Low-velocity motor vehicle collisions are a common cause

Transcription

1 ORIGINAL ARTICLE Electromyographic and Kinematic Exploration of Whiplash-Type Left Anterolateral Impacts Shrawan Kumar, PhD, DSc, FErgS,* Robert Ferrari, MD, FRCPC, FACP, and Yogesh Narayan, BSc(EE), PEng* Background: Volunteer studies of the cervical muscular response and head neck kinematics in response to frontal impacts are uncommon. Moreover, the effect of a frontal impact offset to the left on the resultant muscle responses is unknown. Objective: The purpose of this study was to determine the response of the cervical muscles to increasing low-velocity frontal impacts offset by 45 to the left and to compare the quantitative effects of expected and unexpected impact. Methods: Nine healthy volunteers were subjected to frontal impacts, offset by 45 to the subject s left, of 5.0, 8.7, 11.3, and 15.6 m/s 2 of acceleration at two levels of expectation: expected and unexpected. Bilateral electromyograms (EMGs) of the sternocleidomastoids, trapezii, and splenii capitis were recorded. Triaxial accelerometers recorded the acceleration of the chair, torso at the shoulder level, and head of the participant. Results: Subjects tended to exhibit lower percentages of their maximal voluntary contraction EMG when the impact was expected. When the impact was unexpected, and at an acceleration of 15.6 m/s 2, the splenius capitis muscle contralateral to the impact (ie, right splenius in a left anterolateral impact) generated 95% of its maximal voluntary contraction EMG, whereas the left splenius (ipsilateral to the left anterolateral impact) generated only 43% of this variable. Under these same conditions, the trapezii responded symmetrically, generating approximately 80% of their maximal voluntary contraction EMG. At an acceleration of 15.6 m/s 2, the sternocleidomastoids generated approximately 37% of their maximal voluntary contraction EMG in both the expected and the unexpected impact conditions. EMG (such as time to peak EMG) and kinematic variables were significantly affected by the levels of acceleration (P < 0.001) and expectation status (P < 0.05). The time to onset of the EMG for the splenii capitis and trapezii progressively decreased with increasing levels of acceleration. In response to left anterolateral impacts, muscle responses were greater with higher levels of acceleration and Received for publication June 20, 2003; accepted October 14, From the *Department of Physical Therapy, Faculty of Rehabilitation Medicine, and Department of Medicine, University of Alberta, Edmonton, Alberta, Canada. Reprints: Dr. S. Kumar, Department of Physical Therapy, Faculty of Rehabilitation Medicine, 3-75 Corbett Hall, University of Alberta, Edmonton, AB, Canada T6G 2G4 ( shrawan.kumar@ualberta.ca). Copyright 2004 by Lippincott Williams & Wilkins greatest for the splenius capitis muscle contralateral to the side of impact. The head acceleration response was greater in the unexpected than in the expected condition (P < 0.05). Conclusions: Because the muscular component of the head neck complex plays a central role in the abatement of higher acceleration levels, this is likely a primary site of injury in the whiplash phenomenon in frontal collisions. More specifically, when a frontal impact is offset to the subject s left, it not only results in increased EMG generation in both trapezii, but the splenius capitis contralateral to the direction of impact also bears part of the force of the neck perturbation. Expecting or being aware of imminent impact may play a role in reducing muscle responses in low-velocity anterolateral impacts. Key Words: cervical muscles, electromyography, motor vehicle collisions, frontal impacts, whiplash injury (J Spinal Disord Tech 2004;17: ) Low-velocity motor vehicle collisions are a common cause of whiplash-associated disorders, accounting for up to 50% of claims. 1,2 Although most experimental studies have focused on the mechanism of whiplash-type injury in low rear impacts, 2 9 frontal impacts are also an important source for whiplash claims. 10 It thus seems as important to investigate the mechanism of injury in frontal collisions. One of the clinical authors has noted that real-world frontal impacts may be more heterogeneous than rear impacts. That is, patients reporting symptoms after frontal impacts typically describe a scenario where the collision was not head on but rather offset to the right or left. Anecdotally, at least, these patients tend to emphasize the unilateral nature of their neck pain. Yet, there is no evidence to date that an offset impact has any effect on the resulting muscle response. It is not even clear why this view is so commonly held by patients, as there does not appear to be a preceding scientific groundwork to have led to this tendency. We know relatively little about the individual cervical muscle responses to neck perturbation, especially in terms of asymmetry of responses to neck perturbations. Nevertheless, that asymmetry of a frontal impact is relevant to patients and perhaps clinicians suggests it is worthy of investigation. 412 J Spinal Disord Tech Volume 17, Number 5, October 2004

2 J Spinal Disord Tech Volume 17, Number 5, October 2004 EMG of Whiplash Injury Given ethical concerns with subjecting volunteers to injurious neck perturbations, however, most volunteer collision experiments have been conducted with military personnel and members of the research team. 11 A few other experiments have been done with other volunteer groups, but necessarily limited to low- or very-low-velocity collisions. 2,8,9 To overcome this ethical dilemma and yet conduct investigations to elucidate the kinematics and electromyographic (EMG) response to neck perturbation in human volunteers, Kumar et al 6,7,12 have used surface EMG combined with regression techniques modeled on very-low-velocity collisions. Using this approach, we have conducted rear- and frontal-impact studies, and the regression models are in good agreement with the available data that have been gathered in previous small studies of higher-velocity collisions. 6,7,12 Further, we have been able to show that in straight-on rear impacts, the greatest muscle response arises from the sternocleidomastoids rather than the trapezii or splenius capitis muscles. 6,7 The sternocleidomastoid muscles thus appear to be at the greater risk for injury in low-velocity rear impacts. On the other hand, in straight-on frontal impacts, the trapezius muscles show the greatest muscle response and are most likely to be the first injured. 12 To assess the types of frontal impacts perceived to have a unilateral effect on cervical muscle injury, we undertook a study to assess the cervical muscle response to frontal impacts where the direction of impact is offset 45 to the subject s left. MATERIALS AND METHODS The methods for this study of left anterolateral impacts are the same as those used previously for our straight-on impact studies. 6,7,12 The details have thus been described elsewhere and will be given in brief here. Subjects Nine healthy normal subjects with no history of whiplash injury and no cervical spine pain during the preceding 12 months volunteered for the study. The nine subjects (six women, three men) had a mean age of 25.0 ± 2.5 years, a mean height of 174 ± 3.6 cm, and a mean weight of 73 ± 11.0 kg. All were right hand dominant. The study was approved by the University of Alberta Health Research Ethics Board. Tasks With use of a sled device described elsewhere, 6,7,12 seated and stabilized subjects were exposed to frontal impacts with their seating offset 45 from the direction of sled accelerations of 5.0, 8.7, 11.3, and 15.6 m/s 2 in a random order by the pneumatic piston. The accelerations were delivered under two conditions: Volunteers were either expecting (expected group) or not expecting (unexpected group) the impact. All subjects underwent all levels of accelerations under both expectation conditions. Experimental Setup The acceleration device consisted of an acceleration platform and a sled. The full details of the device are given by Kumar et al. 6,7,12 An abbreviated description follows. The acceleration platform had parallel tracks cm long, mounted lengthwise 60 cm apart. These tracks permitted smooth gliding of the sled on the rails, with a low coefficient of friction (0.03). This assembly allowed a maximum linear speed up to 36 km/h. At one end of the platform, a pneumatic cylinder with a piston stroke length of 30 cm was connected to an air supply and mounted rigidly on the acceleration platform. The device was calibrated for the delivery of known forces causing acceleration of 5.0, 8.7, 11.3, and 15.6 m/s 2. The opposite end of the platform was equipped with a high-density rubber stopper in the sled s path to prevent it from sliding off the platform. The sled consisted of a molded plastic seat with a backrest and four legs mounted to a rectangular sliding board coupled with the tracks for friction-reduced travel on impact. The sled was equipped with a footrest and four buckled straps to stabilize the lower extremities. The seat was fitted with a four-point seat restraint system. The volunteers faced 45 from the direction of travel for all experimental trials. Three highperformance triaxial accelerometers with a full-scale nonlinearity of 0.2% were used in the study. They had a dynamic range of ±5 g, a sensitivity of 500 mv/g, and a resolution of 5 mg within bandwidth DC 100 Hz. Data Acquisition The data acquisition system consisted of an analog-todigital board with a 100-kHz sampling capacity. Each of the nine acceleration channels and six EMG channels as well as the force channel were sampled at 1 khz in real time. The sampled signals were stored on a computer with a large hard disc for storage and processing. The EMG and acceleration data were collected during the experimental trials. The peak and average EMG and acceleration values obtained from these sets of data were subjected to quantitative and statistical analysis. Test Protocol After the experiment was discussed and informed consent obtained, the age, weight, and height of each volunteer were recorded. The volunteers then were seated on the chair and stabilized in neutral spinal posture. Two triaxial accelerometers were fixed to the volunteer: one immediately inferior to the seventh cervical vertebra at the level of the shoulder and the other immediately superior to the glabella region of the frontal bone of the skull. The accelerometers were affixed to the volunteers with strong self-adhesive tapes. The axes of the three accelerometers were aligned with the path of the chair. The pneumatic cylinder was aligned such that the piston head 2004 Lippincott Williams & Wilkins 413

3 Kumar et al J Spinal Disord Tech Volume 17, Number 5, October 2004 of the cylinder and the baseboard of the front of the sled were in contact at an angle of 45. The pneumatic piston delivered the appropriate acceleration to the sled. The subjects in the expected group were informed about the forthcoming impact magnitude in qualitative terms: very slow, slow, medium, and fast. The subjects in the unexpected group were blindfolded and provided a portable stereo with engaging music playing loud enough to block any auditory cues. The data collection was initiated, and after 1 second, the pneumatic piston was fired to accelerate the sled. Data Analysis In the analysis, the sample of volunteers was collapsed across gender because preliminary analysis showed no statistically significant differences in the peak EMG amplitudes between the men and women. The velocity and acceleration of the sled subsequent to the pneumatic piston impact and the rubber stopper impact were calculated. The time of the peak acceleration from the firing of the piston was measured. The data on the peak and average accelerations in all three axes of the sled, shoulder, and head for all four levels of accelerative impacts and for both levels of expectation (expected and unexpected) were measured. For the amplitude analysis, the magnitudes of the full wave-rectified, averaged, and linear envelope-detected EMG signals were subjected to 7-point segment polynomial smoothing repeated once. From such traces, peak EMG, average EMG, and the slope depicting the rise of the EMG traces were obtained. Also, the time relations of the time to onset and peaking of the EMG in relation to the piston firing were measured and analyzed. EMG amplitudes were normalized against the subjects maximal voluntary contraction EMG, these voluntary contraction EMGs having been determined previously with these subjects, providing strength measurement results in newtons for each muscle. 13,14 The ratio (percentage) of the EMG amplitude versus the maximal contraction normalized EMG activity for that subject allowed us to determine the force equivalent generated owing to the impact for each muscle. A statistical analysis was performed using the SPSS statistical package (SPSS, Chicago, IL) to calculate descriptive statistics, correlation analysis between EMG and head acceleration, analysis of variance of the EMG slope, time of peak EMG, time to onset of EMG, time to peak EMG, average EMG, and the force equivalents. RESULTS EMG Amplitude in Right Anterolateral Impacts The mean peak (normalized) EMG amplitudes of the cervical muscles tested in this experiment for the expected and unexpected impacts at each applied acceleration level are presented in Table 1. Figure 1 illustrates the EMG recorded under these conditions. As the level of applied acceleration in a left anterolateral impact increased, the magnitude of the EMG recorded from the right splenius capitis increased progressively and disproportionately compared with the left splenius capitis. The EMG showed a trend toward being greater during unexpected impacts, but this did not reach statistical significance. The normalized EMG showed that the percentage of sternocleidomastoid, splenius capitis, and trapezius magnitude increased steadily with the increasing magnitude of the impact TABLE 1. Mean Normalized Peak EMG of Cervical Muscles in Response to Simulated Left-Offset Frontal Impacts Normalized Muscle EMG Acceleration (m/s 2 ) Sternocleidomastoid (% MVC) Splenius Capitis (% MVC) Trapezius (% MVC) Left Right Left Right Left Right Unexpected (2) 8 (4) 21 (16) 26 (11) 32 (13) 40 (20) (11) 10 (8) 19 (7) 42 (18) 46 (32) 49 (30) (13) 18 (12) 32 (24) 61 (16) 66 (53) 71 (41) (38) 37 (48) 43 (40) 95 (66) 77 (44) 82 (47) Expected (3) 9 (1) 14 (4) 26 (16) 28 (13) 34 (25) (2) 9 (0) 16 (7) 29 (15) 32 (8) 43 (33) (3) 9 (1) 14 (5) 37 (20) 47 (12) 61 (37) (9) 19 (17) 25 (18) 70 (42) 55 (12) 74 (66) Values in parentheses represent 1 SD. MVC, maximal voluntary contraction Lippincott Williams & Wilkins

4 J Spinal Disord Tech Volume 17, Number 5, October 2004 EMG of Whiplash Injury FIGURE 1. Means of peak EMG activity (µv) at two levels of expectation and four levels of applied acceleration. LSCM, left sternocleidomastoid; RSCM, right sternocleidomastoid; LSPL, left splenius capitis; RSPL, right splenius capitis; LTRP, left trapezius; RTRP, right trapezius Lippincott Williams & Wilkins 415

5 Kumar et al J Spinal Disord Tech Volume 17, Number 5, October 2004 FIGURE 2. Normalized average and peak EMG (percentage of isometric maximal voluntary contraction), force equivalent of EMG (N), and levels of expectation and applied acceleration. LSCM, left sternocleidomastoid; RSCM, right sternocleidomastoid; LSPL, left splenius capitis; RSPL, right splenius capitis; LTRP, left trapezius; RTRP, right trapezius Lippincott Williams & Wilkins

6 J Spinal Disord Tech Volume 17, Number 5, October 2004 EMG of Whiplash Injury TABLE 2. Mean Time to Onset (ms) of Acceleration and of Muscle EMG from Firing of Solenoid of Pneumatic Piston in Left Anterolateral Impact Muscle Acceleration (m/s 2 ) Sled Shoulder Head Sternocleidomastoid Splenius Capitis Trapezius Left Right Left Right Left Right Unexpected (8) 119 (23) 135 (22) 901 (1096) 508 (383) 410 (165) 410 (314) 305 (154) 244 (73) (18) 82 (7) 74 (17) 591 (836) 167 (88) 176 (139) 144 (15) 161 (16) 156 (22) (23) 69 (25) 74 (45) 940 (1309) 164 (81) 175 (124) 132 (41) 127 (29) 132 (38) (6) 90 (25) 65 (12) 933 (1099) 809 (1148) 125 (81) 135 (32) 134 (20) 145 (30) Expected (11) 128 (40) 137 (26) 2030 (298) 551 (763) 1058 (1686) 181 (43) 1590 (1638) 117 (29) (17) 73 (18) 69 (10) 1080 (1088) 528 (888) 1012 (1822) 130 (33) 120 (19) 116 (29) (7) 86 (24) 69 (8) 1533 (1321) 709 (1268) 823 (1418) 110 (57) 112 (53) 114 (37) (4) 85 (27) 72 (31) 115 (52) 1144 (1172) 642 (1016) 126 (20) 128 (14) 124 (33) Times for the sled, shoulder, and head represent the times at which acceleration in z axis (direction of travel) began. Times for the cervical muscles represent the time to onset of EMG activity. Values in parentheses represent 1 SD. and generally for both the expected and the unexpected conditions (Table 1; Fig. 2). In a left anterolateral impact, with unexpected condition, at an acceleration of 15.6 m/s 2, the right splenius capitis exerted 95% and the left splenius capitis 43% of the mean normalized maximal voluntary contraction, whereas the left trapezius exerted 77% and the right trapezius 82% of the mean normalized maximal voluntary contraction. Meanwhile, the sternocleidomastoids never exceeded 40% of this variable. In terms of force equivalents, the splenii capitis contralateral to the side of impact were required to resist the impact at a level near its maximal voluntary contraction capability. Timing The time to onset of the sled, shoulder, and head acceleration in the z axis (axis along impact direction) and the EMG signals of the six muscles examined are presented in Table 2. The time to onset was measured from the firing of the pneu- TABLE 3. Mean Time (ms) at which Peak EMG Occurred After Firing of Solenoid of Pneumatic Piston Muscle EMG Acceleration (m/s 2 ) Sternocleidomastoid Splenius Capitis Trapezius Left Right Left Right Left Right Unexpected (427) 456 (287) 732 (387) 307 (89) 262 (92) 445 (330) (1099) 1585 (1126) 1222 (699) 230 (40) 225 (31) 232 (53) (86) 457 (296) 735 (1125) 249 (61) 215 (27) 246 (48) (762) 772 (769) 441 (465) 253 (22) 256 (33) 267 (73) Expected (751) 813 (566) 802 (781) 628 (791) 324 (135) 251 (32) (925) 607 (536) 1432 (969) 224 (108) 1036 (953) 229 (28) (700) 1056 (929) 905 (1212) 299 (22) 213 (22) 240 (27) (287) 620 (687) 497 (473) 224 (50) 214 (26) 289 (62) Values in parentheses represent 1 SD Lippincott Williams & Wilkins 417

7 Kumar et al J Spinal Disord Tech Volume 17, Number 5, October 2004 FIGURE 3. Head acceleration in the x, y, and z axes of one subject in response to the level of expectation and applied acceleration. The z axis is parallel, the x axis orthogonal, and the y axis vertical to the direction of travel. Head X, head acceleration in the x axis; head Y, head acceleration in the y axis; head Z, head acceleration in the z axis Lippincott Williams & Wilkins

8 J Spinal Disord Tech Volume 17, Number 5, October 2004 EMG of Whiplash Injury matic piston. The time of the sled, torso, and head acceleration onset decreased with increased applied acceleration. Similarly, the time to onset of the EMG decreased with increased applied acceleration, but only for the splenii capitis and trapezii. The mean times at which peak EMG occurred for all the experimental conditions are presented in Table 3. Head Acceleration The kinematic response of the head to the four levels of applied acceleration in the expected and unexpected conditions is shown in Fig. 3. As anticipated, an increase in applied acceleration resulted in an increase in excursion of the head and accompanying accelerations. The head acceleration response was greater in the unexpected than in the expected condition (P < 0.05). The relationship between the forceequivalent EMG response of each muscle and the head acceleration is shown in Table 4. Statistical Analyses The applied acceleration and the muscles examined had significant main effects on the peak EMG activity (P < 0.01). To justify the combination of the male and female EMG responses to applied acceleration, gender was entered into the analysis, and the results were nonsignificant, indicating that gender did not confound the results. The levels of acceleration, the muscles examined, and the level of expectation had significant main effects on the time to onset of EMG (P < 0.05). Initially, regression analyses were performed only up to 15.6 m/s 2 using linear, quadratic, cubic, power, and exponential functions. The kinematic variables of head displacement, velocity, and acceleration in response to applied acceleration were calculated. (See Figs. 4 and 5 for splenius muscles.) With use of the obtained regression equations, the responses of the left and right muscle groups were extrapolated to more than twice the applied acceleration value. DISCUSSION In a straight-on front impact, the trapezii primarily respond to the burden of the impact, and the sternocleidomastoids are not significantly active. 12 When the frontal impact is offset 45 to the subject s left, however, part of the impact burden is experienced more by the right splenius capitis, the contralateral muscle. Thus, direction of impact determines which muscles respond and the proportionality of the response among the different muscle groups. In this experiment, although there was a symmetrical response from the trapezii, the splenii capitis muscles behaved asymmetrically. As in straight-on frontal impacts, we also found that being unaware of the impact (unexpected condition) significantly increased aspects of the muscle response and the head acceleration. Studies suggest a central role may be played by the cervical muscles in injury causation during low-velocity collisions. 3,4 Because the muscles are the first in the line of defense for the cervical region, they are likely to be the first in casualty as well. In our study, the data reveal that the splenius capitis muscle contralateral to the side of impact is at greater risk for injury in low-velocity collisions that are offset. The current authors propose that whiplash injuries are complex and progressive. Muscles, ligaments, facet joints, and the brain may be injured in sequence with increasing magnitude of impact. In the series of anterolateral impacts described in this report, although the EMG magnitudes for low acceleration levels were TABLE 4. Mean Force Equivalents and Mean Head Accelerations at Time of Maximal EMG in Direction of Travel for Left Anterolateral Impact Chair Acceleration (m/s 2 ) Head Acceleration (m/s 2 ) Force Equivalents for Muscle (N) Sternocleidomastoid Splenius Capitis Trapezius Left Right Left Right Left Right Unexpected (0.5) 3 (3) 5 (2) 18 (8) 19 (7) 13 (3) 15 (4) (0.8) 5 (3) 5 (2) 20 (4) 25 (11) 16 (7) 17 (7) (0.9) 7 (2) 8 (5) 23 (9) 37 (8) 22 (12) 23 (12) (1.9) 10 (4) 10 (7) 29 (17) 50 (15) 23 (4) 22 (15) Expected (0.3) 5 (6) 4 (12) 11 (10) 23 (17) 17 (12) 18 (9) (0.6) 4 (6) 3 (4) 12 (11) 21 (12) 16 (5) 22 (12) (0.4) 4 (6) 4 (4) 11 (11) 29 (15) 25 (16) 34 (18) (0.8) 8 (7) 6 (8) 18 (16) 48 (33) 29 (22) 43 (35) Values in parentheses represent 1 SD Lippincott Williams & Wilkins 419

9 Kumar et al J Spinal Disord Tech Volume 17, Number 5, October 2004 FIGURE 4. Extrapolated regression plots of the effect that applied acceleration has on the head motion variables of displacement (mm), velocity (m/s), and acceleration obtained (m/s 2 ). small, they rose rapidly as a power function in a nonlinear fashion. At lower acceleration levels, the unexpected conditions engendered a higher level of EMG activity than the expected conditions. This changed even further at the highest accelerations. As we have discussed in detail elsewhere with a review of the relevant EMG data, 6,7 studies in neck perturbations suggest that the cervical muscle response is triggered by peripheral input of muscle stretch. Stretching is likely to be a very effective mechanism for triggering stretch receptors. This does not rule out any central input. Driving a motor vehicle is a learned behavior that involves significant training and conditioning of the somatosensory mechanisms in particular. Therefore, it is likely to have a role in strengthening or modifying the peripheral response. This is clearly shown by the data reported in the current experiment. The time to EMG onset for some muscles in unexpected conditions was later than those in the expected conditions and durations were higher. The regression analysis showed a function relation between the motion variables of the head and the applied and Lippincott Williams & Wilkins

10 J Spinal Disord Tech Volume 17, Number 5, October 2004 EMG of Whiplash Injury FIGURE 5. Extrapolated regression plots of the effect that applied acceleration has on the left and right splenius capitis muscles for the variables of peak EMG (V), normalized EMG (percentage of isometric maximal voluntary contraction), and force equivalent of EMG (N). LSPL, left splenius capitis; RSPL, right splenius capitis. projected acceleration. The projected values are hypothetical and likely to be affected by the ligaments and joint geometry in a manner different from that recorded in the experiment. Nonetheless, the experiment provided a sense of the head s behavior. With additional experiments uncovering more information regarding the modulus of elasticity of various ligaments and capsules, it may be possible to estimate the threshold level or range of acceleration at which injuries are likely to occur. Thus, surface EMG is providing valuable data in conducting whiplash collision experiments. 2,5,6 9 In the setting of controlled collision events and in combination with other objective measurements, surface EMG studies have helped to model the mechanism of acute muscle injury in low-velocity collisions. Ideally, one would devise experiments in which volunteers are subjected to collisions of progressively higher velocities where the injury threshold is reached and surpassed, while EMG data are collected. There have been limited experiments with volunteers where this has been done, but usually the volunteers have been members of the research team or military volunteers. 11 With other volunteer groups, the collision velocities have been necessarily kept at 8 km/h. 8,9 Given the ethical considerations and the difficulties in subjecting volunteers to injury, another approach to the problem that we have been investigating has included the use of regression techniques modeled on very-low-velocity collisions. As the regression models are in good agreement with the 2004 Lippincott Williams & Wilkins 421

11 Kumar et al J Spinal Disord Tech Volume 17, Number 5, October 2004 available data that have been gathered in previous small studies of higher-velocity collisions, the use of regression extrapolation methods may have a role, providing more understanding of what happens to the neck muscles in various collision types and yet avoiding subjecting volunteers to injury itself. In the current study, EMG information has been used to examine the muscle response to very-low-velocity impact and to extrapolate that response to higher-velocity impacts. EMG studies also allow one to examine muscle group responses and patterns, rather than simply describe head or other body region accelerations. The experimental design we have used to study neck perturbations in response to very-low-velocity change is not intended to mimic a collision in a vehicle but rather to allow for the initial exploration of the role of EMG in assessing neck perturbations. Because it is not yet possible to objectively identify the acute whiplash injury thought to underlie grade 1 and 2 whiplash-associated disorders, current injury models are based on evaluation of volunteers in collisions. With so many parameters available for modulation in attempting to approximate road collisions, the task of developing a model for the acute whiplash injury is daunting. One starting place, however, is the use of objective measurements such as EMG in a laboratory setting where other confounding variables have been accounted for or eliminated. In time, more variables can be introduced and studied with this approach. There is no direct way to measure forces exerted by muscles due to neck perturbation and subsequent muscle activity, but examining EMG activity generated allows one to compare this with EMG activity in voluntary contractions. This in turn allows one to relate the muscle responses to normal muscle forces in various physiologic ranges of activity. It is not surprising that subjects in our experiment experienced no adverse symptoms in relation to the experimental design. The velocities used were, by design, meant to avoid potential injury. Moreover, we have previously measured the force exertions of neck muscles in healthy volunteers, and we see that voluntary cervical force exertions exceed the force-equivalent exposures we observed in these low-velocity experiments. Nevertheless, using very low-velocity experimental design, we are able to extrapolate through linear regression to predict the head accelerations and forces likely to be experienced by neck perturbations at higher velocities. Our extrapolations closely match those from small volunteer studies where higher velocities were used with symptoms produced. 11 This suggests that regression techniques may allow for extrapolations into low-velocity ranges and may obviate the need for exceeding ethical concerns with experimental designs that could cause volunteer injury. Additional studies will examine the muscle responses to right anterolateral impacts, right and left lateral impacts, and posterolateral (ie, offset) rear impacts. REFERENCES 1. Spitzer WO, Skovron ML, Salmi LR, et al. Scientific monograph of the Quebec Task Force on Whiplash-Associated Disorders. Spine. 1995;120 (suppl 8):1S 73S. 2. Castro WH, Schilgen M, Meyer S, et al. Do whiplash injuries occur in low-speed rear impacts? Eur Spine J. 1997;6: Szabo TJ, Welcher J. Dynamics of low speed crash tests with energy absorbing bumpers. Proceedings of the Stapp Car Crash Conference, Warrendale, PA, Society of Automotive Engineers. 1992;101: Paper Szabo TJ, Welcher J, Anderson RD. Human occupant kinematic response to low speed rear-end impacts. Proceedings of the Thirty Eighth Stapp Car Crash Conference, Warrendale, PA, Society of Automotive Engineers. 1994: Paper Magnusson ML, Pope MH, Hasselquist L, et al. Cervical electromyographic activity during low-speed rear-end impact. Eur Spine J. 1998;8: L118 L Kumar S, Narayan Y, Amell T. Role of awareness in head-neck acceleration in low velocity rear-end impacts. Accid Anal Prev. 2000;32: Kumar S, Narayan Y, Amell T. An electromyographic study of lowvelocity rear-end impacts. Spine. 2002;27: Siegmund GP, Sanderson DJ, Myers BS, et al. Awareness affects the response of human subjects exposed to a single whiplash-like perturbation. Spine. 2003;28: Brault JR, Wheeler JB, Siegmund GP, et al. Clinical response of human subjects to rear-end automobile collisions. Arch Phys Med Rehabil. 1998; 79: Cassidy JD, Carroll LJ, Cote P, et al. Effect of eliminating compensation for pain and suffering on the outcome of insurance claims for whiplash injury. N Engl J Med. 2000;342: Ferrari R. The Whiplash Encyclopedia. The Facts and Myths of Whiplash. Gaithersburg, MD: Aspen Publishers; 1999: Kumar S, Narayan Y, Amell T. Analysis of low-velocity frontal impacts. Clin Biomech. 2003;18: Kumar S, Narayan Y, Amell T. Cervical strength of young adults in sagittal, coronal, and intermediate planes. Clin Biomech. 2001;6: Kumar S, Narayan Y, Amell T, et al. Electromyography of superficial cervical muscles with exertions in sagittal, coronal, and oblique planes. Eur Spine J. 2002;11: Lippincott Williams & Wilkins