Journal of Biomechanics 33 (2000) 601}608

E!ect of resistance load on biomechanical characteristics of racing

wheelchair propulsion over a roller system
John W. Chow!,*, Tim A. Millikan", Les G. Carlton!, Woen-sik Chae!, Marty I. Morse"
!Department of Kinesiology, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA
"Division of Rehabilitation Education Services, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA
Accepted 3 September 1999

Abstract

The purpose of this study was to examine how resistance load in#uenced the kinematic characteristics and the activity of selected
muscles (#exor and extensor carpi radialis, biceps brachii, triceps brachii, antero-middle and postero-middle deltoids, pectoralis
major, and upper trapezius) during maximum e!ort racing wheelchair stroking using 3D videographic and surface EMG techniques.
Fifteen male experienced wheelchair racers served as subjects and three consecutive stroke cycles were analyzed for two load
conditions. In contrast to previous studies where variations in speed were a result of variations in pushing e!ort or disability
classi"cation, a reduction in stroking speed caused by increasing load did not result in a decrease in stroking frequency. Increases in
load signi"cantly in#uenced the push and recovery times but not the stroke time or frequency. The vertical ranges of motion and
vertical velocities at initial hand contact of the upper extremity joints decreased signi"cantly from light to heavy resistance conditions.
These results suggest that the vertical motion is in#uenced greatly by the load. Various degrees of muscle co-contractions were
observed in most phases of the stroke cycle. The activation pattern of the deltoid muscle was di!erent from what has been previously
reported, probably because of the exaggerated forward lean trunk position adopted by our subjects. Although the overall EMG
activity remained the same or decreased when the resistance was increased, stroking under a heavy resistance load is likely to be more
demanding physiologically because of the greater push time}recovery time (work}rest) ratio with increasing resistance. ( 2000
Elsevier Science Ltd. All rights reserved.

Keywords: Kinematics; Electromyography; Muscle activation; Wheelchair athletics

1. Introduction Under the same stroking e!ort, the stroking speed

Wheelchair racing is one of the most popular indi-
vidual sports among atheltes with locomotor disabilities.
Wheelchair races are held either on roads or on standard
400 m tracks. A major di!erence between road and track
races is the hills involved in road races. When pushing
a wheelchair up a hill, the weight of the wheelchair plus
racer has a component that is parallel to the inclined
surface and serves as a resistance to the forward motion
of the wheelchair. It is worth noting that no attempt has
been made to explore the di!erences in stroke mechanics
and muscle activation during level and uphill stroking. It
is apparent that this information would be useful for
instructional purposes and the design of training programs.
* Corresponding author. Tel.: #1217-244-3987; fax:#1217-244-7322.
E-mail address: j-chow1@uiuc.edu (J.W. Chow)
decreases with increasing slope when pushing uphill.
Previous studies on the mechanics of racing wheelchair
propulsion have indicated that greater speed is asso-
ciated with greater stroking frequency and reductions in
stroke, push, and recovery times (Goosey et al.,
1997,1998; O'Connor et al., 1998; Ridgway et al., 1988;
Wang et al., 1995; Woude et al., 1988). Variations in the
procedures used in these studies, however, make it di$-
cult to determine whether these changes are a result of
variations in e!ort, load, or speed. Studies of muscle
functioning using electromyography (EMG) have been
mostly qualitative in nature (Ma( sse et al., 1992; Wang,
1991). In general, the pectoralis major, anterior deltoid,
biceps and triceps brachii have been shown to be more
active in the push phase, while the posterior deltoid is
more active during the recovery phase.
The purpose of this study was to examine the in#uence
of resistance load on stroking characteristics and muscle

0021-9290/00/$ - see front matter ( 2000 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 2 1 - 9 2 9 0 ( 9 9 ) 0 0 2 1 1 - 0
602 J.W. Chow et al. / Journal of Biomechanics 33 (2000) 601}608
activation during maximum e!ort racing wheelchair
stroking. Two loads were used that simulated level and
uphill stroking. Because all trials involved maximum
e!ort, variations in speed in this study were due to
variations in load and not to di!erences in e!ort or
disability classi"cation. Based on the relationships be-
tween speed and stroking characteristics reported in the
literature, it was hypothesized that an increase in resist-
ance (a decrease in speed) would cause a decrease in
stroking frequency and increases in stroke, push, and
recovery times. In addition to the relations among di!er-
ent stroking characteristics, this study also provided
a systematic examination of muscle functioning under
maximum e!ort wheelchair propulsion.
2. Methods
Fifteen male (26$8 yr, injury level T5-L3, track classi-
"cation T3-T4) highly trained experienced racers
served as subjects and they signed informed consent
documents that had been approved by the Institu-
tional Review Board. All trials were performed on a
computerized drum roller system with a maximum brak-
ing resistance of 15 N m (Eagle Sports Chair). Two S-
VHS camcorders (Panasonic AG-455, 60 Hz) were
located in front and behind, and to the left of the subject.
A calibration frame (17 control points, 1.3]1.1]0.9 m3),
a plumbline, and two markers were used for spatial
reference and de"ning a global reference frame, respec-
tively.
Eight pairs of surface EMG electrodes with on-site
preampli"cation circuitry (Liberty Technology MYO115
electrodes, gain "1000, input impedance '1014 ),
CMRR '90 dB) were attached to the right side of the
body: #exor carpi radialis (FCR), approximately 1 of the
3
forearm length from the elbow on the medial side; exten-
sor carpi radialis (ECR), approximately 1 of the forearm
4
length from the elbow on the lateral side; biceps brachii
(BBR), approximately midway between shoulder and el-
bow on the anterior side; triceps brachii (TBR), approx-
imately midway between shoulder and elbow on the
posterior side; antero-middle deltoid (AMD), midway
between the anterior and middle deltoids, postero-middle
deltoid (PMD), midway between the middle and poste-
rior deltoids, pectoralis major (PCM), midway between
the armpit and suprasternal notch, and upper trapezius
(UTR), approximately midway between the acromion of
scapula and C7/T1 vertebrae. To obtain maximum EMG
levels of the selected muscles, maximal e!ort isometric
exercises were performed before the experimental trials.
The EMG signals were further magni"ed using a gen-
eral-purpose ampli"er (Biocommunication Electronics,
Model 215, input impedance "109 ), CMRR
'100 dB) before A/D conversion (12-bit) at a sampling
rate of 1000 Hz.
Each subject, using his own racing wheelchair and
gloves, was asked to perform a maximum e!ort trial at
each of the two resistance loads * 30% (light) and 50%
(heavy) of the maximum resistance o!ered by the roller
system. In each trial, the subject stroked for 30 s and was
videotaped. The EMG signals were collected for 5 s dur-
ing the second half of the trial duration. To synchronize
the video and EMG data, a synchronization unit was
manually activated during the time the EMG signals
were sampled. The unit activated a large light-emitting
diode (LED), which was visible in both camera views,
and forwarded a 3 V signal to the A/D converter.
The direct linear transformation (DLT) procedure (Ab-
del-Aziz et al., 1971) was used to obtain three-dimen-
sional coordinates. The average root-mean-square
calibration error was 2.55 mm. For each subject, three
stroke cycles of EMG and the corresponding video re-
cordings were analyzed. A stroke cycle starts at the
instant of initial hand contact with the pushrim and ends
at the instant of the next initial hand contact. A Peak
manual digitizing system (Englewood, CO) was used to
extract two-dimensional coordinates of the supra-sternal
notch, midpoint between two hips, left shoulder, elbow,
wrist, and third knuckle from the video recordings.
A coordinate transformation was performed so that the
principal axes of the global reference frame were aligned
with the antero-posterior, vertical, and medio-lateral di-
rections (positive for forward, upward, and medial direc-
tions). For each stroke cycle, selected kinematic
parameters of the body landmarks (relative location and
velocity) and body segments (inclination and angular
speed) at the instants of initial hand contact (HC), hand
release (the instant the hand breaks contact with the
pushrim, HR), and maximum elbow height (MEH) were
determined. Time durations, both in seconds and as
Fig. 1. Angular parameters used to indicate the locations of hand
relative to the wheel center at the instants of initial hand contact
(contact angle) and release (release angle). Both angles are measured
clockwise from a line joining the wheel center and top dead center.
 J.W. Chow et al. / Journal of Biomechanics 33 (2000) 601}608
a fraction of the stroke time, were determined for the
push (from HC to HR), ascending recovery (HR to
MEH), descending recovery (MEH to HC), and recovery
(HR to HC) phases. In addition, contact, release, and
push angles (Fig. 1) were determined for each stroke.
The raw EMG signals were "ltered using a recursive
digital "lter (Matlab Elliptic "lter, 10}500 Hz band pass)
and full-wave recti"ed. The data were smoothed using
a moving average of 30 ms before normalizing to the
respective maximum isometric EMG levels. For each
stroke cycle, average EMG values were determined for
di!erent phases and the entire cycle. In addition, the
instant of peak EMG activity was determined for each
muscle. For each parameter in each trial, the average
value over the three stroke cycles was used for sub-
sequent analysis. For each resistance load, means and
standard deviations were computed for each parameter.
An ANOVA with repeated measures was used to test for
signi"cant di!erences between the two resistance loads
(p)0.05).
3. Results
Despite a decrease in speed at the heavy resistance
condition, the subjects maintained the same stroke time
and frequency by increasing the push time and decreas-
ing the recovery time (Table 1). Only the hypothesis that
the push time increases with increased load was sup-
ported. The decrease in descending recovery time con-
tributed more to the decrease in recovery time with
increasing resistance than the decrease in ascending re-
covery time did.
The relative locations of the joints indicate that the
upper arm and forearm positions at HC were about the
same at both resistance loads (Table 2). Signi"cant di!er-
ences in arm positions at MEH were found between the
Table 1
Temperal characteristics of a stroke cycle!
Resistance load
Light
Speed (m s~1)" 9.39(1.03)
Sroke time (s) 0.539 (0.051)
Stroke frequency (Hz) 1.873 (0.180)
Push time (s) 0.094 (0.018)
Recovery push time (s) 17.77(3.7)
Recovery time (s) 0.444 (0.056)
Relative recovery time (%) 82.3 (3.7)
Ascending recovery time (s) 0.336 (0.043)
Relative ascending recovery time (%) 62.2 (3.7)
Desceding recovery time (%) 0.109 (0.023)
Relative desceding recovery time (%) 20.0 (3.5)
!Note: Standard deviations in parentheses.
"Speedometer reading.
603
two loads. A qualitative analysis of joint trajectories
(Fig. 2) indicates that, for all subjects, both the shoulder
and elbow reached their maximum heights at the same
instant, the wrist attained its maximum height shortly
before the MEH, and the shoulder reached its lowest
position at HR regardless of the load.
With an increase in resistance, there was a decrease in
joint ROMs and these decreases were greater in the
vertical than in the antero-posterior and medio-lateral
directions (Table 3). The release angle decreased signi"-
cantly with increasing resistance (Table 4). Since the push
angle remained relatively constant across resistance con-
ditions, the reduction in ROM with increasing load must
be a result of greater changes in arm motion during the
recovery phase than in the push phase.
In general, joint velocities at HC and HR decreased
with increasing resistance (Table 5). Regardless of the
load, the wrist was moving forward, downward, and
inward at MEH. The di!erences in segment inclinations
at the two resistance conditions are consistent with the
other measures as they demonstrate a decrease in arm
excursion with increasing resistance (Table 6). Regardless
of the resistance, the change in angular speed from HC to
HR was greater in the upper arm than in the forearm.
Signi"cantly higher EMG levels were observed in the
light resistance trials in the FCR for the stroke, ascending
recovery, and recovery phases (p)0.05), and in the ECR
for the stroke, descending recovery, and recovery
(p)0.03) phases (Fig. 3). Peak activity was observed in
the second half of the push and early part of the ascend-
ing recovery phases for the FCR and at the end of the
ascending recovery and beginning of the descending re-
covery phases for the ECR. Co-contractions of the FCR
and ECR were observed in all phases.
Signi"cantly higher EMG levels were observed in the
light resistance trials in the TBR for the ascending recov-
ery and recovery phases (p)0.03) (Fig. 3). Peak activity
p-value
Heavy
7.56 (0.89) (0.001
0.535 (0.056) 0.645
1.889 (0.197) 0.565
0.118 (0.023) (0.001
22.2 (4.2) (0.001
0.417 (0.055) 0.005
77.8 (4.2) (0.001
0.325 (0.049) 0.126
60.6 (4.3) 0.048
0.092 (0.020) 0.003
17.2 (3.6) 0.002
604 J.W. Chow et al. / Journal of Biomechanics 33 (2000) 601}608

Table 2
Relative joint locations (m) at di!erent critical instants!

Resistance load

Light Heavy

Antero/posterior Vertical Medio/lateral Antero/posterior Vertical Medio/lateral

Hand contact
Elbow/wrist 0.030 0.290 !0.066 0.031 0.289 !0.066
(0.045) (0.027) (0.030) (0.045) (0.027) (0.030)
Elbow/shoulder !0.223 0.035 !0.203 !0.229 0.026 !0.202
(0.034) (0.045) (!0.234) (0.038) (0.040) (!0.238)
Hand release
Elbow/wrist 0.084 0.287" 0.033 0.076 0.296" 0.022
(0.034) (0.020) (0.031) (0.036) (0.023) (0.027)
Elbow/shoulder !0.223 !0.193 !0.143 !0.224 !0.192 !0.146
(0.036) (0.034) (0.023) (0.033) (0.037) (0.021)
Maximum elbow height
Elbow/wrist 0.239" 0.135" 0.052" 0.205" 0.207" !0.005"
(0.035) (0.073) (0.064) (0.042) (0.052) (0.050)
Elbow/shoulder !0.165" 0.194" !0.200" !0.203" 0.165" !0.221"
(0.055) (0.030) (0.038) (0.047) (0.043) (0.038)

!Note: Standard deviations in parentheses. Positive value indicates that the "rst joint is located in front of, above, or medial to the second joint.
!Signi"cantly di!erent between loads (p)0.05).

Signi"cantly higher EMG levels were observed in the

Fig. 2. Sagittal views of the paths of the shoulder, elbow and wrist
joints over three stroke cycles at (a) light and (b) heavy loads. The stick
"gures show the upper arm, forearm, and hand positions at the instants
of hand contact, hand release and maximum elbow height.
was observed during the end of the ascending recovery
and "rst half of the descending recovery phase for the
BBR and during the late descending recovery and early
part of the push phases for the TBR. Noticeable co-
contractions of the BBR and TBR were observed in all
phases except the push phase.
heavy resistance trials in the AMD for the descending
recovery phase (p)0.01) (Fig. 3). In contrast, signi"-
cantly higher EMG levels were observed in the light
resistance trials in the PMD for the stroke, ascending
recovery, and recovery phases (p)0.01). Peak activity
was observed during the second half of the ascending
recovery phase and throughout the descending recovery
phase for the AMD and during the "rst half of the
ascending recovery phase for the PMD. Various degrees
of co-contractions of the AMD and PMD were observed
in all phases.
Signi"cantly higher EMG levels were observed in the
light resistance trials in the UTR for the stroke, ascend-
ing recovery, and recovery phases (p)0.02) (Fig. 3).
Peak activity was observed during the descending recov-
ery phase for the PCM and in the "rst half of the ascend-
ing recovery phase for the UTR.
4. Discussion
The purpose of this study was to compare the stroking
characteristics and activation levels of selected muscles
during maximum e!ort racing wheelchair stroking over
a roller system at two resistance loads. In contrast to our
hypothesis, which was based on previous studies where
variation in speed were a result of variations in pushing
e!ort or disability classi"cation (Gelhsen et al., 1990;
Ridgway et al., 1988; Wang et al., 1995; Woude et al.,
1988), the subjects maintained nearly identical stroke
 J.W. Chow et al. / Journal of Biomechanics 33 (2000) 601}608 605

Table 3
Joint range of motion (m)!

Resistance load

Light Heavy

Antero/posterior Vertical Medio/lateral Antero/posterior Vertical Medio/lateral

Shoulder 0.066 0.184" 0.054" 0.065 0.153" 0.045"

(0.027) (0.042) (0.016) (0.026) (0.043) (0.016)
Elbow 0.175" 0.574" 0.106 0.145" 0.511" 0.099
(0.045) (0.061) (0.039) (0.040) (0.065) (0.034)
Wrist 0.372 0.750" 0.140" 0.367 0.612" 0.110"
(0.079) (0.110) (0.072) (0.066) (0.086) (0.058)
Knuckle III 0.510 0.799" 0.180" 0.513 0.638" 0.148"
(0.082) (0.124) (0.083) (0.064) (0.102) (0.067)

!Note: Standard deviations in parentheses.

"Signi"cantly di!erent between laods (p)0.05).

Table 4 same (Table 4). It is interesting to note that, although

Contact, release, and push angles!
Resistance load
Light Heavy
Contact angle (3) 33.3 (12.1) 29.8 (14.6)
Release angle (3) 183.1 (24.6) 177.1 (24.2)
Push angle (3) 149.7 (21.5) 147.3 (23.8)
Angular speed" (3s~1) 1,601.6 (162.2) 1,266.6 (178.5)
!Note: Standard deviations in parentheses.
"Angular speed of the wheel during the push phase"push angle/push
time.
frequencies at di!erent loads by manipulating push and
recovery times. Thus, the direct relation between stroking
speed and stroke frequency does not hold for maximum
e!ort pushing with variations in load. Decreases in joint
vertical ROMs and minimal changes in the ascending
recovery time with increasing load (Tables 1 and 3) imply
that the average speeds during the ascending recovery
phase for di!erent joints were higher at the light resist-
ance condition. One explanation is that at the light
resistance condition, the subject's trunk has greater
downward momentum than at the heavy resistance con-
dition toward the end of the push phase. The greater
downward momentum may enhance the speed of the
trunk during the ascending recovery due to the e!ects of
(1) stretch shortening cycle in the back muscles and (2)
greater strain/elastic energy stored in the soft/connective
tissues around the hips when the trunk was lowered to an
extreme #exed position at the end of the push phase. As
a result, the trunk `rebounded o! the bottom of the
strokea quicker when the resistance was light.
Despite the insigni"cant di!erence in contact angle
and signi"cant di!erence in release angle between the
two resistance loads, the push angles were about the
a large range of push angle values has been reported in
wheelchair propulsion articles, the push angle seems to
be relatively constant for a given set of resistance mecha-
p-value nisms and chair types (Woude et al., 1988 * racing
wheelchair on a treadmill; Vanlandewijck et al., 1994
0.208 * basketball wheelchair on a treadmill; Veeger et al.,
0.025 1989 * everyday wheelchair on a treadmill; Veeger et al.,
0.467
(0.001
1992 * wheelchair ergometer; Wang et al., 1995 * rac-
ing wheelchair on a roller system).
The trunk inclinations found in this study (Table 6) are
much smaller than the corresponding values reported in
the literature (Gelhsen et al., 1990; Goosey et al., 1997;
Ridgway et al., 1988; Wang et al., 1995; Woude et al.,
1988). A possible explanation is that the recent change in
racing wheelchair design allows an athlete to tuck the
shanks underneath the thighs. In terms of supporting the
body weight, the new design shifts the main area of
support from the buttocks to the shin areas. As a result,
the athletes can adopt a #at (near-horizontal) trunk posi-
tion throughout a stroke cycle. Sanderson and Sommer
(1985) suggested that an increase in forward lean would
promote the ability to transfer power to the pushrim
from the trunk.
The strategy of using a tighter stroke pattern at the
high resistance condition appears to be a compromise
between the loss in body movement speed at HC and the
loss in angular momenta of the wheels during the recov-
ery phase. When the resistance is high, a compact stroke
allows shorter path of the hand and therefore less time o!
the pushrim during the recovery phase. This serves to
prevent signi"cant loss in angular momenta of the wheels
during the recovery due to the high resistance. This
"nding should be valuable for both teaching and training
purposes.
Activity of the forearm muscles during racing wheel-
chair propulsion has not been previously reported. Our
606 J.W. Chow et al. / Journal of Biomechanics 33 (2000) 601}608

Table 5
Joint velocities (m s~1) at di!erent critical instants!

Resistance load

Light Heavy

Antero/posterior Vertical Medio/lateral Antero/posterior Vertical Medio/lateral

Hand contact
Shoulder !0.07" !1.09" !0.39 0.12" !0.75" !0.33
(0.31) (0.37) (0.15) (0.31) (0.30) (0.19)
Elbow !0.74" !4.29" 0.07" !0.33" !3.47" !0.14"
(0.72) (0.74) (0.41) (0.58) (0.77) (0.36)
Wrist 1.33" !4.75" 0.21" 1.71" !3.77" 0.03
(0.75) (0.82) (0.33) (0.62) (0.74) (0.21)
Hand release
Shoulder !0.15 !0.34" 0.01 !0.11 !0.16" 0.05
(0.17) (0.36) (0.10) (0.20) (0.42) (0.14)
Elbow !0.07 !0.87" 0.52 !0.01 !0.59" 0.60
(0.38) (0.96) (0.39) (0.31) (0.78) (0.39)
Wrist !2.28" !0.21 !0.58" !1.95" !0.18 !0.44"
(0.64) (1.09) (0.37) (0.54) (0.90) (0.33)
Maximum elbow height
Shoulder 0.34 !0.09 0.04" 0.33 !0.09 0.12"
(0.17) (0.18) (0.15) (0.19) (0.19) (0.12)
Elbow 0.69 !0.03 !0.12 0.70 !0.04 !0.80
(0.36) (0.09) (0.28) (0.37) (0.10) (0.20)
Wrist 1.53 !1.39 0.41 1.75 !1.04 0.52
(0.63) (0.60) (0.53) (0.48) (0.53) (0.46)

!Note: Standard deviations in parentheses. A positive value indicates anterior, upward, or medial direction.
"Signi"cantly di!erent between loads (p)0.05).

Table 6 results showed that both FCR and ECR activity de-
Segment inclinations and angular speeds at di!erent critical instants! creased with increasing load (Fig. 3). This may suggest
Inclination (3)" Angular speed (3s~1) that the role of the wrist action changed from propulsion
to stabilization as the resistance increase. The moderate
Resistance load
ECR and low FCR activity during the recovery phase
Light Heavy Light Heavy suggest that the wrist extension is needed to position the
hand for its initial contact with the pushrim.
Hand contact
Trunk 14.1# 16.4# 109.8# 88.8# The activation patterns of the BBR, AMD and PMD
(8.4) (7.5) (40.3) (32.1) observed in this study do not match with the correspond-
Upper arm 6.2 4.8 511.7# 428.9# ing patterns reported in the literature (Ma( sse et al., 1992;
(8.1) (7.2) (146.2) (157.0)
Forearm !73.4 !73.3 515.0# 442.2# Wang, 1991). The di!erences may be related to the exag-
(5.8) (6.6) (112.3) (54.1) gerated forward lean trunk position adopted by our
Hand release subjects. There is a consensus that the functions of
Trunk 7.0 8.2 60.1 57.4 muscles in the shoulder region change when there is
(9.2) (8.5) (29.8) (27.3) a change in upper arm position relative to the trunk
Upper arm !36.0 !35.5 295.4# 219.9#
(6.7) (5.9) (120.0) (89.4) (Smith et al., 1996).
Forearm !72.1# !74.5# 497.1# 434.7# The EMG activity of the PCM demonstrates an ad-
(6.7) (6.5) (78.6) (75.4) vantage of dividing the recovery phase into ascending
Maximum elbow height and descending portions. When the recovery phase is
Trunk 24.4 23.1 30.0 32.7 considered as a single phase, the average EMG level of
(7.9) (7.6) (15.2) (17.2)
Upper arm 37.0# 29.0# 143.2 139.8 the PCM in the push phase is greater than the same
(8.9) (9.3) (44.4) (49.9) parameter in the recovery phase (Fig. 3). However, when
Forearm !27.5# !44.2# 281.0 270.4 the recovery phase is divided into two phases, the
(14.8) (11.5) (94.7) (91.3)
greatest average EMG level was found in the descending
!Note: Standard deviations in parentheses. recovery phase.
"Inclination is the smallest angle between a segment and the horizon- Using "xed loads instead of varying the loads to match
tal plane, positive if the distal endpoint is higher than the proximal
endpoint. subjects capabilities might be considered a limitation of
#Signi"cantly di!erent between loads (p)0.05). this study. However, because the subjects are all
 J.W. Chow et al. / Journal of Biomechanics 33 (2000) 601}608 607

Fig. 3. Average normalized EMG levels during di!erent phases of a stroke cycle at light and heavy resistance loads for di!erent muscles. Signi"cant
di!erences were found between the two loads (*). The error bars indicate standard deviations.

high-level wheelchair racers and were considered to be
a homogenous group in terms of pushing capability and
because a within subject design was used, the use of "xed
loads should not a!ect the major "ndings of this study.
Another limitation is that, because only two loads were
used in this study, the e!ect of load might not be fully
explored.
In summary, the results of this study did not support
the hypothesis we proposed. The kinematics changes that
occur as speed change depends on whether the speed
variations are associated with changes in load or are due
to changes in e!ort. The overall EMG activity remained
the same or decreased when the resistance was increased
from light to heavy (judging from the average EMG over
a stroke cycle, Fig. 3). These results are not surprising
since the subjects performed with maximum e!ort and
the muscle shortening speeds were not the same under
the two load conditions. It is our experience that the peak
EMG recorded during maximum e!ort isometric con-
tractions are often lower in amplitude than the EMG
recorded during maximum e!ort rapid contractions for
the same muscle action. Although the resistance had
minimum e!ects on overall EMG activity, stroking un-
der a heavy resistance is likely to be more demanding
608 J.W. Chow et al. / Journal of Biomechanics 33 (2000) 601}608
physiologically because of the increase in push time and
decrease in recovery time (decrease in work}rest ratio)
with increasing resistance.
Acknowledgements
The authors would like to thank Sarah Beckman,
Brian Case, Dan Johnson, and Young-tae Lim for their
assistance in data collection. This study was funded in
part by the University of Illinois Research Board.
References
Abdel-Aziz, Y.I., Karara, H.M., 1971. Direct linear transformation from
comparator coordinates into object space coordinates in close-
range photogrammetry. Proceedings of the ASP Symposium on
Close Range Photogrammetry. American Society of Photogram-
metry, Falls Church, VA, pp 1}18.
Gelhsen, G.M., Davis, R.W., Bahamonde, R., 1990. Intermittent velo-
city and wheelchair performance characteristics. Adapted Physical
Activity Quarterly 7, 219}230.
Goosey, V.L., Campbell, I.G., 1998. Pushing economy and propulsion
technique of wheelchair racers at three speeds. Adapted Physical
Activity Quarterly 15, 36}50.
Goosey, V.L., Fowler, N.E., Campbell, I.G., 1997. A kinematic analysis
of wheelchair propulsion techniques in senior male, senior female,
and junior male athletes. Adapted Physical Activity Quarterly 14,
156}165.
Ma( sse, L.C., Lamontagne, M., O'Riain, M.D., 1992. Biomechanical
analysis of wheelchair propulsion for various seating positions.
Journal of Rehabilitation Research and Development 29 (3), 12}28.
O'Connor, T.J., Robertson, R.N., Cooper, R.A., 1998. Three-dimen-
sional kinematic analysis and physiological assessment of racing
wheelchair propulsion. Adapted Physical Activity Quarterly 15, 1}14.
Ridgway, M., Pope, C., Wilkerson, J., 1988. A kinematic analysis of
800-meter wheelchair racing techniques. Adapted Physical Activity
Quarterly 5, 96}107.
Sanderson, D.J., Sommer III, H.J., 1985. Kinematic features of wheel-
chair propulsion. Journal of Biomechanics 18, 423}429.
Smith, L.K., Weiss, E.L., Lehmkuhl, L.D., 1996. In: Brunnstrom's
Clinical Kinesiology. F.A. Davis Company, Philadelphia, pp. 245.
Vanlandewijck, Y.C., Spaepen, A.J., Lysens, R.J., 1994. Wheelchair
propulsion e$ciency: movement pattern adaptations to speed
changes. Medicine and Science of Sports and Exercise 26,
1373}1381.
Veeger, H.E.J., Woude, L.H.V. van der, Rozendal, R.H., 1989. Wheel-
chair propulsion technique at di!erent speeds. Scandinavian Jour-
nal of Rehabilitative Medicine 21, 197}203.
Veeger, H.E.J., Woude, L.H.V. van der, Rozendal, R.H., 1992. E!ect of
handrim velocity on mechanical e$ciency in wheelchair propulsion.
Medicine and Science in Sports and Exercise 24, 100}107.
Wang, Y.T., 1991. Relationship between kinematic factors and muscle
activation during wheelchair propulsion. Ph.D. Thesis. University
of Illinois at Urbana-Champaign.
Wang, Y.T., Deutsch, H., Morse, M., Hedrick, B., Millikan, T., 1995.
Three-dimensional kinematics of wheelchair propulsion across rac-
ing speeds. Adapted Physical Activity Quarterly 12, 78}89.
Woude, L.H.V. van der, Veeger, D., Rozendal, R.H., Ingen Schenau,
G.J., Rooth, F. van, Nierop, P. van, 1988. Wheelchair racing: e!ects
of rim diameter and speed on physiology and technique. Medicine
and Science in Sports and Exercise 20, 492}500.