| | An Observational Electromyography Study of the Effect of Trunk Flexion in Low-Velocity Frontal Whiplash-Type ImpactsAbstract Kumar S, Ferrari R, Narayan Y. An observational electromyography study of the effect of trunk flexion in low-velocity frontal whiplash-type impacts. ObjectiveTo examine the effect of forward and lateral trunk flexion on the cervical electromyogram and head kinematic response to whiplash-type frontal impacts. DesignObservational study of sled impacts. ParticipantsTwenty healthy volunteers. InterventionTwenty volunteers were subjected to increasing low-velocity (<8km/h) frontal impacts of 4.4, 7.6, 10.3, and 13.3m/s2 acceleration with trunk forward flexed by 45° and laterally flexed to the right and left by 45°. Main Outcome MeasuresBilateral electromyography of the sternocleidomastoids, trapezii, and splenii capitis and acceleration of the sled, torso, and head were recorded. ResultsWith either direction of lateral trunk flexion at impact, the trapezii electromyographic activity increased with increasing acceleration (P<.05). With the trunk flexed to the left, the left trapezius generated 39% of its maximal voluntary contraction (MVC) electromyographic activity, while the right trapezius generated 31% of its MVC electromyographic activity. The left splenius (ipsilateral to leftward trunk flexion) generated 24% of its MVC electromyographic activity, with all other muscles generating 15% or less of this measure. With the trunk flexed to the right, the right trapezius generated 38% of its MVC electromyographic activity, while the left trapezius generated 32% of this value. Again, the ipsilateral (to trunk flexion) splenius capitis generated 27% of its MVC electromyographic activity, and all other muscles 11% or less of this measure. ConclusionsWhen subjects sit with trunk flexed out of neutral posture at the time of frontal impact, the cervical muscle response is low and unlikely to be injurious.
PHYSIATRISTS, OTHER CLINICIANS, engineers, and the legal profession have increasingly been interested in the biomechanics of whiplash-type impacts, partly because of the need for correlation to clinical measures (impact severity to symptoms and outcome). In particular, physiatrists are frequently assessing and providing medicolegal opinions on people who have sustained whiplash injuries. Whiplash injury is an important health problem with a significant economic and health burden.1 While there has been considerable research on the cervical response to rear-end impacts using volunteers,2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18 much less research has been conducted on frontal impacts, and those that have been done have been conducted mainly with military personnel.19, 20, 21, 22, 23, 24 There is a need for more understanding of frontal impact biomechanics, because frontal collisions are as common a cause of whiplash claims as rear-end collisions.25 Understanding the mechanism of whiplash injury in frontal impacts is relevant also if preventive efforts are to be considered, or if more specific treatment approaches may be designed according to the site of injury.
Given ethical concerns with subjecting volunteers to potentially injurious neck perturbations, to conduct investigations to elucidate the kinematics and electromyographic response to neck perturbation in volunteers, researchers have in recent years used surface electromyography combined with regression techniques applied to data from very-low to low velocity impacts. They have applied this approach to the problem of frontal impacts,24 and our review of the data suggests the extrapolations are in good agreement with the available data that has been gathered in previous, small studies of higher velocity impacts.26 This approach has the advantage of systematically unraveling the effects of the many parameters involved—for example, direction of impact, head position, body position, anthropometrics, bracing, seat design, and restraint design—in a fashion that will avoid the ethical concerns with subjecting volunteers to injurious neck perturbations, although the data only apply to low-velocity (on the order of 8–15km/h) collisions.
The previously reported studies of frontal impact mainly considered the problem of whiplash injury with head and trunk in the neutral posture (ie, the recommended driving posture). The reality is that vehicle occupants are not always positioned in this neutral position at the time of impact. Foret-Bruno et al27 have observed that whiplash victims may be in the trunk-flexed position, and that, at least from collision experiments with anthropometric dummies, this may affect injury risk. They concluded that this may increase the risk not only from impact with the vehicle interior, but through effects of increased cervical extension when the occupant is seated with most of the torso away from the seat and rebounds into the seat after the impact. However, there is as yet no volunteer data that examine the cervical responses of volunteers when they are not seated in the standard, neutral head and trunk posture. To address this void in current knowledge, we undertook an electromyography study to assess the cervical muscle response for frontal impacts where the trunk is flexed forward and to the subject’s right and left to mimic circumstances of “out-of-position” vehicle occupants. By utilizing previous researchers’ methodologies, but altering only this variable, the effect of this variable should be measurable.
Methods  Sample The methods for this study of frontal impacts with trunk flexion are similar to that used by others reporting frontal impact studies with the subject in neutral posture.24 Twenty healthy subjects (10 men, 10 women) with no history (by self-report) of whiplash injury and no cervical spine pain during the preceding 12 months volunteered for the study. This was a sample of convenience. The 20 subjects had a mean age of 23.6±3.0 years, a mean height of 172±7.7cm, and a mean weight of 69±13.9kg. Subjects were not examined by radiography. All subjects were right-hand dominant. The study was approved by the university research ethics board. Tasks and Data Collection Active surface electrodes with 10 times on-site amplification were placed on the belly of the sternocleidomastoids, the upper trapezius at C4 level, and the splenius capitis in the triangle between the sternocleidomastoids and trapezii bilaterally. The fully isolated amplifier had additional gain settings up to 10,000 times with a frequency response direct current bandwidth of 0 to 5kHz and a common mode rejection ratio of 92dB. Before calibrating sled acceleration, the cervical strength of the volunteers was measured to develop a force-electromyography calibration factor, utilizing a method reported in the literature.28, 29 The seated and stabilized subjects exerted their maximum isometric effort in attempted flexion, extension, and lateral flexion to the left and the right for force-electromyography calibration, as described in the literature on maximal voluntary contraction (MVC) measurements.28, 29 The acceleration device consisted of an acceleration platform and a sled. The device is as shown in figure 1. After the experiment was discussed and informed consent obtained, the age, weight, and height of each volunteer was recorded. Volunteers then were seated on the chair and restrained with a lap-only seat belt to allow subjects to be positioned with trunk flexion. The chair was rigid, so as to minimize any effect of elastic properties of the chair following acceleration. Although this chair differs substantially from an automobile seat, we have conducted experiments in rear and lateral impacts where no differences were found in the electromyographic responses under similar experimental conditions for the 2 seat types.30 Thus, it is unlikely that the chair itself has an important effect on the subject response under these conditions. Subjects were then outfitted with triaxial accelerometersa on their glabella and the first thoracic spinous process. Another triaxial accelerometer was mounted on the sled, not the chair. The accelerometers had a full scale nonlinearity of 0.2%, dynamic range of ±5g, with a sensitivity of 500mV/g, resolution of 5mg within a direct current bandwidth of 100Hz, and a silicon micro-machined capacitive beam that was quite rugged and extremely small in die area. We then exposed subjects to frontal impacts with their trunk flexed forward and either to their left and right at accelerations of 4.4, 7.6, 10.3, and 13.3m/s2 generated in a random order by a pneumatic piston. The random variation in magnitude of acceleration was used to avoid acclimation of the subjects to the impact exposures. We asked subjects to assume a position of trunk flexion (forward and lateral) and to look down at their right or left foot. There was no attempt to have the subjects completely relaxed with the neck fully flexed (ie, slumped posture), because we expected this would not be typical of road collisions. Subjects were positioned in 45° of flexion and 45° of flexion either to the left or to the right (fig 2). There was no blocking of visual or auditory cues, which is the same as the “expected” impact data researchers have reported previously,24 but the 4 available accelerations for impact were randomly varied for the subjects’ exposures, so they could not anticipate the acceleration severity for an impending impact. To release the piston, the solenoid of the pneumatic system was activated by an electronic impulse, which was recorded for timing reference. On delivery of impact by the pneumatic piston, the sled moved on 2 parallel tracks mounted 60cm apart. The coefficient of friction of the tracks was .03, which allowed for smooth gliding of the sled on the rails. The opposite end of the track was equipped with nonlinear springs and a high density rubber stopper to prevent the subject from sliding off the platform. Each subject effectively underwent 4 levels of accelerative impacts under 2 conditions of trunk flexion, for 1 direction of impact (total impacts, 8). The trunk flexion itself did not appear to place the head in a more forward position. Although the subjects were asked to have a flexed trunk prior to impact, nothing was done to fix the position, and the head and trunk were left free to move after impact. The accelerations involved in this experiment were low enough that injury was not expected. The acceleration was delivered in a way that mimicked the time course seen in motor vehicle collisions and occurred fast enough to produce eccentric muscle contractions. The acceleration impulse reached its peak value in 33ms similar to acceleration impulse of vehicle to vehicle impact.8, 9 The sled acceleration impulses are shown in figure 3. Subjects were asked to report any headache or other aches they experienced in the days after the impacts for a period of 6 months. None were reported. Data Analysis We measured the data on the peak and average accelerations in all 3 axes of the sled, shoulder, and head for all 4 levels of accelerative impacts. The gravity bias was eliminated by subtracting the base value for the posture from the accelerometer readings. However, we made no effort to compensate for motion subsequent to the impact. The onset of acceleration was measured by dropping the ascending slope line on the baseline. The point of intersection of these lines was considered as onset of acceleration and time was measured with reference to the solenoid firing. In the analysis, the sample of volunteers was collapsed across sex because preliminary analysis of normalized peak electromyographic activity showed no statistically significant differences in the electromyographic amplitudes between the men and women. The sled velocity and its acceleration subsequent to the pneumatic piston impact were measured. All timing data (time to onset of electromyographic and peak electromyographic activity) were referred to the solenoid of the piston firing. The time of the peak accelerations of sled and head were measured. Also, the time relations of the onset and peak of the electromyographic activity were measured and analyzed. The time to onset was determined when the electromyographic perturbation reached 2% of the peak electromyographic value to avoid false positives due to tonic activity. This method was in agreement with projection of the line of slope on the baseline. Electromyographic amplitudes were normalized against the subjects’ MVC electromyograms. The ratio percentage of the electromyographic amplitude versus the maximal contraction normalized electromyographic activity for that subject allowed us to determine the force equivalent generated due to the impact for each muscle. We performed statistical analysis using the SPSS statistical packageb to calculate descriptive statistics, correlation analysis between electromyographic activity and head acceleration, analysis of variance of the time to electromyographic activity onset, time to peak electromyographic activity, and average electromyographic activity. Results Electromyographic Amplitude In a frontal impact, with the trunk flexed 45° to the right or left, both the trapezius and splenius capitis muscles ipsilateral to the direction of trunk flexion showed a significant effect of acceleration (P=.02). These muscles showed a greater electromyographic response compared with other muscles, and the trapezii were the most active of all muscles (P<.02). The normalized electromyographic activity for the sternocleidomastoid, splenius capitis, and trapezius muscles are shown in figure 5. As the level of applied acceleration in the impact increased, the magnitude of the electromyographic activity recorded from the trapezii and ipsilateral splenius capitis increased progressively and disproportionately compared with other muscles (P<.05). In a frontal impact, with trunk flexed to the right, the most significant finding is that the muscle responses were generally of low magnitude (≤38% for all muscles), the right trapezius muscle having the highest electromyographic activity at 38% of MVC, followed by the left trapezius (32%), right splenius capitis (26%), left splenius capitis (11%), and both right and left sternocleidomastoids at 5% of their MVC. The ipsilateral splenius capitis (the right one in right trunk flexion) thus had a greater electromyographic response than its counterpart. The same occurred when the trunk was flexed to the left, where the muscles had a 39% or less normalized peak electromyographic response, the left trapezius muscle having the highest electromyographic activity at 39% of MVC, followed by the right trapezius (31%), the left splenius capitis (24%), the right splenius capitis (12%), and both the right and left sternocleidomastoids at 5% of their MVC. Again, the ipsilateral (to trunk flexion) splenius capitis had a higher response than its contralateral counterpart. To place the magnitude of the electromyographic responses in perspective, we compared the normalized electromyographic percentages of the trapezii for 3 conditions: head and trunk in neutral posture, and trunk flexed forward and laterally right or left. The comparison data are from a study whose methodology we have here replicated, except that we have added the variable of occupant positioning.24 Compared with the state of the head and trunk in neutral posture, trunk flexion significantly reduces the trapezius electromyographic response (P<.05), except for the 1 instance of the left trapezius in impacts with left trunk flexion showing a similar electromyographic response to when the trunk is in neutral posture. Although the data concerning electromyographic responses with the head in neutral posture are from a different group of subjects, the methodology of always normalizing the electromyographic response to an individual’s MVC helps to adjust for these variables (ie, sex, stature, and age affects MVC, and electromyographic responses should thus be normalized before making comparisons among individuals or groups). Timing The time to onset of the sled, torso, and head acceleration (the timing data is in relation to firing of the solenoid of the piston) showed a trend toward being decreased with increased applied acceleration. Similarly, the time to onset of the electromyographic activity decreased with increased applied acceleration. The mean times at which peak electromyographic activity occurred for all the experimental conditions also showed a trend (P=.06) to earlier times of peak activity with increasing acceleration, but this again did not reach statistical significance. The kinematic responses show that very low velocity impacts produce less force equivalent than the MVC for the same subject, and thus this experimental approach allows us to gather valuable data without exposing subjects to any foreseeable injury. The head accelerations were correspondingly lower than the sled accelerations in this experiment. For very low velocity impacts, this is to be expected, as it is usually only when the sled acceleration exceeds 5g that head acceleration begins to exceed sled acceleration. This experiment involved less than 2g of accelerations. Discussion In frontal impacts, whiplash victims may be leaning forward or leaning over as a result of watching for traffic or speaking with other occupants, reaching for an object on the floor, and so forth. In the current study, where we have used electromyographic measurements to study the cervical muscle response when the trunk is flexed forward and to the right or left at the time of impact, we find that the trapezii generate the greatest electromyographic response, but in general all muscles show low electromyographic magnitudes. Only the trapezii, bilaterally, and the ipsilateral splenius capitis showed significant interaction with acceleration (P<.02), indicating that other muscles acted as stabilizers. One may compare these results to previous studies we have conducted with exactly the same methodology, other than the fact that we examined the magnitudes of the electromyographic response with the head and trunk in neutral posture.24 These subjects did not differ from the current study in mean age, but, moreover, because we normalize all peak electromyographic measurements to the subject’s MVC, we can create useful comparison groups with this methodology. When the head and trunk are in neutral posture, under similar study protocols, both the trapezii generate electromyographic activity on the order of 50% of their MVC electromyographic values.24 As seen in this experiment, even the most active muscles do not exceed 39% of their maximal electromyographic contraction magnitude. That the splenius capitis ipsilateral to the direction of trunk flexion is more active than its counterpart may reflect the fact that it is already stretched by this posture, and thus primed to respond to acceleration by a greater contraction than the contralateral splenius capitis. The sternocleidomastoid muscles, by their attachment and action, are least likely to undergo eccentric contraction in the presence of an expected reduction in head-torso lag in the trunk-flexed posture. In contrast, the attachment and action of the trapezii, with cervical extension being 1 such action, are likely to be in a prestretched position in the trunk flexed posture with the subject looking downward. The reason why the ipsilateral trapezius undergoes more electromyographic activity is unclear, and more study of this phenomenon is needed. Even lower than expected head-torso lag in this posture is thus expected to generate more response and a higher likelihood of eccentric contraction in the trapezii than the sternocleidomastoids. It is suggested that the forward-flexed trunk posture does not increase the likelihood of cervical muscle injury as compared with impacts with the trunk in neutral position, at least not for low-velocity impacts. This is contrary to previous research findings.28 Previous research, however, focused on dummy responses, which may explain the difference in our findings, and also some of the dummy experiments were of much higher velocity impacts. Nevertheless, symptoms are reported even after low-velocity impacts, and these lead to as many as 60% of injury claims.16 With low-velocity impacts, one does not expect any significant rebounding of the subject back into the seat. Conclusions The generalizability of the observations from this study are limited by the fact that it was necessary to use a lap-seat belt only in order to test the effect of trunk flexion: the scenario may not mimic many cases of frontal road collisions. An additional limitation of the study is that due to the lack of kinetic data, accurate gravity compensation could not be applied. Further studies with this approach are needed to determine its full potential, but ultimately it may be necessary to verify this data with higher-velocity gravity adjusted impact studies. Suppliers
References 1.
1
Spitzer WO
, Skovron ML
, Salmi LR
, et al.
Scientific monograph of the Quebec Task Force on Whiplash-Associated Disorders
(redefining “whiplash” and its management)
.
Spine
. 1995;120(Suppl 8):1S–73S
.
2.
2
West DH
, Gough JP
, Harper GT
.
Low speed rear-end collision testing using human subjects
.
Accid Reconstr J
. 1993;5:22–26
.
3.
3
McConnell WE
, Howard RP
, Guzman HM
, et al.
Analysis of human test subject kinematic responses to low velocity rear end impacts
.
In:
Proceedings of the Thirty-Seventh Stapp Car Crash Conference
. Warrendale: Society of Automotive Engineers; 1993;p. 21–30
.
4.
4
McConnell WE
, Howard RP
, Van Poppel J
, Krause R
, Guzman HM
, Bomar JB
.
Human head and neck kinematics after low velocity rear-end impacts—understanding “whiplash”
.
In:
Proceedings of the Thirty-Ninth Stapp Car Crash Conference
. Warrendale: Society of Automotive Engineers; 1995;p. 215–238
.
5.
5
Scott MW
, McConnell WE
, Guzman HM
, Howard RP
, Bomar JB
, Smith HL
.
Comparison of human and ATD head kinematics during low-speed rearend impacts
.
In:
Proceedings of the Thirty-Seventh Stapp Car Crash Conference
. Warrendale: Society of Automotive Engineers; 1993;p. 1–8
.
6.
6
Siegmund GP
, Bailey MN
, King DJ
.
Characteristics of specific automobile bumpers in low-velocity impacts
.
Proceedings of the Thirty-Eighth Stapp Car Crash Conference
. Warrendale: Society of Automotive Engineers; 1994;
.
7.
7
Siegmund GP
, Williamson PB
.
Speed change (Δ v) of amusement park bumper cars
.
In:
Proceedings of the Canadian Multidisciplinary Road Safety Conference VIII
. 1993;p. 299–308
.
8.
8
Szabo TJ
, Welcher J
, Anderson RD
.
Human occupant kinematic response to low speed rear-end impacts
.
In:
Proceedings of the Thirty-Eighth Stapp Car Crash Conference
. Warrendale Society of Automotive Engineers; 1994;p. 23–35
.
9.
9
Szabo TJ
, Welcher J
.
In:
Dynamics of low speed crash tests with energy absorbing bumpers
. Warrendale: Society of Automotive Engineers; 1992;p. 1–9
.
10.
10
Matsushita T
, Sato TB
, Hirabayashi K
, Fujimura S
, Asazuzma T
, Takatori T
.
X-ray study of the human neck motion due to head inertia loading
.
In:
Proceedings of the Thirty-Eighth Stapp Car Crash Conference
. Warrendale: Society of Automotive Engineers; 1994;p. 55–64
.
11.
11
Rosenbluth W
, Hicks L
.
Evaluating low-speed rear-end impact severity and resultant occupant stress parameters
.
J Forensic Sci
. 1994;39:
.
12.
12
Nielsen GP
, Gough JP
, Little DM
, West DH
, Baker VT
.
Human subject responses to repeated low speed impacts using utility vehicles
.
In:
Proceedings of the Forty-First Stapp Car Crash Conference
. Warrendale: Society of Automotive Engineers; 1997;p. 189–212
.
13.
13
Brault JR
, Wheeler JB
, Siegmund GP
, Brault EJ
.
Clinical response of human subjects to rear-end automobile collisions
.
Arch Phys Med Rehabil
. 1998;79:72–80
.
Abstract |
Full-Text PDF (1080 KB)
|
CrossRef
14.
14
Howard RP
, Bowles AP
, Guzman HM
, Krenrich SW
.
Head, neck, and mandible dynamics generated by “whiplash”
.
Accid Anal Prev
. 1998;30:525–534
.
MEDLINE |
CrossRef
15.
15
Magnusson ML
, Pope MH
, Hasselquist L
, et al.
Cervical electromyographic activity during low-speed rear-end impact
.
Eur Spine J
. 1998;8:118–125
.
MEDLINE |
CrossRef
16.
16
Castro WH
, Schilgen M
, Meyer S
, Weber M
, Peuker C
, Wortler K
.
Do “whiplash injuries” occur in low-speed rear impacts?
.
Eur Spine J
. 1997;6:366–375
.
MEDLINE |
CrossRef
17.
17
Castro WH
, Meyer SJ
, Becke ME
, et al.
No stress—no whiplash? Prevalence of “whiplash” symptoms following exposure to a placebo rear-end collision
.
Int J Legal Med
. 2001;114:316–322
.
MEDLINE |
CrossRef
18.
18
Kumar S
, Narayan Y
, Amell T
.
An electromyographic study of low-velocity rear-end impacts
.
Spine
. 2002;27:1044–1055
.
CrossRef
19.
19
Ewing CL
, Thomas DJ
.
Torque versus angular displacement response of human head to 2Gx impact acceleration
.
In:
Proceedings of the Seventeenth Stapp Car Crash Conference
. Warrendale: Society of Automotive Engineers; 1973;p. 309–342
.
20.
20
Ewing CL
, Thomas DJ
, Lustick L
, Becker E
, Willems G
, Muzzy WH
.
The effect of the initial position of the head and neck on the dynamic response of the human head and neck to Gx impact acceleration
.
In:
Proceedings of the Nineteenth Stapp Car Crash Conference
. Warrendale: Society of Automotive Engineers; 1975;p. 487–512
.
21.
21
Wagner R
.
A 30 mph front/rear crash with human test persons
.
In:
Proceedings of the Twenty-Third Stapp Car Crash Conference
. Warrendale: Society of Automotive Engineers; 1979;p. 827–840
.
22.
22
Wismans J
, Philippens M
, van Oorschot E
, Kallieris D
, Mattern R
.
Comparison of human volunteer and cadaver head-neck response in frontal flexion
.
In:
Proceedings of the Thirty-First Stapp Car Crash Conference
. Warrendale: Society of Automotive Engineers; 1987;p. 1–11
.
23.
23
Grunsten RC
, Gilbert NS
, Mawn SV
.
The mechanical effects of impact acceleration on the unconstrained human head and neck complex
.
Contemp Orthop
. 1989;18:199–202
.
24.
24
Kumar S
, Narayan Y
, Amell T
.
Analysis of low-velocity frontal impacts
.
Clin Biomech (Bristol, Avon)
. 2003;18:694–703
.
Abstract | Full Text |
Full-Text PDF (519 KB)
|
CrossRef
25.
25
Cassidy JD
, Carroll LJ
, Cote P
, Lemstra M
, Berglund A
, Nygren A
.
Effect of eliminating compensation for pain and suffering on the outcome of insurance claims for whiplash injury
.
N Engl J Med
. 2000;342:1179–1186
.
MEDLINE |
CrossRef
26.
26
Ferrari R
.
In:
The whiplash encyclopedia
(the facts and myths of whiplash)
. Gaithersburg: Aspen; 1999;p. 449–470
.
27.
27
Foret-Bruno JY
, Tarriere C
, Le Coz JY
, Got C
, Guillon F
.
Risk of cervical lesions in real-world and simulated collisions
.
In:
Proceedings of the Thirty-Fourth Conference of the American Association of Automotive Medicine
. Warrendale: Society of Automotive Engineers; 1990;p. 373–389
.
28.
28
Kumar S
, Narayan Y
, Amell T
.
Cervical strength of young adults in sagittal, coronal, and intermediate planes
.
Clin Biomech (Bristol, Avon)
. 2001;6:380–388
.
29.
29
Kumar S
, Narayan Y
, Amell T
, Ferrari R
.
Electromyography of superficial cervical muscles with exertions in sagittal, coronal, and oblique planes
.
Eur Spine J
. 2002;11:27–37
.
MEDLINE |
CrossRef
30.
30
Kumar S
, Ferrari R
, Narayan Y
, Jones T
.
The effect of seat belt use on the cervical electromyogram response to whiplash-type impacts
.
J Manipulative Physiol Ther
. 2006;29:115–125
.
Abstract | Full Text |
Full-Text PDF (1312 KB)
|
CrossRef
a Department of Physical Therapy, Faculty of Rehabilitation Medicine, University of Alberta, Edmonton, AB, Canada b Department of Medicine, University of Alberta, Edmonton, AB, Canada  Reprint requests to Shrawan Kumar, PhD, FRS(C), 3-75 Corbett Hall, Dept of Physical Therapy, Faculty of Rehabilitation Medicine, University of Alberta, Edmonton, AB T6G 2G4, Canada
No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. PII: S0003-9993(05)01533-9 doi:10.1016/j.apmr.2005.12.034 © 2006 American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation. Published by Elsevier Inc. All rights reserved. | |
|
FULL TEXT01533-9/journalimage?img=PIIS0003999305015339.gr1.lrg.gif&fig=fig1&kwhquery=&issn=0003-9993&ishighres=false&allhighres=false&free=yes','journalimage');)
01533-9/journalimage?img=PIIS0003999305015339.gr2.lrg.gif&fig=fig2&kwhquery=&issn=0003-9993&ishighres=false&allhighres=false&free=yes','journalimage');)
01533-9/journalimage?img=PIIS0003999305015339.gr3.lrg.gif&fig=fig3&kwhquery=&issn=0003-9993&ishighres=false&allhighres=false&free=yes','journalimage');)
01533-9/journalimage?img=PIIS0003999305015339.gr4.lrg.gif&fig=fig4&kwhquery=&issn=0003-9993&ishighres=false&allhighres=false&free=yes','journalimage');)
01533-9/journalimage?img=PIIS0003999305015339.gr5.lrg.gif&fig=fig5&kwhquery=&issn=0003-9993&ishighres=false&allhighres=false&free=yes','journalimage');)