion * including the high-#exibility neutral
zone region. Highly reproducible characteristic hysteresis
#exibility curves were obtained by plotting the relative
angular displacement of L3 with respect to L4 versus
moment at L3 (Fig. 4). These #exibility curves allow the
biomechanical properties of the spine to be determined
* including the clinically important neutral zone prop-
erties. Fig. 3. The applied #exion}extension moment data as transduced by
the load cell compared with the #exion}extension moment data after
transformation into the local body coordinate system of L3. Note the
4. Discussion substantial di!erence between the transduced and transformed mo-
ments particularly at low loads (i.e. within the neutral zone). Trans-
formation of the primary transduced moment data into the local body
The apparatus described in this paper applies pure coordinate system is required to determine the true loading condition at
moment loads to multi-segment spine specimens L3. These are the "nal three of the "ve cycles reported in the text.
 J.T. Lysack et al. / Journal of Biomechanics 33 (2000) 765}770
Fig. 4. A typical set of #exibility curves for the L3}L4 porcine spine
subjected to: (a) #exion}extension, and (b) right}left lateral bending
moments demonstrates excellent reproducibility. Preconditioning
cycles (three and four for #exion}extension and four only for lateral
bending) are shown in gray and the "fth cycle is overlain in black.
continuously and bi-directionally about the neutral posi-
tion. The advantage of this approach is that it permits the
analysis of the behaviour of the spine within the high-
#exibility neutral zone region. This is of particular inter-
est since the clinical instability of a spinal segment has
been related to its behaviour within the neutral zone
(Panjabi et al., 1994).
The limitations of this apparatus are encountered
when testing very long or unstable spine specimens. In
these cases, the o!-axis moments may increase substan-
tially secondary to the coupled behaviour of the spine.
The apparatus has been designed to minimize this e!ect,
however, by allowing the specimen to #oat relatively
unconstrained in space. For example, during lateral be-
nding testing, the specimen naturally tends to #ex for-
ward as the lateral bending moment increases. This
coupled #exion motion causes the cephalad end of the
specimen to translate anteriorly. The low-friction linear
bearings mounted to the frame of the apparatus allow the
crosshead, cables, and top pot to translate anteriorly with
little opposition. In this example, an o!-axis posterior
shear force and extension moment are thereby mini-
769
mized. Furthermore, the e!ect of these o!-axis loads are
accounted for by transforming the raw load cell data
(which are transduced at the caudal end of the specimen)
into the local body coordinate system of interest.
Many devices which have been designed to impart
pure moment loads to spinal specimens do not have the
capacity to monitor o!-axis loads. Often, the purity of
the moments is either assumed based on the mass of
hanging weights (Goel et al., 1985; Panjabi, 1988), or is
calculated from uni-axial load cell data (Chiba et al.,
1996; Kotani et al., 1996). The use of a six-degree-of-
freedom load cell to monitor the o!-axis loads is there-
fore arguably a signi"cant advantage of the present ap-
proach in itself.
To the best of our knowledge, there exists no data on
the o!-axis loads produced by continuous pure moment
loading devices with which to compare the performance
of our apparatus. In one study using uni-directional quasi-
static pure moments (Eggli et al., 1991), the peak o!-axis
moments were as high as 25% of the primary #exion
moment (o!-axis forces were not reported). In another
study involving combined #exion-compression loading
(Yoganandan et al., 1995), the peak o!-axis moments reach-
ed 8% of the primary #exion moment and the peak o!-axis
shear forces were as high as 27.5 N. The corresponding
values for the present study of 7% and 8 N demonstrate
a similar performance regarding o!-axis loads.
In conclusion, the primary advantage of this loading
apparatus is its capability of producing automated con-
tinuous cyclic pure moment loading between speci"ed
#exion and extension (or right and left lateral bending)
maximum load endpoints. This allows the plotting of
complete hysteresis curves from which the characteristic
#exibility parameters of interest can be calculated for
each multi-segment spine specimen. The ability to calcu-
late neutral zone parameters such as neutral zone #exibil-
ity and laxity angle (Tencer et al., 1995) is of particular
signi"cance given that the clinical instability of a spinal
segment is related to its behaviour within the neutral
zone (Panjabi et al., 1994).
Acknowledgements
The authors gratefully acknowledge: Gerald Saunders,
David Siu, and Julie Howes of the Clinical Mechanics
Group, Department of Surgery, Queen's University,
Kingston, Ontario, for assistance in the development of
the testing apparatus and associated software; Xunhua
Yuan, Department of Orthopaedics, University Hospital,
Lund, Sweden, for assistance with data collection and
analysis; and Hye Won Yang, independent artist, for
producing the graphical illustrations of the apparatus.
This research has been supported by NSERC Research
Grant OGP0006858, NSERC Equipment Grant
EQP0157230, and PSI Grant 95-36.
770 J.T. Lysack et al. / Journal of Biomechanics 33 (2000) 765}770
References
Adams, M.A., 1995. Mechanical testing of the spine. An appraisal of
methodology, results, and conclusions. Spine 20, 2151}2156.
Asazuma, T., Stokes, I.A., Moreland, M.S., Suzuki, N., 1990. Interseg-
mental spinal #exibility with lumbosacral instrumentation. An in
vitro biomechanical investigation. Spine 15, 1153}1158.
Ashman, R.B., Bechtold, J.E., Edwards, W.T., Johnston, C.E., II,
McAfee, P.C., Tencer, A.F., 1989. In vitro spinal arthrodesis implant
mechanical testing protocols. Journal of Spinal Disorders 2,
274}281.
Chiba, M., McLain, R.F., Yerby, S.A., Moseley, T.A., Smith, T.S.,
Benson, D.R., 1996. Short-segment pedicle instrumentation. Bi-
omechanical analysis of supplemental hook "xation. Spine 21,
288}294.
Cholewicki, J., Crisco, J.J., III, Oxland, T.R., Yamamoto, I., Panjabi,
M.M., 1996. E!ects of posture and structure on three-dimensional
coupled rotations in the lumbar spine. A biomechanical analysis.
Spine 21, 2421}2428.
Crawford, N.R., Brantley, A.G.U., Dickman, C.A., Koeneman, E.J.,
1995. An apparatus for applying pure nonconstraining moments to
spine segments in vitro. Spine 20, 2097}2100.
Dickey, J.P., (1998) Anatomy and mechanics of the human and porcine
lumbar interspinous ligament. Ph.D. Thesis, Queen's University,
Kingston, Ontario, Canada.
Eggli, S., Schlapfer, F., Schneider, E., Aebi, M., 1991. Assess-
ment of a new load test device for measuring the rigidity of the
thoraco-lumbar "xation systems. Orthopaedic Transactions 15,
371.
Goel, V.K., Goyal, S., Clark, C.R., Nishiyama, K., Nye, T.A., 1985.
Kinematics of the whole lumbar spine. E!ect of Discectomy. Spine
10, 543}554.
Kotani, Y., Cunningham, B.W., Cappuccino, A., Kaneda, K., McAfee,
P.C., 1996. The role of spinal instrumentation in augmenting lum-
bar posterolateral fusion. Spine 21, 278}287.
Kunz, D.N., McCabe, R.P., Zdeblick, T.A., Vanderby Jr., R., 1994.
A multi-degree of freedom system for biomechanical testing. Jour-
nal of Biomechanical Engineering 116, 371}373.
Panjabi, M.M., 1977. Experimental determination of spinal motion seg-
ment behavior. Orthopaedic Clinics of North America 8, 169}180.
Panjabi, M.M., 1988. Biomechanical evaluation of spinal "xation devi-
ces: I. A conceptual framework. Spine 13, 1129}1134.
Panjabi, M.M., 1994. Lumbar spine instability. A biochemical chal-
lenge. Current Orthopaedics 8, 100}105.
Panjabi, M.M., Lydon, C., Vasavada, A., Grob, D., Crisco, J.J., III,
Dvorak, J., 1994. On the understanding of clinical instability. Spine
19, 2642}2650.
Tencer, A.F., Hampton, D., Eddy, S., 1995. Biomechanical properties of
threaded inserts for lumbar interbody spinal fusion. Spine 20,
2408}2414.
Tupling, S.J., Pierrynowski, M.R., 1987. Use of cardan angles to locate
rigid bodies in three-dimensional space. Medical and Biological
Engineering and Computing 25, 527}532.
Wilke, H.J., Claes, L., Schmitt, H., Wolf, S., 1994. A universal spine
tester for in vitro experiments with muscle force simulation. Euro-
pean Spine Journal 3, 91}97.
Winter, D.A., 1979. Processing of raw kinematic data. In: Biomechanics
of Human Movement. Wiley, Toronto, pp. 25}39.
Yamamoto, I., Panjabi, M.M., Crisco, T., Oxland, T., 1989. Three-
dimensional movements of the whole lumbar spine and lumbosacral
joint. Spine 14, 1256}1260.
Yoganandan, N., Cusick, J.F., Pintar, F.A., Droese, K., Voo, L., 1995.
An experimental technique to induce and quantify complex cyclic
forces to the lumbar spine. Neurosurgery 36, 956}964.