7/30/2019 Gait research
1/13
This article was downloaded by: [HINARI]On: 10 November 2011, At: 05:47Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House37-41 Mortimer Street, London W1T 3JH, UK
ErgonomicsPubl icat i on detai l s, i nc luding instr uct ions for authors and subscript ion infor mat i on:h t t p : / / www. tand fonl ine. com/ lo i / t e rg20
Balance control dur ing lateral load t ransfers over a
slippery surfaceRobert D. Caten a
a, Angela DiDomenico
b, Jacob J. Banks
b& Jack T. Dennerl ein
cd
aDepart ment of Physical Therapy, University of Evansvil l e, Evansvil l e, IN, USA
bCent er f or Physical Ergonomi cs, Liber t y Mut ual Research Inst it ut e for Safet y, Hopkint on,
MA, USAc
Department of Environmental Health, Harvard School of Public Health, Boston, MA, USAd
Departm ent of Ort hopaedic Surgery, Brigham and Women's Hospit al and Harvard MedicaSchool , Bost on, MA, USA
Available online: 25 Oct 2011
To cite t his art icle: Robert D. Caten a, Angela DiDomeni co, Jacob J. Banks & Jack T. Dennerl ein (2011): Balance controlduring lateral load transfers over a slippery surface, Ergonomics, 54:11, 1060-1071
To link t o thi s arti cle: ht t p : / / dx .doi .org/ 10.1080/ 00140139.2011.618229
PLEASE SCROLL DOWN FOR ARTICLE
Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditionsThis article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form toanyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses shouldbe independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims,proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly inconnection with or arising out of the use of this material.
http://www.tandfonline.com/page/terms-and-conditionshttp://dx.doi.org/10.1080/00140139.2011.618229http://www.tandfonline.com/page/terms-and-conditionshttp://dx.doi.org/10.1080/00140139.2011.618229http://www.tandfonline.com/loi/terg207/30/2019 Gait research
2/13
Balance control during lateral load transfers over a slippery surface
Robert D. Catenaa
, Angela DiDomenicob
*, Jacob J. Banksb
and Jack T. Dennerleinc,d
aDepartment of Physical Therapy, University of Evansville, Evansville, IN, USA; bCenter for Physical Ergonomics, Liberty MutualResearch Institute for Safety, Hopkinton, MA, USA;
cDepartment of Environmental Health, Harvard School of Public Health, Boston,
MA, USA;d
Department of Orthopaedic Surgery, Brigham and Womens Hospital and Harvard Medical School, Boston, MA, USA
(Received 5 November 2010; final version received 23 August 2011 )
Few studies have measured balance control during manual material handling, and even fewer with environmentalcofactors. This study examined the effect of different surface frictions during a stationary manual material handlingtask. Thirty-six healthy participants completed 1808 lateral transfer tasks of a load over high- and low-frictionsurfaces (m 0.86 and m 0.16, respectively). Balance measures, stance kinematics and lower extremity muscleactivities were measured. Success during the novel slippery surface dichotomised our population, allowing us toinvestigate beneficial techniques to lateral load transfers over the slippery surface. Stance width reduction by 8 cmand 158 of additional external foot rotation towards the load were used to counter the imbalance created by theslippery surface. There was no clear alteration to lower extremity muscular control to adapt to a slippery surface.
Changes in stance seemed to be used successfully to counter a slippery surface during lateral load transfers.
Statement of Relevance: Industries requiring manual material handling where slippery conditions are potentiallypresent have a noticeable increase in injuries. This study suggests stance configuration, more so than any othermeasure of balance control, differentiates vulnerability to imbalance during material handling over a slipperysurface.
Keywords: falls; balance control; manual materials handling; slips; friction
Introduction
According to the US Bureau of Labor Statistics (BLS
2009), there is a particularly high rate of falls in
industries requiring manual material handling, such as
delivery personnel and transportation cargo handlers.Fall incidence rates are about 100% higher in ground
freight transportation compared with ground passen-
ger transportation occupations. Activities within man-
ual material handling industries require a variety of
movements and manoeuvres with a load. Balance and
fall related research has primarily addressed biome-
chanical aspects of walking during load-carrying
situations (Holbein-Jenny et al . 2007, Birrell and
Haslam 2009, Gillette et al. 2010). Load transfer
manoeuvres have not been examined extensively as a
challenge to balance control and we are not aware of
any detailed epidemiological data to indicate the risk
of different load transfer manoeuvres. The previous
balance control studies that have examined load
transfers have emphasised sagittal plane reaching
(Kozak et al. 2003, Row and Cavanagh 2007, Liao
and Lin 2008) and lifting (Commissaris and Toussaint
1997, Toussaint et al. 1998, Kollmitzer et al. 2002),
where the centre of gravity (COG) is not expected to
excessively deviate towards the lateral directions
(Kollmitzer et al. 2002).
Our research group has recently endeavoured to
examine the lateral load transfer manoeuvre. This task
requires the acquisition, movement and placement ofa load from one side of the body to the other. This is
similar to tasks performed by air and truck cargo
handlers and deliverers. The body COG is required to
travel closer to the sides of the base of support (BOS)
as load distance is increased during such a manoeuvre.
Relocation of the COG near the edges of the BOS
may be an important cause of imbalance during
manual material handling (Streepey and Angulo-
Kinzler 2002). Our first examination (Catena et al.
2010) focussed on the effects of weight during the
lateral load transfer. We found that increased weight
led to greater vulnerability to a loss of balance.
Individuals in turn constricted movement about the
ankle joints by stiffening the joints through
co-contraction of antagonist shank musculature.
Stiffening the bodys posture is a commonly used
conservative adaptation to environmental
perturbations (Santello and McDonagh 1998, Lark
et al. 2003, Nielsen et al. 2004).
*Corresponding author. Email: [email protected]
Ergonomics
Vol. 54, No. 11, November 2011, 10601071
ISSN 0014-0139 print/ISSN 1366-5847 online
2011 Taylor & Francis
http://dx.doi.org/10.1080/00140139.2011.618229http://www.tandfonline.com
7/30/2019 Gait research
3/13
Besides highlighting the importance of co-contrac-
tion to control balance, these previous studies highlight
the need to examine the effects of a variety of
environmental factors (e.g. icy surface, unstable loads
or distracting attention) in balance control, as not all
factors are countered with the same control mechan-
isms. This remains true for occupational tasks such asthe lateral load transfer, where there has been limited
research involving environmental factors such as
reduced friction. Load transferring in slippery condi-
tions has focussed more on basic tasks such as walking
(Myung and Smith 1997) and simple grasping tasks
(Cooper et al. 2005), as opposed to lateral movements
with the load. One study (Holbein and Redfern 1997)
suggests that holding loads to the side in more slippery
conditions probably leads to choosing a more con-
servative posture, by reducing the travel area of the
COG within the BOS. The more dynamic task of
walking with a load in slippery conditions has been
shown to result in imbalance (Myung and Smith 1997,Cham and Redfern 2002).
While imbalance is certainly more prevalent in
walking, since the COG has greater motion and is
more prevalently located outside of the BOS, a fall
during walking is commonly in the anterior or poster-
ior (A/P) direction (Bakken et al. 2006). A fall in the A/
P direction has easily employed countermeasures such
as stepping manoeuvres, joint flexions and bracing by a
leg and arms (Bakken et al. 2006). Imbalance in a
medial or lateral direction does not permit the same
countermeasures (Lo and Ashton-Miller 2008). Step-
ping manoeuvres to counter a fall to same side as the
loaded leg are blocked by the loaded leg requiringrotation of the entire body around it. Lower extremity
lateral joint movements are limited to hip and ankle
abductions and bracing for impact after a fall is limited
to the arm on the side of the lateral fall. A fall in the A/
P direction can, however, be more easily countered
during lateral movements of the body. Stepping
manoeuvres could be accomplished with either foot
(depending on the distribution of the body weight over
each foot). During this stepping manoeuvre, all lower
extremity joints are available to help decelerate the
body. If the step does not succeed in preventing a fall,
then both arms are available to help brace for impact
with the ground. Occupational tasks that require
lateral motions of the whole body must be examined
more closely if we hope to get a full understanding offalls in the workplace.
In this study, we evaluated how balance control
during lateral load transfers was affected by the
friction of the standing surface. The first attempt on a
novel slippery surface was of particular interest. Based
upon our previous analyses of weight during the lateral
load transfer, we hypothesised that rather than
demographic characteristics or kinematic factors,
inappropriate muscular control would be the
predominant factor in loss of balance. Secondly, we
sought to describe how individuals adapt to a slippery
surface given enough practice. We hypothesised that
muscular activity in the lower extremities might bealtered to control imbalance sufficiently. Findings from
this research will demonstrate how a slippery surface
can be negotiated safely when a lateral load transfer is
required and how individuals might alter normal
lateral load transfer performance when a slippery
surface is not necessarily expected, but occasionally
present.
Methods
Participants
Thirty-six healthy working age (2066 years of age)
adults without material handling experienceparticipated in this study (Table 1). Informed consent
was approved by the Institutional Review Boards of
both Harvard School of Public Health and Liberty
Mutual Research Institute for Safety. The informed
consent was read and signed by each participant prior
to data collection.
Participants wore a tank top, shorts, below-ankle
socks and below-ankle hiking shoes (Nike Bandolier
Table 1. Demographic variables across the two groups: those who successfully completed LCOF-1 (SL-1) and those who wereunsuccessful (UL-1).
Variable SL-1 UL-1 p-value
Gender: female, male 8, 15 6, 7 0.501Age (years): mean (SD) 44.6 (14.8) 43.0 (15.9) 0.871Height (m): mean (SD) 1.70 (0.10) 1.68 (0.08) 0.598Weight (kg): mean (SD) 77.6 (17.7) 70.6 (11.0) 0.150BMI (kg/m2): mean (SD) 26.6 (4.0) 25.0 (3.4) 0.202Arm length (% body height): mean (SD) 34.9 (1.1) 35.3 (1.2) 0.353Pelvis width (% body height): mean (SD) 15.4 (1.6) 15.2 (1.4) 0.654Hand dominance: right, left 18, 5 11, 2 0.644Foot dominance: right, left 20, 3 10, 3 0.438
Ergonomics 1061
7/30/2019 Gait research
4/13
II) for instrumentation purposes. All shoes were
purchased at the beginning of this study and were
only worn for this study; therefore, the tread was that
of brand new hiking shoes. A harness connected to an
overhead fall arrest system was worn during box
transfer tasks for safety purposes. The harness was
adjusted to allow for unobstructed side-to-side motion.Slack in the harness cable was adjusted to provide
resistance only when needed to catch the person from
falling (Figure 1).
Procedure
Each participant completed 1808 lateral transfers of a
41.5 (width) cm 6 41.5 (depth) cm 6 32.5 (height) cm
box loaded to 5% (+ 0.01 kg) of their body weight
with semi-oval, reinforced cut-out handles. Two
different flooring surfaces were presented to the
participants. A high friction surface (m 0.86) was
created between the rubber shoe soles and plywoodflooring painted with a granular white paint high
coefficient of friction surface (HCOF). A low friction
surface (m 0.16) was created between Teflon tapeadhered to the bottom of the shoes and Teflon flooring
low coefficient of friction surface (LCOF). The two
surfaces were the same colour and had a similar gloss
so that they were nearly indistinguishable. The surface
and shoe modification was performed in such a way as
to not alert the participants about the change.
The task required the participants to transfer the
box from one pedestal to the other, whose heights wereadjusted such that the boxs handles were at the
participants standing elbow height. The pedestals top
surfaces were 11.5 cm wider and longer than the box
and covered with perforated rubber matting, which
encouraged the participants to lift rather than slide the
box off the surface. Prior to data collection, each
participant was given the opportunity to practice the
lateral load transfer over the high friction surface.
Participants were instructed to perform the transfer at
a self-selected pace. The purpose of this practice
session was to familiarise them with the process and
instructions. At the same time, we asked that they
determine their self-imposed maximum transferdistance that could be completed safely while still
adhering to our instructions.
For consistency, specific instructions on how to
complete the task were provided to all participants.
Figure 1. The start position for a HCOF trial. The objective is for the participant to transfer the box from its pedestal tothe opposing pedestal without walking. A researcher monitors the feet from behind during the trial. Lasers on the floor mark thestarting stance configuration. This figure has been adopted from Catena et al. (2010) with written consent from the authors andpublisher.
1062 R.D. Catena et al.
7/30/2019 Gait research
5/13
These were: (1) you are not allowed to lift a foot
completely off of the ground, drag your feet or slide
your lead foot, but you are allowed to pivot on your
feet. (2) When lifting the box, make sure you have your
little fingers in the handles before lifting, dont drag the
box, dont rest your weight on it, and dont push
yourself up by pushing into it. (3) When placing thebox, make sure that the box is completely on the
pedestal and again dont push against the box to push
yourself up. The final purpose of the practice sessions
was for the participants to determine their preferred
starting stance configuration. The starting stance
configuration was marked. After each collected trans-
fer trial, the participant would be instructed to return
to the starting stance configuration.
Data collection started with six successful trials
performed at the participants practiced maximum
distance over a HCOF. While the participants were
sequestered away from the testing area, researchers
modified the flooring surface to the Teflon surface.When the participants returned to the testing area they
put on the modified shoes, which were not slippery
unless in contact with the Teflon surface. The second
test condition required a single transfer of the box at
the same distance and starting stance configuration
over the new and unidentified LCOF-1. There were no
practice trials for this condition. The third condition
started after adequate familiarisation and practice with
the modified surface. This practice session included
modifications to the transfer distance and starting
stance configuration as deemed necessary by the
participant. Once the modifications and practice were
completed, the new start stance configuration wasmarked and the participants completed six successful
transfer trials LCOF-2. The conditions were not
randomised because the novelty of the first low friction
condition and adaptation to the slippery surface were
of particular interest.
Measurements and parameters
All COFs were measured with a Brungraber Mark II
tribometer (Chang et al. 2003). Testing for the high
friction condition involved a cut section of a shoe sole
against the plywood floor surface. Testing for the low
friction condition involved a cut section of a shoe sole
with Teflon taped to the bottom against the Teflon
floor surface.
A load cell (Scaime C9418) was located underneath
the top of each pedestal and sampled at 1000 Hz to
capture the timing of the box transfer. The gradual
acquisition and release of the box (per cent of box
weight on pedestal) was measured from these data,
allowing for calculations of the total time to complete
the box transfer and inclusion of the box weight into
the body centre of mass (COM). The distance the box
was transferred was calculated from the displacement
of the box COM as measured by the motion analysis
system described below, and normalised to the stature
of the participant.
Three-dimensional marker data were collected by a
12-camera motion capture system (MotionAnalysis,Santa Rosa, CA). Thirty-four markers were placed on
the body to identify 13 individual body segments,
similar to previous marker sets used for measuring
balance control (Hahn and Chou 2004, Catena et al.
2009, Catena et al. 2010). Eight markers (four for each
foot) attached to the shoes at the most anterior,
posterior, medial and lateral projections of the sole
allowed us to measure elevation of each part of the
foot off of the ground. These same markers were used
for BOS measurements. Marker data were collected at
100 Hz for a time sufficient for the participant to
complete the task and filtered with a zero-lag forth
order low pass Butterworth filter at 8 Hz.Thirteen body segment COMs were calculated
based on anthropometric data of segment COM
locations and weights from previous research using the
appropriate age, gender and body mass index
information for each participant (Plagenhoef et al.
1983, Winter 1990, de Leva 1996, Pavol et al. 2002,
Durkin and Dowling 2003). The weighted sum of
segment COMs, harness and electromyography
(EMG) transmitter determined the whole body COM
at each time frame for the entire trial. The weight of
the box was proportionally accounted for in the COM
calculation as load cell measures in each pedestal
changed.The BOS was calculated from the four markers on
each foot. As a portion of a foot was raised, BOS sides
connected to the corresponding marker were
disregarded. If a foot was raised entirely off the ground
then that trial was excluded from our analyses. Stance
kinematics was measured through calculations of
displacement of the ankle joint centres in the lateral
direction (stance width) and the external rotation of
the support and contralateral feet with respect to pelvis
rotation (combination of foot and hip external
rotation). The minimum distance of the COG to each
of the six edges (front, back and lateral forefoot and
rear foot of the right and left feet) of the BOS (E1:
where x2, y2 and x1, y1 are coordinates of adjacent BOS
markers and x0, y0 are coordinates for the COG at a
single point in time) was used to indicate balance
control. Variability in balance control was determined
by measuring the total area of COG path (estimated as
an oval using the most A/P and M/L points)
normalised to the area of the BOS and the COM path
distance during the transfer normalised by the
displacement of the box during the transfer.
Ergonomics 1063
7/30/2019 Gait research
6/13
dx2 x1y1 y0 x1 x0y2 y1ffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffiffiffi
x2 x12 y2 y1
2q : E1
Bilateral activity of the biceps femoris, vastus
medialis, gastrocnemius and tibialis anterior were
measured by a telemetry electromyography system(Noraxon, Scottsdale, AZ). All electrode placements
were in accordance with those described in the literature
(Hermens et al. 1999) and confirmed via palpation and
observing the EMG signal during isometric contrac-
tions. The skin was prepared beforehand by treatment
with alcohol and gentle abrasion. One author was
responsible for all skin preparation and electrode
placement for all participants during this study.
Data from maximum voluntary contractions
(MVCs) were collected to normalise the amplitude of
the EMG signal. A MVC for 5 s was collected while
the lower limb was manually restrained by a researcher
at a specific joint angle (08 ankle plantar flexion forgastrocnemius and tibialis anterior; 308 knee flexion
for biceps femoris; 158 knee flexion for vastus
medialis). This was performed three times for each
muscle with 1 min breaks in between.
All EMG signals were pre-amplified, band passed
filtered (10500 Hz), sampled at 1000 Hz and synchro-
nised with motion capture data. EMG signals were
rectified and smoothed using a 100-ms moving window
about each frame. The signals collected during the
lateral transfers were normalised by the maximum
signal amplitude measured during the MVC protocols.
The eight muscle activities were grouped into four
agonistantagonist pairs: the support and contralateralshank (gastrocnemius and tibialis anterior) and the
support and contralateral thigh (biceps femoris and
vastus medialis). The support leg was defined as the leg
on the ipsilateral side of the load at a particular time
and the contralateral leg was opposite of the support
leg. For each pair, we calculated the average intensity
(average muscle activity of the segment muscles) and
co-contraction percentage (lesser muscle activity as a
percent of greater muscle activity) of agonistantago-
nist muscle pairs for 800 ms time windows around
acquisition and release of the box to determine support
leg muscle activity and muscle activity of the contral-
ateral leg. We previously used this method to measure
co-contraction activity and percentage separately (Ca-
tena et al. 2010). Our separate equations are derived
from the single encompassing equation proposed by
Winter (1990).
Statistical analyses
This study initially had one independent variable
(task), which had three levels [high friction (HCOF),
novel low friction (LCOF-1), and practiced low
friction (LCOF-2)]. However, our balance control
variables were unobtainable for LCOF-1 in a group of
13 individuals who were unable to successfully
complete the first slippery trial. Therefore, we
dichotomised groups by successful (SL-1) vs.
unsuccessful (UL-1) completion of LCOF-1. Fouranalyses were performed: (1) group comparison, (2)
between-group comparison during HCOF, (3) within-
group comparison of LCOF-1 and HCOF in the SL-1
group and (4) a group-by-task comparison with the
two groups and HCOF and LCOF-2 tasks.
All assumptions for analysis of variance (ANOVA)
were considered appropriate except for muscle activity
variables and age, which displayed non-parametric
distribution. Muscle activity was examined using the
KruskalWallis test for group comparisons, Savage
one-way analyses for task comparison main effects and
Wilcoxon two-sample tests for task pair-wise
comparisons in SAS (Cary, NC). Age was analysedwith a MannWhitney test in SPSS 12.0 (Chicago, IL).
Parametrically distributed stance and balance
variables were examined using the Proc MIXED
function of SAS with age as a covariate to fit
ANCOVA models with subject as a random effect.
Pair-wise comparisons with Bonferroni adjustments
were performed when statistical significance was
calculated from main effects. Analyses performed on
group demographics and the first slippery trials were
analysed with chi-squared analyses and t-tests, when
respectively appropriate, in SPSS 12.0. Alpha levels for
all statistical analyses were set a priori at 0.05.
Results
First slippery trial analyses
Group comparisons
By analysing performance during the first slippery
trial, we hoped to determine how balance is affected by
lateral load transfer over a novel slippery surface. Of
the 36 participants, 13 of them were unable to
successfully complete the first slippery trial. Two
individuals fell (fall arrested by harness) during their
first trial with the slippery surface (Figure 2). Six
individuals took a step to maintain balance during
their first trial with the slippery surface. The five
remaining individuals did not lose balance in any way,
but chose to stop the trial after first attempting it,
stating that they believed they would be unable to
complete the task given the instructions described
previously without falling. This group of 13 individuals
made up the unsuccessful during first low friction trial
(UL-1) group. We found no significant differences
between genders, hand dominance, foot dominance,
1064 R.D. Catena et al.
7/30/2019 Gait research
7/13
age, height, weight, body mass index (BMI), pelvis
width normalised by height or arm length normalised
by height for the two groups (Table 1).
Between-group comparison during HCOF
Our next attempt to determine what differentiatedgroup success examined initial performance of
individuals in the HCOF trials. By doing this, we
hoped to learn what tactics each group was initially
employing and would supposedly continue to employ
if unaware of work environment changes. Independent
sample t-test results indicated that there were no
differences between groups in how far the box was
transferred or the transfer time for HCOF trials.The UL-1 group started with an 8.3-cm wider
stance than SL-1 (p 5 0.001). Stance width remained
statistically wider for the UL-1 group throughout the
load transfer (Figure 3). Table 2 shows the results only
for start and load acquisition times, but results were
similar at load release and the end of the trial. While
the two groups did not differ significantly in the
amount of foot external rotation that they initially
started with, the UL-1 group used 138 less external
rotation when they manoeuvred to acquire the box
compared with the SL-1 group (p 5 0.001). During
this manoeuvre, the load side foot rotates out about
the hip and then the pelvis rotates about thelongitudinal body axis, and often past that of the
externally rotated foot, but does even more so in the
UL-1 group as this group keeps the foot in a relatively
static position compared with SL-1 individuals. The
UL-1 group did not contract agonistantagonist
muscles as uniformly across lower extremity joints
(co-contraction) as the SL-1 group (Table 3); however,
Figure 2. Example raw data of a LCOF-1 trial for an UL-1individual transferring a box from right to left. Top view ofthe COG movement throughout the trial within the BOS.The BOS shown is specific to the instance of pickup. The symbol shows the COG location at pickup. Notice how theCOG quickly trails straight backwards at the end of the trial.This individual failed due to a loss of balance backward atthe end of the trial that resulted in a fall into the harness.
Figure 3. The average BOS transition from start (A) to pickup (B) of the box between SL-1 (black) and UL-1 (grey) forthe HCOF trials. Stance width for the two groups is indicated in the middle of the feet and the difference between lead footrotation angles is indicated between the dashed lines.
Ergonomics 1065
7/30/2019 Gait research
8/13
the UL-1 group did use more agonistantagonist
muscular intensity at the support thigh (p 0.006),contralateral thigh (p 5 0.001) and contralateral
shank (p 0.046).Ultimately, only a few outcome balance measures
proved to be different between the two resultant
groups in the HCOF condition (Table 4). We did
find that the COG of the UL-1 group travelled 1.4 cm
closer to the edge of their lateral-rear foot side of the
BOS during box acquisition compared with the SL-1
group (p 0.037). On the other hand, the UL-1 groupstayed 1.2 cm further away from the front boundary of
their BOS compared with the SL-1 group (p 0.024).
Within-group comparison of LCOF-1 and HCOF
in the SL-1 group
Our final attempt to determine what differentiated
group success examined initial performance of
successful individuals in the HCOF trials compared
with their performance during LCOF-1. By doing this,
we hoped to learn whether initial tactics used by this
group either continued to be used with success or were
abandoned in favour of more appropriate tactics
specifically for the slippery surface. Box transfer
distance and start stance properties did not change
from HCOF to LCOF-1 because they were held
constant between tasks to emulate unawareness to a
change in the work environment.
While there were no differences in start stance,
there was a change in stance almost immediately as the
individuals manoeuvred to pick the load up (Table 2).
There was over a 2-cm increase in stance width at load
acquisition in LCOF-1 compared to HCOF
(p 5 0.001), which became a 5-cm increase by the
end of the trial in LCOF-1 (p 0.037). During LCOF-1, the SL-1 group also rotated their feet more with the
pelvis as it rotated towards the direction of the pedestal
Table 2. Stance kinematic variables across the three conditions.
Variable Group HCOF LCOF-1 LCOF-2
Stance width at start (cm) SL-1 54.4 (9.9)B 55.5 (10.4) 51.6 (13.7)UL-1 62.7 (17.1)A X 54.9 (17.7)Y
Stance width at load acquisition (cm) SL-1 60.4 (8.1)B Y 63.0 (8.5)X 63.3 (10.8)X
UL-1 65.7 (16.0)A 62.8 (15.8)
External foot rotation* of lead foot at start (degree) SL-1 15.94 (9.03) 14.86 (10.8) 16.78 (9.83)UL-1 18.49 (15.0) 19.34 (14.4)External foot rotation* of lead foot at load acquisition (degree) SL-1 716.84 (22.3)A Y 75.90 (14.2)X 72.58 (13.6)X
UL-1 729.89 (23.6)B 712.15 (16.6)X
Notes: *Relative to perpendicular position of pelvis ASIS; A group 4 B group during the specified task (p 5 0.05); X task 4 Y task inthe specified group (p 5 0.05). Significant pairwise comparison results are presented with superscript letters and shaded. Values are given as mean(SD.)
Table 3. Agonistantagonist muscle pair co-contraction parameters over the 800 ms window for the support leg andcontralateral leg.
Variable Group HCOF LCOF-1 LCOF-2
Support shank co-contraction intensity{ SL-1 53.7 (24.7)A 55.1 (24.4) 52.3 (24.7)A
UL-1 49.5 (23.5)B
49.3 (26.8)B
Support shank average intensity{ SL-1 26.5 (13.0) 28.7 (13.2) 25.8 (12.2)UL-1 29.5 (24.9) 35.6 (33.5)
Support thigh co-contraction intensity{ SL-1 50.2 (24.8)A 54.7 (25.0) 50.2 (22.8)A
UL-1 44.4 (22.8)B 47.2 (26.3)B
Support thigh average intensity{ SL-1 21.0 (13.2)B Y 24.1 (14.2) 23.6 (13.2)B X
UL-1 24.2 (13.1)A Y 27.1 (17.2)A X
Contralateral shank co-contraction intensity{ SL-1 32.1 (26.5)A X 21.4 (17.7) 23.7 (22.3)A Y
UL-1 17.3 (18.7)B 18.9 (20.7)B
Contralateral shank average intensity{ SL-1 9.8 (6.5)B 10.6 (7.9) 8.8 (7.7)B
UL-1 17.1 (10.3)A 16.5 (17.1)A
Contralateral thigh co-contraction intensity{ SL-1 37.4 (26.9) 40.3 (26.8) 38.4 (24.8)UL-1 35.3 (23.7) 38.1 (26.4)
Contralateral thigh average intensity{ SL-1 9.4 (7.2)B 10.3 (8.3) 9.7 (7.9)B
UL-1 11.0 (9.1)A 11.2 (11.4)A
Notes: A group 4 B group during the specified task (p 5 0.05); X task 4 Y task in the specified group (p 5 0.05);{
% of the larger muscleactivity; {% MVC. Significant pairwise comparison results are presented with superscript letters and shaded. Values are given as mean (SD).
1066 R.D. Catena et al.
7/30/2019 Gait research
9/13
to either pick up or drop off the box (p 5 0.001).
There were no statistical differences between lower
extremity muscle involvements in each task for theSL-1 group (Table 3).
The outcome balance measures of individuals after
introduction of a slippery surface changed significantly
from a high friction surface for the successful group
(Table 4). The SL-1 group took 22% longer (more
than half a second longer) to complete the box transfer
in LCOF-1 compared with the HCOF condition
(p 0.010). During this time, there was an increasein COM travel distance (p 0.003). The COG of theSL-1 group stayed 1.6 cm further from the edge of
their lateral-fore foot side of the BOS during box
acquisition with the slippery surface (p 5 0.001). On
the other hand, the COG travelled closer to theanterior and posterior sides of the BOS during the
task with a slippery surface (p 5 0.001 and p 0.017,respectively).
Group-by-task comparison with HCOF and
LCOF-2 tasks
By analysing performance during the HCOF and
LCOF-2, we hoped to determine how the two groups
adapted to a slippery surface to accomplish the lateral
load transfer. This in turn provided some additional
information of what differentiated success in the first
slippery trial. Eventually, all individuals from the UL-1
group made sufficient modifications to accomplish the
lateral load transfer over the slippery surface. This
occurred after several practice trials during which the
pedestal distance was initially shortened, and then
increased as a new transfer technique was employed
successfully. By the end of the practice trials, there
were no statistical differences in transfer distance
between the two groups or between surface conditions.
For both groups, the maximum transfer distance for
LCOF-2 was determined with half as many practice
trials as for HCOF (p 5 0.001), but there was no
group difference in the number of practice trials used(p 0.815).
While the UL-1 group had a significantly wider
start stance than SL-1 in the high friction condition (as
documented in a previous section), the group
difference was not present in the adapted low friction
condition (Table 2). The UL-1 group had decreased
their start stance width by 7.8 cm during LCOF-2
compared to HCOF (p 0.002), while there was nosignificant change in the SL-1 group. As previously
mentioned, the SL-1 group had more external rotation
of the foot towards the box just before picking it up
during HCOF compared to the UL-1 group. This
remained true during LCOF-2 (p 0.027), but bothgroups increased external rotation of the pickup foot
even more so during this task (p 5 0.001). Likewise,
when individuals turned to drop the box off on the
opposing pedestal, both groups increased the external
rotation of the foot on that side by more than 158
(p 5 0.001). The two groups modified muscular
control going from HCOF to LCOF-2 similarly and
group difference remained similar between the two
tasks. The one difference between the two groups was
that the SL-1 group decreased contralateral shank
co-contraction during LCOF-2 (p 0.008), while theUL-1 group did not (Table 3).
Occasional imbalances still occurred after
adaptation to the slippery surface. There was no
difference in the number of unsuccessful trials between
the two groups during HCOF (Table 4), but the UL-1
group had significantly more unsuccessful trials
during LCOF-2 than the SL-1 group (p 0.026). Wefound an increase in the area of the BOS covered by
the COG for both groups during LCOF-2 compared to
HCOF (p 5 0.001). The COG of all individuals
travelled closer to the back (p 5 0.001) and front
Table 4. Balance control variables across the three conditions.
Variable Group HCOF LCOF-1 LCOF-2
Total area of COG (% of start BOS area) SL-1 25.0 (9.8)Y 27.7 (9.8) 28.8 (9.5)X
UL-1 24.6 (9.0)Y 29.8 (9.8)X
COM path distance during transfer (% of box displacement) SL-1 28.5 (4.4)Y 30.0 (4.3)X 30.0 (4.6)X
UL-1 29.7 (4.9) 29.3 (4.4)
Minimum distance between COG and front of BOS (cm) SL-1 7.18 (3.8)
A X
5.19 (3.2)
Y
4.52 (3.0)
Y
UL-1 8.38 (3.4)B X 6.22 (3.0)Y
Minimum distance between COG and back of BOS (cm) SL-1 7.88 (2.4)X 7.13 (2.2)Y 5.67 (2.4)Y
UL-1 7.83 (2.2)X 6.66 (2.0)Y
Practice trials prior to the given task SL-1 13.83 (5.8)X 7.09 (1.5)Y
UL-1 13.69 (5.9)X 7.38 (2.3)Y
Average number of foot lifts(trials excluded during analysis of other variables)
SL-1 0.74 (1.1) 0.39 (0.8)B
UL-1 0.69 (1.0) 1.15 (1.1)A
Notes: A group 4 B group during the specified task (p 5 0.05); X task 4 Y task in the specified group (p 5 0.05). Significant pairwisecomparison results are presented with superscript letters and shaded. Values are given as mean (SD).
Ergonomics 1067
7/30/2019 Gait research
10/13
(p 5 0.001) of the BOS during LCOF-2 compared to
HCOF.
Discussion
The purpose of this study was to evaluate balance
control during lateral load transfers to determine (1)characteristics of individuals able to maintain balance
on a novel slippery surface and (2) to see how
individuals adapted to a slippery surface, particularly
when at first they were unable to maintain balance. We
had originally hypothesised that the key to maintain-
ing balance during the lateral load transfer over a
slippery surface was increased joint stiffness as
scientific literature and our own past research had
suggested. We found that this was not the case for a
slippery surface. The most beneficial technique to
maintaining balance was to increase intended move-
ment towards and with the box, as evident in the
success individuals had after allowing greater footrotations with stance width adjustments. These find-
ings suggest that individuals already incorporating
increased body motion towards the direction of
interest will be less likely to succumb to a loss of
balance on the occasional low friction surface experi-
enced during a lateral material handling task.
Balance control
Balance control was measured several ways in this
study, the most direct being a loss of balance. A
complete loss of balance was rare. Foot lifts to quickly
readjust the BOS to encompass the body COG weremuch more prevalent. All individuals at some point in
testing or practice trials used such a strategy. However,
the UL-1 group had a significant increase in these
needed BOS adjustments during the required comple-
tion of six successful low COF trials. This indicated
that adaptation to the slippery surface did not
completely eliminate imbalance and that the UL-1
individuals may face an increased chance of a loss of
balance. In fact, we never observed UL-1 performance
completely equivalent to SL-1 performance. Future
studies might examine performance over more trials to
determine if there is ever complete convergence of the
two groups.
By examining balance through measures of the
COM, we were able to identify how balance was
possibly lost in many of these situations. While the
lateral load transfer task has large movements from
side-to-side, the BOS is appropriately widened for this
movement. However, by widening the BOS to such a
degree, anterior and posterior control is possibly
reduced through a significant alteration in normal
muscle functioning. We were unable to find any
scientific literature describing balance with a very
wide stance (450 cm), but when stance width is
increased to such a degree muscle pathways and
lengths across the hip, knee and ankle are drastically
altered. Some muscles such as the gluteals and the
tibialis anterior are shortened when stance is wide,
while hamstrings and peroneal muscles are lengthened.The lengthtension relationship then dictates that these
muscles will have reduced force production (Nordin
and Frankel 2001). We suspect muscle tensions played
a role in anterior/posterior imbalance seen.
When individuals stood with a 45-cm stance width,
there were no deleterious effects on balance (Kirby
et al. 1987). However, when individuals rotate their
pelvis towards the box to pick it up, they are now
actually standing with less mediolateral stance width
and more A/P foot displacement. When individuals
stood with 30 cm of A/P foot separation, there were
significant increases in mediolateral deviation of the
centre of pressure (Kirby et al. 1987), similar to A/Pmeasures in our study because our measures are based
on a global reference. Since anterior and posterior
COG movement increased in the slippery surface
conditions, it seems apparent that movements in these
directions are important to examine as a direct cause of
loss of balance during lateral load transfers. Something
about the SL-1 group allowed them to explore these
areas of the BOS, similar to the UL-1 group, without
the same detrimental balance effects that the UL-1
groups faced.
Stance kinematicsOne possible area to analyse group difference in
imbalance in the A/P direction is kinematics of the
lateral load transfer manoeuvre. There was clear
evidence of group differences in transfer manoeuvres.
The UL-1 group chose a wider BOS to transfer this
box over a high friction surface. A wider BOS may
indicate even further reductions in force production of
lower extremity muscles. When used during the
slippery surface condition, this wider BOS may not be
as easily controlled as a narrower BOS when friction is
not supplying a lateral stopping force against the sole
of the foot and hip adductor muscles are solely
responsible for maintaining the stance width. This is
similar to hip adductor use to decelerate the foot in
push off during skating (Chang et al. 2009). One of the
adaptive techniques that the UL-1 group chose to
employ was a narrowed stance width in LCOF-2. The
hypothesis that stance width determines success then
leads to questions about strength as a contributing
factor. A narrower stance width would require more
low-back extensor activity to reach a load because of a
need for increased torso flexion. Greater hip adductor
1068 R.D. Catena et al.
7/30/2019 Gait research
11/13
strength would allow individuals to maintain a slightly
wider BOS over a slippery surface to avoid torso
flexion.
Another important kinematic change was the
external foot rotation towards the load. In the high
friction conditions, the SL-1 group rotated the
ipsilateral foot more towards the load than the UL-1group. External rotation was even greater during the
first slippery trial. UL-1 individuals similarly increased
the amount of ipsilateral foot rotation to the load in
the adaptive low friction trials. One possible reason
this may be beneficial is that this manoeuvre positions
the lower extremities in such a way that flexion at the
ankle, knee and hip is towards the load rather than
only having a hip abduction moment available to
control movement in the direction of interest as would
be the case if the body remained directed forward with
the load still by the side. This hypothesis emphasises
the importance of muscles that were unfortunately not
measured in our study (e.g. gluteus medius and tensorfascia latae) so we can only speculate that the rotation
manoeuvre decreased the reliance on less effective hip
abductors. A second reason external rotation might be
significant is not that it just signifies what is specifically
needed for better performance in this particular task,
but it might also be an indicator that the groups
benefited from a strategy that utilised body movement
to their advantage. Foot rotations may be just one
indicator of a surfer movement over slippery surfaces;
a motion altered to glide with body momentum over
the slippery surface, rather than fighting against the
motion unreduced by a lack of surface friction
(Marigold and Patla 2002). This alteration reducesthe need for shear forces applied to the surface for
movement. To completely analyse this hypothesis,
future studies will look at whole body kinematics.
The unfavourable effect of rotating feet towards to
the load is that the body is put into a more tandem-like
stance configuration. This form of stance has been
demonstrated to be more unstable than side-by-side
stance (Winter 1995). We believe that this contributed
to the increased motion towards the forward and
backward directions of the BOS. Doing this would
require increased activation of lateral balance control
musculature that was unfortunately not measured in
our testing.
Muscular control
Initial and continued group differences in lower
extremity muscle activity would seem to indicate that
the UL-1 group stiffened lower extremity joints in an
apparent conservative strategy to avoid falling. This is
similar to previous literature that shows that many
different populations regard joint stiffening as a way to
avoid falls (Santello and McDonagh 1998, Lark et al.
2003, Nielsen et al. 2004). However, muscular control
that we measured had no apparent effect on success in
the lateral load transfer over the slippery surface. If
anything, it was beneficial to loosen the joints with less
lower-extremity muscle activity. Their continued slight
increase in joint stiffness may be an explanation as towhy the UL-1 group required more step manoeuvres to
avoid falling even after adaptation to the slippery
surface. This was contradictory to our hypothesis and
to our previous findings that indicated stiffening of the
ankle joint to counter increased load weight during the
lateral load transfer (Catena et al. 2010). This was also
contradictory to others interpretation of the benefit of
ankle stiffness. However, some researchers have found
that while joint stiffening is employed, it is not always
successful in maintaining postural control (Reeves
et al. 2006, Cenciarini et al. 2010, Reynolds 2010),
which indicates task dependency, or that success may
depend on the joint and task being examined (Chamand Redfern 2001). Future studies of the lateral load
transfers should include measurements of muscles
involved in coronal and transverse plane movements to
complete the analysis of muscular control.
Demographic characteristics
We found no indication that simple demographic
factors determined successful performance in lateral
load transfer over a low friction surface. Among the
examined factors, age (Winter 1995) and BMI (Hue
et al. 2007, Teasdale et al. 2007) have been touted as
correlating factors to balance control. Our sample ofindividuals contained a range of BMIs (17.734.2 kg/
m2) that we think was adequate to assess it as a non-
factor in balance during the lateral load transfer for
most of the working population performing such tasks.
We had a normal distribution of BMIs that made up
our sample. Unfortunately, we do not believe our
sample of age (2066 years) was completely adequate
to state for sure that it is a non-factor in balance
during the lateral load transfer. Rather than a normal
distribution, ours was a uniform distribution of age. It
also failed to include enough individuals of a balance-
effected age, older than 50 (Winter 1995). We do
believe that age will certainly be a factor past 60, but
our findings do show that the gradual age increase was
not a factor in the working age population.
Limitations
(1) A list of instructions were given to the
participants and specified in the Procedure
section. These instructions were meant to limit
Ergonomics 1069
7/30/2019 Gait research
12/13
the variability between individuals. While this
process allows us to specifically analyse the
variables of interest (low friction and balance),
we risked missing the natural or desired
manoeuvres individuals might use if not con-
strained. Specifically, not allowing stepping
would be highly unlikely to occur in a naturalsetting, especially for such a transfer distance.
Our requirement that the maximum box
transfer distance be attempted was also meant
to reduce the variability between individuals.
This requirement meant that individuals at-
tempted lateral load transfer much further than
they would normally. Observed compensations
for this unnatural distance were a wider stance
in the UL-1 group and more trunk flexion in
the SL-1 group. An interesting future study
might use information from this study to
compare with natural performance of the
same task without said restrictions.(2) There are a couple of limitations to using the
study cohort. We have no reason to believe
that our cohort is anthropometrically charac-
teristic of the manual material handling popu-
lation. Application of our results should
consider this. Also, we did not attempt to
control was athletic ability. We had one
individual in our study that was a collegiate
athlete, and several others that would be
considered athletic. These individuals were
able to complete the first slippery trial success-
fully, but since there were very few of them we
could not analyse them in a statistical analysis.Athletes have been previously shown to recover
from imbalance situations better than non-
athletes (Brauer et al. 2008). Increased strength
and coordination of an athlete would imagin-
ably have some beneficial effect on perfor-
mance of the lateral load transfer.
(3) Muscle activity was a key factor in our previous
study of lateral load transfers (Catena et al.
2010). Specifically, flexor/extensor co-contrac-
tions were found to be a factor in performance.
Our current study continued with these mea-
sures but was not able to draw any significant
conclusions from them. Since we did not
anticipate transverse and coronal plane kine-
matics as being important, we made no
measurements of major contributors to move-
ment in these planes. The result was a lack of
information that would possibly be beneficial
to our analysis. Likewise, no measure of
strength was recorded, which would be helpful
in making more concrete conclusions about the
consequences of stance width adjustments.
Conclusions
This study has shown that a slippery surface has
adverse effects on individuals using a wide stance and
constricted movement to perform a lateral load
transfer. Even though the task is mostly side-to-side
movement, the balance control results suggest that loss
of balance will likely occur in the A/P directions. Theoften employed strategy of maintaining balance
through increased joint stiffness was not used
successfully to maintain balance in the slippery
condition, was not increased any further in adaptation
to the slippery condition, and in some cases was even
decreased during the slippery condition. Only stance
kinematic performance was able to distinguish success
in the slippery conditions. Whole body kinematics will
be analysed in future studies to determine the extent
that body kinematics plays in successful completion of
the lateral load transfer over a slippery surface.
Having shown that simple modifications to stance
kinematics can determine balance control over aslippery surface during lateral material handling,
further research strategies could be taught to
individuals having to occasionally perform such an
occupational task. It might also be the case that a
decreased reliance on muscle control might be
beneficial for such a task. However, we caution any
immediate use of the techniques described in this
article without further analysis on whole body joint
kinematics to detect any additional modifications to
the lateral manoeuvre that might be needed
accompaniments.
AcknowledgementsThis research was supported by the Liberty Mutual-Harvard School of Public Health Postdoctoral Program,as a postdoctoral fellowship was awarded to Dr. Catena.The authors also thank Simon Matz for his help withstatistical analyses.
References
Bureau of Labor Statistics (BLS), 2009. Survey ofoccupational injuries and illnesses in cooperation withparticipating state agencies [online]. US Department ofLabor. Available from: http://www.bls.gov/ [Accessed 26September 2011].
Bakken, G.M., et al., 2006. Slips, trips, missteps and theirconsequences. Tucson, AZ: Lawyers and JudgesPublishing.
Birrell, S.A. and Haslam, R.A., 2009. The effect of militaryload carriage on 3-D lower limb kinematics andspatiotemporal parameters. Ergonomics, 52, 12981304.
Brauer, S.G., Neros, C., and Woollacott, M., 2008. Balancecontrol in the elderly: do Masters athletes show moreefficient balance responses than healthy older adults?Aging Clinical and Experimental Research, 20, 406411.
Catena, R.D., et al., 2010. The effect of load weight onbalance control during lateral box transfers. Ergonomics,53, 13591367.
1070 R.D. Catena et al.
7/30/2019 Gait research
13/13
Catena, R.D., van Donkelaar, P., and Chou, L.S., 2009.Different gait tasks distinguish immediate vs. long-termeffects of concussion on balance control. Journal ofNeuroengineering and Rehabilitation, 6, 25.
Cenciarini, M., et al., 2010. Stiffness and damping in posturalcontrol increase with age. IEEE Transactions on Biome-dical Engineering, 57, 267275.
Cham, R. and Redfern, M.S., 2001. Lower extremity
corrective reactions to slip events. Journal of Biomecha-nics, 34, 14391445.
Cham, R. and Redfern, M.S., 2002. Changes in gait whenanticipating slippery floors. Gait Posture, 15, 159171.
Chang, R., Turcotte, R., and Pearsall, D., 2009. Hipadductor muscle function in forward skating. SportsBiomechanics, 8, 212222.
Chang, W.-R., Cotnam, J.P., and Matz, S., 2003. Fieldevaluation of two commonly used slipmeters. AppliedErgonomics, 34, 5160.
Commissaris, D.A. and Toussaint, H.M., 1997. Loadknowledge affects low-back loading and control ofbalance in lifting tasks. Ergonomics, 40, 559575.
Cooper, S.A., et al., 2005. Reducing stability of supportstructure for a target does not alter reach kinematics
among younger adults. Perceptual & Motor Skills, 100,831838.
de Leva, P., 1996. Adjustments to Zatsiorsky-Seluyanovssegment inertia parameters. Journal of Biomechanics, 29,12231230.
Durkin, J.L. and Dowling, J.J., 2003. Analysis of bodysegment parameter differences between four humanpopulations and the estimation errors of four popularmathematical models. Journal of Biomechanical Engineer-ing, 125, 515522.
Gillette, J.C., et al., 2010. The effects of age and type ofcarrying task on lower extremity kinematics. Ergonomics,53, 355364.
Hahn, M.E. and Chou, L.-S., 2004. Age-related reduction insagittal plane center of mass motion during obstacle
crossing. Journal of Biomechanics, 37, 837844.Hermens, H.J., et al., 1999. European recommendations forsurface electromyography, deliverable of the SENIAMproject. The Netherlands: Roessingh Research andDevelopment.
Holbein, M.A. and Redfern, M.S., 1997. Functional stabilitylimits while holding loads in various positions. Interna-tional Journal of Industrial Ergonomics, 19, 387395.
Holbein-Jenny, M.A., et al., 2007. Kinematics of heelstrikeduring walking and carrying: implications for slipresistance testing. Ergonomics, 50, 352363.
Hue, O., et al., 2007. Body weight is a strong predictor ofpostural stability. Gait Posture, 26, 3238.
Kirby, R.L., Price, N.A., and MacLeod, D.A., 1987. Theinfluence of foot position on standing balance. Journal ofBiomechanics, 20, 423427.
Kollmitzer, J., et al., 2002. Postural control during lifting.Journal of Biomechanics, 35, 585594.
Kozak, K., Ashton-Miller, J.A., and Alexander, N.B., 2003.The effect of age and movement speed on maximumforward reach from an elevated surface: a study inhealthy women. Clinical Biomechanics, 18, 190196.
Lark, S.D., et al., 2003. Joint torques and dynamic jointstiffness in elderly and young men during stepping down.Clinical Biomechanics, 18, 848855.
Liao, C.F. and Lin, S.I., 2008. Effects of different movementstrategies on forward reach distance. Gait Posture, 28,1623.
Lo, J. and Ashton-Miller, J.A., 2008. Effect of pre-impactmovement strategies on the impact forces resulting from
a lateral fall. Journal of Biomechanics, 41, 19691977.Marigold, D.S. and Patla, A.E., 2002. Strategies for dynamic
stability during locomotion on a slippery surface: effectsof prior experience and knowledge. Journal ofNeurophysiology , 88, 339353.
Myung, R. and Smith, J.L., 1997. The effect of load carryingand floor contaminants on slip and fall parameters.Ergonomics, 40, 235246.
Nielsen, J.F., Andersen, H., and Sinkjaer, T., 2004.Decreased stiffness at the ankle joint in patients withlong-term Type 1 diabetes. Diabetic Medicine, 21, 539544.
Nordin, M. and Frankel, V.H., 2001. Basic biomechanics ofthe musculoskeletal System. Philadelphia: LippincottWilliams & Wilkins.
Pavol, M.J., Owings, T.M., and Grabiner, M.D., 2002. Bodysegment inertial parameter estimation for the generalpopulation of older adults. Journal of Biomechanics, 35,707712.
Plagenhoef, S., Evans, F.G., and Abdelnour, T., 1983.Anatomical data for analyzing human motion. ResearchQuarterly for Exercise and Sport, 54, 169178.
Reeves, N.P., et al., 2006. The effects of trunk stiffness onpostural control during unstable seated balance.Experimental Brain Research, 174, 694700.
Reynolds, R.F., 2010. The ability to voluntarily control swayreflects the difficulty of the standing task. Gait Posture,31, 7881.
Row, B.S. and Cavanagh, P.R., 2007. Reaching upward ismore challenging to dynamic balance than reaching
forward. Clinical Biomechanics, 22, 155164.Santello, M. and McDonagh, M.J., 1998. The control oftiming and amplitude of EMG activity in landingmovements in humans. Experimental Physiology, 83,857874.
Streepey, J.W. and Angulo-Kinzler, R.M., 2002. The roleof task difficulty in the control of dynamic balance inchildren and adults. Human Movement Science, 21, 423438.
Teasdale, N., et al., 2007. Reducing weight increases posturalstability in obese and morbid obese men. InternationalJournal of Obesity (London), 31, 153160.
Toussaint, H.M., et al., 1998. Scaling anticipatory posturaladjustments dependent on confidence of load estimationin a bi-manual whole-body lifting task. ExperimentalBrain Research, 120, 8594.
Winter, D.A., 1990. Biomechanics and motor control of humanmovement. New York, NY: Wiley-Interscience.
Winter, D.A., 1995. A.B.C. (Anatomy Biomechanics andControl) of balance during standing and walking.Waterloo, ON: University of Waterloo.
Ergonomics 1071