- Email: [email protected]
Self-initiating a seated perturbation modifies the neck postural responses in humans
Neuroscience Letters 347 (2003) 1–4 www.elsevier.com/locate/neulet
Self-initiating a seated perturbation modifies the neck postural responses in humans Jean-Se´bastien Blouin, Martin Descarreaux, Ariane Be´langer-Gravel, Martin Simoneau, Normand Teasdale* Universite´ Laval, Division de Kine´siologie, Faculte´ de Me´decine, Que´bec, Canada Received 28 October 2002; received in revised form 15 February 2003; accepted 13 April 2003
Abstract When seated subjects are submitted to a linear acceleration, reports indicate that the kinematic and electromyographic (EMG) responses of the head – neck system can be modulated with the magnitude of the linear acceleration. There is no evidence, however, that head kinematics or neck EMG activity can be modulated when specific knowledge and active control about the onset of platform acceleration are available. Sixteen seated subjects were given forward linear accelerations in two different conditions nested within subjects: reactive and predictive. In the reactive condition, the acceleration was initiated following a variable delay unknown to the subjects whereas in the predictive condition, subjects manually self-initiated the perturbation. All neck muscle activities were decreased 50 –100 ms after platform movement onset in the predictive condition relative to the reactive condition, whereas head and neck peak angular positions and velocities were not different between the two conditions. These results suggest that feedforward control could use the self-generated timing information of platform movement onset to scale the appropriate neck motor output. q 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Seated perturbation; Timing; Cognition; Neck electromyographic activity attenuation
Head control and stability rely upon the integration of proprioceptive, visual and vestibular sensory inputs [7,8]. When subjects in a seated position are submitted to linear accelerations (simulating rear-end car accidents), the kinematic and EMG responses of the head – neck system are modulated with the magnitude of the acceleration. Generally, with an increased magnitude of the perturbation, shorter-latency EMG responses and greater magnitude of kinematic and EMG responses are observed [2,4,9]. The short latency of these responses (about 70 ms) suggests that they are triggered mainly by postural reflexes [4,9]. The possible role of cognition (expectation) for modulating the head –neck responses remains unclear. First, Kumar et al. [5] showed that when subjects had prior knowledge of the amplitude of the upcoming perturbation, peak head acceleration decreased and time to peak head acceleration increased. In a later experiment, Kumar et al. [4] showed that the amplitude of EMG responses also was attenuated when subjects had prior knowledge of the amplitude of the * Corresponding author. Universite´ Laval, PEPS Local 00232, Division de Kine´siologie, Que´bec, Que´bec, Canada G1K 7P4. Tel.: þ 1-418-6562147; fax: þ 1-418-656-2441. E-mail address: [email protected] (N. Teasdale).
acceleration. Conversely, Siegmund et al. [9] showed that neither the neck muscle responses nor the head kinematic responses were modified by prior knowledge of the amplitude of platform acceleration. These divergent results are difficult to reconcile and leave open the questionable role of cognition on the modulation of seated postural responses. The above experiments [4,9] examined the effects of knowing the magnitude of the upcoming perturbation on the head –neck responses. The temporal onset of the perturbation, however, was unknown to the subjects. Subjects can benefit from specific temporal information for modulating spinal reflexes [3]. There is no evidence that the neck neuromuscular responses can be modulated when subjects self-initiate a forward linear acceleration. The aim of this study was to determine if the neuromuscular responses of the head –neck system can be modified when specific knowledge and active control about the onset of platform acceleration are available to the subjects. We hypothesized that when the onset of the linear acceleration is self-initiated, the nervous system could use this timing information to minimize muscle activation in order to reduce muscle stress at the cervical level. Sixteen healthy subjects, six women and ten men (on
0304-3940/03/$ - see front matter q 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0304-3940(03)00632-3
2
J.-S. Blouin et al. / Neuroscience Letters 347 (2003) 1–4
average ^ SD, 23 ^ 3 years old), without any history of whiplash or neck pain participated in this experiment. All subjects signed a written consent form approved by the local ethics committee. Subjects were submitted to 16 forward whole body linear accelerations in two different conditions: reactive and predictive. For the reactive condition, after a ready verbal signal, the platform moved forward (simulating a small rearend car accident) following a variable delay of 0.5 –5 s unknown to the subjects. For the predictive condition, subjects self-initiated the perturbation by pushing a button with their thumb. Eight trials were performed for each condition. Both conditions were performed on the same day and the time interval between trials was about 1 min for all subjects. The conditions were performed in block (reactive, predictive) and the order of presentation was randomized across subjects. Subjects were comfortably seated in a normal car seat without a seat belt or head rest (see Ref. [1] for a general description of the apparatus). The car seat was mounted on a metal plate fixed on three linear bearings. A steel cable attached to an electromagnet kept the platform from moving. Another steel cable attached at one end of the platform ran through a frictionless pulley and held up a heavy weight. By turning off the electromagnet, the platform accelerated forward with a total displacement of 80 mm at a maximum linear acceleration of 1.1 ^ 0.1 g. An electrical contact at the rear-end of the platform detected the onset of the forward displacement. Total recording lasted 2 s and started with the deactivation of the electromagnet. Platform, neck and head displacements were recorded using a SELSPOT II System (Innovision Systems) with four cameras at a sampling rate of 500 Hz. Infrared diodes were placed on the right side of the subjects: tragus, outer canthus, transverse process of C4 (neck mid-line at mandibular angle level) and spinous process of C7. In addition, one diode was placed on the seat. Position data were filtered using a cubic spline function minimizing residuals. A uni-axial accelerometer (Entran Devices Inc.) was placed on top of each subject’s head using a custom-made head device (total weight: 123 g); it served to date as accurately as possible the onset of the head movement. Linear or rotational movements of the head were calculated using signals from the active markers. Peak angular position and velocity were measured for three head or neck angles: (1) tragus and outer canthus with respect to the horizontal (head angle); (2) tragus and C4 with respect to the vertical (head-C4 angle); and (3) C4 and C7 with respect to the vertical (C4-C7 angle). Time to peak angular position and velocity were also calculated. Bipolar disposable surface Ag-AgCl electrodes (3 mm, Meditrace) were applied bilaterally over the surface of the sternocleidomastoideus (SCM), anterior scalenus (hereafter called scalenus), paraspinals (at the level of C4) and trapezius muscles. The signals were pre-amplified at the source (gain ¼ 500) prior to a second level amplification
(Bortec AMT-8; band-pass of 10 Hz to 1 kHz). Signals from the accelerometer, electrical contact and EMG were sampled at 2 kHz (12-bit A/D resolution). The onset of the platform movement (detected from the electrical contact) was defined as time zero. To examine whether the EMG activities were modulated both prior to and following the perturbation, EMG activities were integrated for nine consecutive 25 ms intervals: from 2 25 ms prior to 200 ms after the platform onset. Integrated EMG (iEMG) values were expressed as a percentage of maximal EMG activity monitored for the first trial in the reactive condition (12.5 ms before and after peak activity of the initial burst). EMG signals were full-wave rectified to determine the onset of the muscular responses. All onsets were determined visually by inspecting the EMG signals. An onset was determined when the activity clearly exceeded that of the baseline variability level for more than ten consecutive samples (5 ms). A similar procedure was used to document the onset of head movement using the accelerometer signal. EMG latency values were measured with respect to platform movement onset. For all dependent variables, within-subject mean of the reactive trials and within-subject mean of the predictive trials were first obtained. Kinematic variables were submitted to a one-way repeated measures ANOVA (Condition – two levels: predictive and reactive). For the EMG latencies, a Side factor (left – right) was added (Condition £ Side). The iEMG data were submitted to a three-way repeated measures ANOVA (Condition £ Side £ Interval). We initially tested for a gender effect and an order effect of conditions. None of these effects were significant (P . 0:05) and they are not presented for the sake of brevity. Prior to each perturbation, the head angular position did not vary across conditions (on average ^ SD, 26.6 ^ 7.78, P . 0:05). The onset of head movement relative to the platform movement also did not vary (on average, 21 ^ 7 ms, P . 0:05). For all three angles, peak angular position and velocity did not vary with onset awareness (P . 0:05 for all analyses). On average, peak angular position and velocity were 9.6 ^ 3.08 and 122.7 ^ 29.78/s for the head angle, 11.9 ^ 3.58 and 160.5 ^ 31.68/s for the head-C4 angle, and 6.5 ^ 2.38 and 81.5 ^ 19.78/s for the C4-C7 angle. Timing values for the head and head-C4 angles, however, showed longer time to peak angular positions for the predictive than the reactive condition (201 vs. 189 ms, F ¼ 6:50, P , 0:05 for the head angle; 193 vs. 175 ms, F ¼ 15:96, P , 0:001 for the head-C4 angle). The time to peak angular velocities did not vary for these angles (on average, 114 ^ 9 ms, P . 0:05 for the head angle; 105 ^ 6 ms, P . 0:05 for the head-C4 angle). Conversely, for the C4-C7 angle the time to peak angular position did not vary between conditions (on average, 162 ^ 41 ms, P . 0:05) but the time to peak angular velocity decreased in the predictive compared to the reactive condition (60 vs. 64 ms, F ¼ 7:20, P , 0:05). Overall, the kinematic responses suggest that the subjects
J.-S. Blouin et al. / Neuroscience Letters 347 (2003) 1–4
were able to maintain the head –neck within similar safety boundaries but modified the timing of the head – neck kinematic responses across conditions. Fig. 1 illustrates the iEMG activities for all muscles across the nine intervals for both conditions. The ANOVAs revealed a significant interaction of Interval £ Condition for all muscles: scalenus (F ¼ 32:17, P , 0:001), SCM (F ¼ 29:80, P , 0:001), paraspinals (F ¼ 9:27, P , 0:001) and trapezius (F ¼ 6:81, P , 0:01). Main effects of Condition and Interval also were significant for all muscles (P , 0:01). Decomposition of the interactions showed that, for all muscles, normalized iEMG activity for the 50 – 75 ms and 75 –100 ms intervals was smaller for the predictive than for the reactive condition (significant Tukey comparison of means for all muscles, P , 0:05). For the paraspinals, a smaller iEMG activity for the predictive condition also was observed for the 100 – 125 ms and 125 –150 ms intervals (significant Tukey comparison of means, P , 0:05). For all muscles, the ANOVAs showed no effect of Side nor any interaction of Side £ Condition, Side £ Interval or Side £ Condition £ Interval (P . 0:05).
3
Hence, this suggests that 50 –100 ms after the perturbation, the activity for all muscles recorded was significantly smaller for the predictive than for the reactive condition. Muscle onsets were computed only for the scalenus, SCM and paraspinals muscles because the EMG activity of the trapezius was difficult to discern from the background activity in the predictive condition. All muscle onsets are reported relative to the platform movement onset since the onset of head movement did not vary for both conditions (see above). For the scalenus and SCM muscles, the latencies for the predictive condition were decreased compared to the reactive condition (45 vs. 52 ms (F ¼ 53:54, P , 0:001) for the scalenus; 46 vs. 54 ms (F ¼ 37:82, P , 0:001) for the SCM). The ANOVA revealed no main effect of Condition for the paraspinals muscles (on average, 58 ^ 8 ms, P . 0:05). No main effect of Side nor any interaction Side £ Condition (P . 0:05) were observed for all muscles. A primary goal of this study was to determine if the nervous system could use the timing information from the onset of platform movement to reduce neck muscular
Fig. 1. Mean normalized integrated EMG activity (for all muscles) between 225 ms prior to platform onset and 200 ms after platform onset are illustrated for the predictive and reactive conditions. Normalized iEMG activities are expressed as a percentage of maximal integrated EMG interval (12.5 ms before and after peak initial burst of activation) calculated for the first reactive trial (see text). Error bars represent 0.95 confidence intervals.
4
J.-S. Blouin et al. / Neuroscience Letters 347 (2003) 1–4
responses. For all muscles, the neck EMG responses were attenuated from 50 to 100 ms after the onset of platform movement for the predictive condition while the preperturbation EMG activity (2 25 to 25 ms after platform movement onset) remained unchanged. This can not be explained by a forward tilt of the head in anticipation of the perturbation since the initial head angular position prior to each perturbation did not vary across conditions. Furthermore, when subjects self-initiated the perturbation, the onset of EMG responses occurred 7 and 8 ms earlier for the scalenus and SCM, respectively. This suggests that the nervous system could use the timing information generated from the thumb motor command initiating the perturbation to scale the appropriate motor output. It supports previous findings by Kumar et al. [4] and suggests that decreasing the uncertainty about the upcoming perturbation could help the nervous system to use the advance information about this upcoming event in order to modify the postural set. It remains to be determined, however, if subjects could still scale their postural responses if an external paced signal was used to inform them about the acceleration onset. Considering the early latency of the responses (45 – 58 ms following platform movement) and the early neck muscle attenuations observed in our study (50 – 100 ms following platform movement), it is unlikely that a cognitive on-line response modulated the neck responses [6]. Cervicocollic and vestibulocollic reflexes or a fast visual feedback loop have all been suggested as a possible trigger of the postural response of seated humans following imposed trunk linear acceleration [1,9,10]. Gating or weighting of one of these primary afferents could lead to a modification of the response observed when uncertainty about the initiation of the linear trunk acceleration is removed. A feedforward strategy that possibly changed the gain of the vestibular nuclei and/or cervical spinal interneurons using the timing information from the motor command sent to the thumb to initiate the perturbation could have been adopted. Fine adjustments of EMG onset were observed only for the neck flexor muscles. These muscles prevented the head from moving into extension. It clearly seems that the feedforward adjustments of the postural set were organized to generate functionally efficient neck motor commands. Peak angular positions and velocities (for head, head-C4 and C4-C7 angles) remained constant for both conditions while all neck muscle activities decreased for the predictive condition. Hence, subjects were able to maintain the head – neck within similar safety boundaries while minimizing the neck EMG activity required to stabilize the head – neck system. In the present experiment, the acceleration was well tolerated and not perceived as noxious by the subjects. A greater perturbation could have yielded different responses, i.e. increasing neck co-contractions in the predictive condition to minimize head movements. For the head and head-C4 angles, time to peak angular positions increased in the predictive condition while time to peak angular velocities remained constant. These results suggest that
head and upper neck acceleration times were similar but that deceleration times increased in the predictive condition. Increasing the deceleration times of peak head and head-C4 angles could result from the smaller EMG activity in order to minimize compressive and shear forces acting on the cervical spine. For the C4-C7 angle, different results were observed: the time to peak angular velocity decreased in the predictive condition suggesting that a stiffening of the lower cervical spine on the trunk occurred. This result, however, has to be analyzed cautiously since the difference between the two conditions was only 4 ms. Altogether, results of the present experiment suggest that the uncertainty about the timing of a perturbation yields higher EMG activities but similar peak head –neck angular positions and velocities. This suggests a feedforward control to scale the appropriate motor output when subjects selfinitiated the perturbation. For non-noxious linear accelerations, it appears that the optimal motor output is shaped in order to diminish the compressive and shear forces acting on the cervical spine and not to keep the head from moving.
Acknowledgements This work was supported by NSERC and CIHR-FCQ.
References [1] J.S. Blouin, M. Descarreaux, A. Be´ langer, J. Marcotte, C. Lamarche, N. Teasdale, Neuromuscular mechanisms underlying head-trunk stabilization: implication for whiplash injuries, J. Neuromusculoskel. Syst. 10 (2002) 125–132. [2] J.R. Brault, G.P. Siegmund, J.B. Wheeler, Cervical muscle response during whiplash: evidence of a lengthening muscle contraction, Clin. Biomech. 15 (2000) 426–435. [3] J.S. Frank, Spinal motor preparation in humans, Electroenceph. clin. Neurophysiol. 63 (1986) 361 –370. [4] S. Kumar, Y. Narayan, T. Amell, An electromyographic study of lowvelocity rear-end impacts, Spine 27 (2002) 1044– 1055. [5] S. Kumar, Y. Narayan, T. Amell, Role of awareness in head-neck acceleration in low velocity rear-end impacts, Accid. Anal. Prev. 32 (2000) 233 –241. [6] L. Mazzini, M. Schieppati, Activation of the neck muscles from the ipsi- or contralateral hemisphere during voluntary head movements in humans. A reaction-time study, Electroenceph. clin. Neurophysiol. 85 (1992) 183 –189. [7] T. Pozzo, A. Berthoz, L. Lefort, Head kinematic during various motor tasks in humans, Prog. Brain Res. 80 (1989) 377 –383. [8] T. Pozzo, A. Berthoz, L. Lefort, Head stabilization during various locomotor tasks in humans. I. Normal subjects, Exp. Brain Res. 82 (1990) 97– 106. [9] G.P. Siegmund, D.J. Sanderson, J.T. Inglis, The effect of perturbation acceleration and advance warning on the neck postural responses of seated subjects, Exp. Brain Res. 144 (2002) 314 –321. [10] N. Vibert, H.G. MacDougall, C. de Waele, D.P. Gilchrist, A.M. Burgess, A. Sidis, A. Migliaccio, I.S. Curthoys, P.P. Vidal, Variability in the control of head movements in seated humans: a link with whiplash injuries? J. Physiol. 532 (2001) 851 –868.
Self-initiating a seated perturbation modifies the neck postural responses in humans Jean-Se´bastien Blouin, Martin Descarreaux, Ariane Be´langer-Gravel, Martin Simoneau, Normand Teasdale* Universite´ Laval, Division de Kine´siologie, Faculte´ de Me´decine, Que´bec, Canada Received 28 October 2002; received in revised form 15 February 2003; accepted 13 April 2003
Abstract When seated subjects are submitted to a linear acceleration, reports indicate that the kinematic and electromyographic (EMG) responses of the head – neck system can be modulated with the magnitude of the linear acceleration. There is no evidence, however, that head kinematics or neck EMG activity can be modulated when specific knowledge and active control about the onset of platform acceleration are available. Sixteen seated subjects were given forward linear accelerations in two different conditions nested within subjects: reactive and predictive. In the reactive condition, the acceleration was initiated following a variable delay unknown to the subjects whereas in the predictive condition, subjects manually self-initiated the perturbation. All neck muscle activities were decreased 50 –100 ms after platform movement onset in the predictive condition relative to the reactive condition, whereas head and neck peak angular positions and velocities were not different between the two conditions. These results suggest that feedforward control could use the self-generated timing information of platform movement onset to scale the appropriate neck motor output. q 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Seated perturbation; Timing; Cognition; Neck electromyographic activity attenuation
Head control and stability rely upon the integration of proprioceptive, visual and vestibular sensory inputs [7,8]. When subjects in a seated position are submitted to linear accelerations (simulating rear-end car accidents), the kinematic and EMG responses of the head – neck system are modulated with the magnitude of the acceleration. Generally, with an increased magnitude of the perturbation, shorter-latency EMG responses and greater magnitude of kinematic and EMG responses are observed [2,4,9]. The short latency of these responses (about 70 ms) suggests that they are triggered mainly by postural reflexes [4,9]. The possible role of cognition (expectation) for modulating the head –neck responses remains unclear. First, Kumar et al. [5] showed that when subjects had prior knowledge of the amplitude of the upcoming perturbation, peak head acceleration decreased and time to peak head acceleration increased. In a later experiment, Kumar et al. [4] showed that the amplitude of EMG responses also was attenuated when subjects had prior knowledge of the amplitude of the * Corresponding author. Universite´ Laval, PEPS Local 00232, Division de Kine´siologie, Que´bec, Que´bec, Canada G1K 7P4. Tel.: þ 1-418-6562147; fax: þ 1-418-656-2441. E-mail address: [email protected] (N. Teasdale).
acceleration. Conversely, Siegmund et al. [9] showed that neither the neck muscle responses nor the head kinematic responses were modified by prior knowledge of the amplitude of platform acceleration. These divergent results are difficult to reconcile and leave open the questionable role of cognition on the modulation of seated postural responses. The above experiments [4,9] examined the effects of knowing the magnitude of the upcoming perturbation on the head –neck responses. The temporal onset of the perturbation, however, was unknown to the subjects. Subjects can benefit from specific temporal information for modulating spinal reflexes [3]. There is no evidence that the neck neuromuscular responses can be modulated when subjects self-initiate a forward linear acceleration. The aim of this study was to determine if the neuromuscular responses of the head –neck system can be modified when specific knowledge and active control about the onset of platform acceleration are available to the subjects. We hypothesized that when the onset of the linear acceleration is self-initiated, the nervous system could use this timing information to minimize muscle activation in order to reduce muscle stress at the cervical level. Sixteen healthy subjects, six women and ten men (on
0304-3940/03/$ - see front matter q 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0304-3940(03)00632-3
2
J.-S. Blouin et al. / Neuroscience Letters 347 (2003) 1–4
average ^ SD, 23 ^ 3 years old), without any history of whiplash or neck pain participated in this experiment. All subjects signed a written consent form approved by the local ethics committee. Subjects were submitted to 16 forward whole body linear accelerations in two different conditions: reactive and predictive. For the reactive condition, after a ready verbal signal, the platform moved forward (simulating a small rearend car accident) following a variable delay of 0.5 –5 s unknown to the subjects. For the predictive condition, subjects self-initiated the perturbation by pushing a button with their thumb. Eight trials were performed for each condition. Both conditions were performed on the same day and the time interval between trials was about 1 min for all subjects. The conditions were performed in block (reactive, predictive) and the order of presentation was randomized across subjects. Subjects were comfortably seated in a normal car seat without a seat belt or head rest (see Ref. [1] for a general description of the apparatus). The car seat was mounted on a metal plate fixed on three linear bearings. A steel cable attached to an electromagnet kept the platform from moving. Another steel cable attached at one end of the platform ran through a frictionless pulley and held up a heavy weight. By turning off the electromagnet, the platform accelerated forward with a total displacement of 80 mm at a maximum linear acceleration of 1.1 ^ 0.1 g. An electrical contact at the rear-end of the platform detected the onset of the forward displacement. Total recording lasted 2 s and started with the deactivation of the electromagnet. Platform, neck and head displacements were recorded using a SELSPOT II System (Innovision Systems) with four cameras at a sampling rate of 500 Hz. Infrared diodes were placed on the right side of the subjects: tragus, outer canthus, transverse process of C4 (neck mid-line at mandibular angle level) and spinous process of C7. In addition, one diode was placed on the seat. Position data were filtered using a cubic spline function minimizing residuals. A uni-axial accelerometer (Entran Devices Inc.) was placed on top of each subject’s head using a custom-made head device (total weight: 123 g); it served to date as accurately as possible the onset of the head movement. Linear or rotational movements of the head were calculated using signals from the active markers. Peak angular position and velocity were measured for three head or neck angles: (1) tragus and outer canthus with respect to the horizontal (head angle); (2) tragus and C4 with respect to the vertical (head-C4 angle); and (3) C4 and C7 with respect to the vertical (C4-C7 angle). Time to peak angular position and velocity were also calculated. Bipolar disposable surface Ag-AgCl electrodes (3 mm, Meditrace) were applied bilaterally over the surface of the sternocleidomastoideus (SCM), anterior scalenus (hereafter called scalenus), paraspinals (at the level of C4) and trapezius muscles. The signals were pre-amplified at the source (gain ¼ 500) prior to a second level amplification
(Bortec AMT-8; band-pass of 10 Hz to 1 kHz). Signals from the accelerometer, electrical contact and EMG were sampled at 2 kHz (12-bit A/D resolution). The onset of the platform movement (detected from the electrical contact) was defined as time zero. To examine whether the EMG activities were modulated both prior to and following the perturbation, EMG activities were integrated for nine consecutive 25 ms intervals: from 2 25 ms prior to 200 ms after the platform onset. Integrated EMG (iEMG) values were expressed as a percentage of maximal EMG activity monitored for the first trial in the reactive condition (12.5 ms before and after peak activity of the initial burst). EMG signals were full-wave rectified to determine the onset of the muscular responses. All onsets were determined visually by inspecting the EMG signals. An onset was determined when the activity clearly exceeded that of the baseline variability level for more than ten consecutive samples (5 ms). A similar procedure was used to document the onset of head movement using the accelerometer signal. EMG latency values were measured with respect to platform movement onset. For all dependent variables, within-subject mean of the reactive trials and within-subject mean of the predictive trials were first obtained. Kinematic variables were submitted to a one-way repeated measures ANOVA (Condition – two levels: predictive and reactive). For the EMG latencies, a Side factor (left – right) was added (Condition £ Side). The iEMG data were submitted to a three-way repeated measures ANOVA (Condition £ Side £ Interval). We initially tested for a gender effect and an order effect of conditions. None of these effects were significant (P . 0:05) and they are not presented for the sake of brevity. Prior to each perturbation, the head angular position did not vary across conditions (on average ^ SD, 26.6 ^ 7.78, P . 0:05). The onset of head movement relative to the platform movement also did not vary (on average, 21 ^ 7 ms, P . 0:05). For all three angles, peak angular position and velocity did not vary with onset awareness (P . 0:05 for all analyses). On average, peak angular position and velocity were 9.6 ^ 3.08 and 122.7 ^ 29.78/s for the head angle, 11.9 ^ 3.58 and 160.5 ^ 31.68/s for the head-C4 angle, and 6.5 ^ 2.38 and 81.5 ^ 19.78/s for the C4-C7 angle. Timing values for the head and head-C4 angles, however, showed longer time to peak angular positions for the predictive than the reactive condition (201 vs. 189 ms, F ¼ 6:50, P , 0:05 for the head angle; 193 vs. 175 ms, F ¼ 15:96, P , 0:001 for the head-C4 angle). The time to peak angular velocities did not vary for these angles (on average, 114 ^ 9 ms, P . 0:05 for the head angle; 105 ^ 6 ms, P . 0:05 for the head-C4 angle). Conversely, for the C4-C7 angle the time to peak angular position did not vary between conditions (on average, 162 ^ 41 ms, P . 0:05) but the time to peak angular velocity decreased in the predictive compared to the reactive condition (60 vs. 64 ms, F ¼ 7:20, P , 0:05). Overall, the kinematic responses suggest that the subjects
J.-S. Blouin et al. / Neuroscience Letters 347 (2003) 1–4
were able to maintain the head –neck within similar safety boundaries but modified the timing of the head – neck kinematic responses across conditions. Fig. 1 illustrates the iEMG activities for all muscles across the nine intervals for both conditions. The ANOVAs revealed a significant interaction of Interval £ Condition for all muscles: scalenus (F ¼ 32:17, P , 0:001), SCM (F ¼ 29:80, P , 0:001), paraspinals (F ¼ 9:27, P , 0:001) and trapezius (F ¼ 6:81, P , 0:01). Main effects of Condition and Interval also were significant for all muscles (P , 0:01). Decomposition of the interactions showed that, for all muscles, normalized iEMG activity for the 50 – 75 ms and 75 –100 ms intervals was smaller for the predictive than for the reactive condition (significant Tukey comparison of means for all muscles, P , 0:05). For the paraspinals, a smaller iEMG activity for the predictive condition also was observed for the 100 – 125 ms and 125 –150 ms intervals (significant Tukey comparison of means, P , 0:05). For all muscles, the ANOVAs showed no effect of Side nor any interaction of Side £ Condition, Side £ Interval or Side £ Condition £ Interval (P . 0:05).
3
Hence, this suggests that 50 –100 ms after the perturbation, the activity for all muscles recorded was significantly smaller for the predictive than for the reactive condition. Muscle onsets were computed only for the scalenus, SCM and paraspinals muscles because the EMG activity of the trapezius was difficult to discern from the background activity in the predictive condition. All muscle onsets are reported relative to the platform movement onset since the onset of head movement did not vary for both conditions (see above). For the scalenus and SCM muscles, the latencies for the predictive condition were decreased compared to the reactive condition (45 vs. 52 ms (F ¼ 53:54, P , 0:001) for the scalenus; 46 vs. 54 ms (F ¼ 37:82, P , 0:001) for the SCM). The ANOVA revealed no main effect of Condition for the paraspinals muscles (on average, 58 ^ 8 ms, P . 0:05). No main effect of Side nor any interaction Side £ Condition (P . 0:05) were observed for all muscles. A primary goal of this study was to determine if the nervous system could use the timing information from the onset of platform movement to reduce neck muscular
Fig. 1. Mean normalized integrated EMG activity (for all muscles) between 225 ms prior to platform onset and 200 ms after platform onset are illustrated for the predictive and reactive conditions. Normalized iEMG activities are expressed as a percentage of maximal integrated EMG interval (12.5 ms before and after peak initial burst of activation) calculated for the first reactive trial (see text). Error bars represent 0.95 confidence intervals.
4
J.-S. Blouin et al. / Neuroscience Letters 347 (2003) 1–4
responses. For all muscles, the neck EMG responses were attenuated from 50 to 100 ms after the onset of platform movement for the predictive condition while the preperturbation EMG activity (2 25 to 25 ms after platform movement onset) remained unchanged. This can not be explained by a forward tilt of the head in anticipation of the perturbation since the initial head angular position prior to each perturbation did not vary across conditions. Furthermore, when subjects self-initiated the perturbation, the onset of EMG responses occurred 7 and 8 ms earlier for the scalenus and SCM, respectively. This suggests that the nervous system could use the timing information generated from the thumb motor command initiating the perturbation to scale the appropriate motor output. It supports previous findings by Kumar et al. [4] and suggests that decreasing the uncertainty about the upcoming perturbation could help the nervous system to use the advance information about this upcoming event in order to modify the postural set. It remains to be determined, however, if subjects could still scale their postural responses if an external paced signal was used to inform them about the acceleration onset. Considering the early latency of the responses (45 – 58 ms following platform movement) and the early neck muscle attenuations observed in our study (50 – 100 ms following platform movement), it is unlikely that a cognitive on-line response modulated the neck responses [6]. Cervicocollic and vestibulocollic reflexes or a fast visual feedback loop have all been suggested as a possible trigger of the postural response of seated humans following imposed trunk linear acceleration [1,9,10]. Gating or weighting of one of these primary afferents could lead to a modification of the response observed when uncertainty about the initiation of the linear trunk acceleration is removed. A feedforward strategy that possibly changed the gain of the vestibular nuclei and/or cervical spinal interneurons using the timing information from the motor command sent to the thumb to initiate the perturbation could have been adopted. Fine adjustments of EMG onset were observed only for the neck flexor muscles. These muscles prevented the head from moving into extension. It clearly seems that the feedforward adjustments of the postural set were organized to generate functionally efficient neck motor commands. Peak angular positions and velocities (for head, head-C4 and C4-C7 angles) remained constant for both conditions while all neck muscle activities decreased for the predictive condition. Hence, subjects were able to maintain the head – neck within similar safety boundaries while minimizing the neck EMG activity required to stabilize the head – neck system. In the present experiment, the acceleration was well tolerated and not perceived as noxious by the subjects. A greater perturbation could have yielded different responses, i.e. increasing neck co-contractions in the predictive condition to minimize head movements. For the head and head-C4 angles, time to peak angular positions increased in the predictive condition while time to peak angular velocities remained constant. These results suggest that
head and upper neck acceleration times were similar but that deceleration times increased in the predictive condition. Increasing the deceleration times of peak head and head-C4 angles could result from the smaller EMG activity in order to minimize compressive and shear forces acting on the cervical spine. For the C4-C7 angle, different results were observed: the time to peak angular velocity decreased in the predictive condition suggesting that a stiffening of the lower cervical spine on the trunk occurred. This result, however, has to be analyzed cautiously since the difference between the two conditions was only 4 ms. Altogether, results of the present experiment suggest that the uncertainty about the timing of a perturbation yields higher EMG activities but similar peak head –neck angular positions and velocities. This suggests a feedforward control to scale the appropriate motor output when subjects selfinitiated the perturbation. For non-noxious linear accelerations, it appears that the optimal motor output is shaped in order to diminish the compressive and shear forces acting on the cervical spine and not to keep the head from moving.
Acknowledgements This work was supported by NSERC and CIHR-FCQ.
References [1] J.S. Blouin, M. Descarreaux, A. Be´ langer, J. Marcotte, C. Lamarche, N. Teasdale, Neuromuscular mechanisms underlying head-trunk stabilization: implication for whiplash injuries, J. Neuromusculoskel. Syst. 10 (2002) 125–132. [2] J.R. Brault, G.P. Siegmund, J.B. Wheeler, Cervical muscle response during whiplash: evidence of a lengthening muscle contraction, Clin. Biomech. 15 (2000) 426–435. [3] J.S. Frank, Spinal motor preparation in humans, Electroenceph. clin. Neurophysiol. 63 (1986) 361 –370. [4] S. Kumar, Y. Narayan, T. Amell, An electromyographic study of lowvelocity rear-end impacts, Spine 27 (2002) 1044– 1055. [5] S. Kumar, Y. Narayan, T. Amell, Role of awareness in head-neck acceleration in low velocity rear-end impacts, Accid. Anal. Prev. 32 (2000) 233 –241. [6] L. Mazzini, M. Schieppati, Activation of the neck muscles from the ipsi- or contralateral hemisphere during voluntary head movements in humans. A reaction-time study, Electroenceph. clin. Neurophysiol. 85 (1992) 183 –189. [7] T. Pozzo, A. Berthoz, L. Lefort, Head kinematic during various motor tasks in humans, Prog. Brain Res. 80 (1989) 377 –383. [8] T. Pozzo, A. Berthoz, L. Lefort, Head stabilization during various locomotor tasks in humans. I. Normal subjects, Exp. Brain Res. 82 (1990) 97– 106. [9] G.P. Siegmund, D.J. Sanderson, J.T. Inglis, The effect of perturbation acceleration and advance warning on the neck postural responses of seated subjects, Exp. Brain Res. 144 (2002) 314 –321. [10] N. Vibert, H.G. MacDougall, C. de Waele, D.P. Gilchrist, A.M. Burgess, A. Sidis, A. Migliaccio, I.S. Curthoys, P.P. Vidal, Variability in the control of head movements in seated humans: a link with whiplash injuries? J. Physiol. 532 (2001) 851 –868.