Abstract
Background Chronic whiplash-associated disorders (WAD) have been shown to be associated with motor dysfunction. Increased electromyographic (EMG) activity in neck and shoulder girdle muscles has been demonstrated during different tasks in participants with persistent WAD. Muscle functional magnetic resonance imaging (mfMRI) is an innovative technique to evaluate muscle activity and differential recruitment of deep and superficial muscles following exercise.
Objective The purpose of this study was to compare the recruitment pattern of deep and superficial neck flexors between patients with WAD and controls using mfMRI.
Design A cross-sectional design was used.
Method The study was conducted in a physical and rehabilitation medicine department. The participants were 19 controls who were healthy (10 men, 9 women; mean [±SD] age=22.2±0.6 years) and 16 patients with WAD (5 men, 11 women; mean [±SD] age=32.9±12.7 years). The T2 values were calculated for the longus colli (Lco), longus capitis (Lca), and sternocleidomastoid (SCM) muscles at rest and following cranio-cervical flexion (CCF).
Results In the overall statistical model for T2 shift, there was a significant main effect for muscle (F=3.906, P=.033) but not for group (F=2.855, P=.101). The muscle × group interaction effect was significant (F=3.618, P=.041). Although not significant, there was a strong trend for lesser Lco (P=.061) and Lca (P=.060) activity for the WAD group compared with the control group. Although the SCM showed higher T2 shifts, this difference was not significant (P=.291).
Limitations Although mfMRI is an innovative and useful technique for the evaluation of deep cervical muscles, consideration is required, as this method encompasses a postexercise evaluation and is limited to resistance types of exercises.
Conclusions Muscle functional magnetic resonance imaging demonstrated a difference in muscle recruitment between the Lco, Lca, and SCM during CCF in the control group, but failed to demonstrate a changed activity pattern in the WAD group compared with the control group. The mild symptoms in the WAD group and the wide variability in T2 values may explain the lack of significance.
Chronic whiplash-associated disorders (WAD) have been shown to be associated with motor dysfunction, which occurs soon after injury and persists into the period of chronicity in some patients.1,2 Changes observed include reduced cervical spine movements, disturbances in cervical kinesthesia as reflected by errors in head and neck repositioning, and increased electromyographic (EMG) activity in neck and shoulder girdle muscles during tasks of functional low-load activities, as well as during a task of supported cranio-cervical flexion (CCF).2–11
The task of CCF aims to address the deep cervical muscles and was developed by Jull and coworkers12 in response to research indicating the importance of the deep cervical flexors in support of the cervical lordosis and motion segments and clinical observations of their impairment with neck pain.12–16 More specifically, CCF is the primary action of the longus capitis (Lca) muscle that attaches to the cranium and the superior portion of the longus colli (Lco) that attaches to the first cervical vertebra.12,17,18 In contrast, superficial cervical flexor muscles such as the sternocleidomastoid (SCM) muscle are not prime movers of CCF and are structurally more suited to assist in flexing the lower cervical spine on the thorax.
Previous studies have demonstrated higher EMG signal amplitudes in the SCM in both patients with WAD and patients with insidious-onset neck pain compared with people who are healthy.2,8,10,16 Inferences were made that the increased activation of the superficial cervical flexors was likely to be a compensation for reduced deep cervical flexor activation, but there was no direct evidence for this assumption. This evidence was provided in a subsequent study that directly measured activity in the deep cervical flexors with a novel surface EMG technique.16,19 It was demonstrated that individuals with chronic neck pain exhibited reduced activity of the deep cervical flexors during this test, which was, in combination with an increased activity in the SCM, an indication of an altered neuromotor control strategy.16
Although this novel EMG technique allows direct measurement of deep cervical flexor activity, it has some limitations. This technique is quite invasive, as electrodes are inserted via the nose with a nasopharygeal suction catheter and suctioned onto the posterior oropharyngeal wall. This method also does not differentiate between the Lco and Lca, and muscle cross-talk cannot be ruled out.20,21 In contrast, muscle functional magnetic resonance imaging (mfMRI) has gained popularity in studies evaluating muscle function to overcome the problems with surface EMG. It has been demonstrated to be a noninvasive method to observe the differential activation of deep and superficial cervical flexor muscle layers and to allow differentiation between the Lco and Lca, as both muscles may have a different anatomical action based on their attachments.22–25 The technique is based on an acute enhancement of the T2 relaxation time (T2) of muscle water due to activity.26 Increased T2 of muscle water is reflected in a significant augmentation of the signal intensity of the activated muscles and provides information regarding the pattern and the intensity of the muscle activation. Previous studies in which the sensitivity of the T2 value for changes in exercise intensity was investigated provided evidence of a linear association between the T2 value and exercise intensity.27,28
Although there is enough evidence of an association between chronic neck pain and impaired cervical flexor muscle performance, mfMRI may provide additional information regarding the recruitment of the cervical flexor muscles in patients with WAD. The purpose of the present study was to compare the recruitment pattern of Lco, Lca, and SCM during CCF between patients with WAD and people who are healthy by use of mfMRI.
Materials and Method
Setting and Participants
Nineteen participants who were healthy (10 men and 9 women), with a mean (±SD) age of 22.2±0.6 years, comprised the control group. Exclusion criteria were recent neck pain, back pain, or headache of cervical origin (<3 months) and contraindications to magnetic resonance imaging (MRI), such as a cardiac pacemaker, claustrophobia, implanted metals, unremovable piercings, aneurysm clips, carotid artery vascular clamp, neurostimulator, cochlear or ear implants, and (possible) pregnancy within the first 3 months.
Patients in the WAD group were 5 men and 11 women with a mean (±SD) age of 32.9±12.7 years. Patients were selected via the Flemish Whiplash Association, and criteria for inclusion in the study consisted of a period of at least 6 months since the accident (mean=6.4 years, SD=4.7), ongoing pain and disability in the neck region, severity of injury classified as Québec Task Force grade 2 or 3,29 and no other injury relating to a previous motor vehicle accident. Patients were excluded if they had either undergone cervical spine surgery or reported any neurologic signs. On the test day prior to MRI, patients were asked to complete the Neck Disability Index (NDI)30 and to indicate their average intensity of neck pain on a 10-cm numerical rating scale (NRS) anchored with “no pain” and “the worst possible pain imaginable.” The NDI (score out of 50) was used to measure the perceived impairments of the patients resulting from their neck pain. The mean (±SD) for average intensity of pain (NRS) and perceived disability (NDI) were 3.9±2.6 and 17.2±7.4, respectively. Written informed consent was obtained from all participants.
General Design
The protocol was designed to compare the recruitment pattern of deep and superficial neck flexors between patients with WAD and controls by use of mfMRI. Therefore, magnetic resonance images were obtained under 2 conditions. They first were obtained at rest. Participants then performed the CCF exercise outside the scanner room. Immediately after the CCF exercise, a second magnetic resonance image was obtained.
Exercise Protocol
At least 1 week prior to the test day, participants were instructed in the action of CCF. In a supine position, participants first were carefully instructed in this action by performing a gentle nodding action without lifting the head off the surface until full CCF range of motion was reached. Analysis of the performance was based on the clinical protocol, as described by Jull et al.12 Once the correct movement pattern was achieved, a pressure cuff (Stabilizer)* was placed suboccipitally behind the participant's cervical spine and inflated until a stable pressure of 20 mm Hg was achieved in order to monitor the slight flattening of the cervical lordosis and register the muscular effort as an increase in pressure. Participants practiced progressive targeting of 5 incremental levels (2-mm-Hg increments) between 22 and 30 mm Hg by steadily maintaining the pressure on each target for 10 seconds, guided by visual feedback of the pressure cuff and controlled by an examiner. As 26 mm Hg was the pressure level that was used on the test day, 10-second holds of this target were repeated 10 times.
On the test day, participants were asked to maintain the pressure level at 26 mm Hg and to execute an isometric hold until fatigue prevented them from sustaining the contraction any longer. The exercise was performed 3 times, with 1 minute of rest between the sets. Participants were instructed to discontinue the test when they perceived an inability to sustain the contraction at the indicated intensity. The mean (±SD) time CCF was held (CCF time) was 107.7±46.7 seconds in the control group compared with 58.7±31.4 seconds in the WAD group.
MRI
Magnetic resonance imaging was performed using a 3-T magnet (MAGNETOM Trio, Tim system, syngo MR VB13 software).† A flexible surface coil, 20 × 50 cm, fixed over the anterior aspect of the participants' neck was combined with the phased-array spine coil as a receiver coil combination.
The participants were placed in a comfortable and relaxed supine position, with their hips flexed to 45 degrees and their legs supported by foam wedges. Their head was positioned in neutral, without rotation, lateral flexion, or exaggerated lordosis. Axial images of the cervical spine were obtained at rest and after the exercise. As the half-life of exercise-induced changes in muscle T2 has been shown to be 7 minutes, participants were placed in the scanner immediately following the exercise.28 The mean (±SD) time between the end of the exercise and the start of scanning was 117±13 seconds.
A sagittal localizing sequence first was performed to identify cervical disk space intervals. Axial images parallel to the consecutive intervertebral disks with a slice thickness of 5 mm were obtained at 5 different cervical levels (C0–C1, C2–C3, C3–C4, C4–C5, and C5–C6). For T2 calculation, a multi-spinecho sequence was used: repetition time=2,500 milliseconds; echo times=10 to 161.6 milliseconds, with steps of 10.1 milliseconds (16 echoes); field of view=256 mm; matrix=128 × 128; and voxel size=2 × 2 × 5 mm. Total acquisition time for 1 scan was 5:12 minutes. Imaging procedures were identical for the resting scan and the exercise scan.
Data Analysis
After scanning, the images were transferred to a computer for calculation of muscle T2 using ImageJ, a Java-based version of the public domain NIH Image software.§ In order to calculate T2 values, regions of interest were identified on the T2 images. On the base of its clearest visualization, the Lca was analyzed at the C0-C1 level, whereas the SCM and Lco were analyzed at the C5–C6 level (Fig. 1).
Axial T2-weighted scan at (a) the C0–C1 level, showing the longus capitis muscle (Lca) surrounded by the oropharynx (OP), the anterior arch of the atlas (AAA), and the spinal cord (SC), and (b) the C5–C6 level, showing the sternocleidomastoid muscle (SCM) and the longus colli muscle (Lco) surrounded by the vertebral body of C5, the common carotid artery (CCA), and the larynx (L). A=anterior, P=posterior, TR=repetition time, and TE=echo time
A region of interest was defined as the greatest outline of the muscle with avoidance of nonmuscular tissue, such as fat, fascia, and vessels. Sixteen echoes were used in T2 calculation, using a Simplex algorithm to fit the values from the specific slice in a T2 image volume to the exponential Sn=S0 exp(−TEn/T2) (n=1:16), where Sn is a signal intensity of n-th echo, S0 is the initial signal intensity, and TEn is the echo interval. The mean T2 value and its standard deviation were derived for each region of interest. The T2 measurements for the Lca, Lco, and SCM have been shown to be highly reliable in our laboratory, with intraclass correlation coefficients ranging from .87 to .94, depending on the muscles evaluated.23
Analysis of the data was performed using SPSS statistical software, version 16.‡ Descriptive statistics (mean±SD) were calculated for T2 values (millisecond) at rest and after exercise for each muscle group. The T2 shifts, which are defined as T2 values after exercise minus T2 values at rest, were used for statistical analysis.
The T2 shifts were evaluated using a repeated-measures analysis of variance with the within-participant factor being muscle (Lca, Lco, and SCM) and the between-participant factor being group (WAD and control). Participant's sex was used as covariate in this analysis. Post hoc comparisons were made when required, and adjustments were used to correct for multiple tests (Holm-Bonferroni method). Pearson correlation coefficients were used to evaluate correlations between T2 shifts and the following parameters: NDI, NRS, CCF time, and time since injury. Statistical significance was accepted at the .05 alpha level.
Results
The T2 values at rest and after exercise calculated for all muscles in the control and WAD groups are shown in the Table. There were no significant differences in the T2 values at rest for all muscles between the control and WAD groups.
Mean (±SD) T2 Values (in Milliseconds) for the Longus Capitis (Lca), Longus Colli (Lco), and Sternocleidomastoid (SCM) Muscles at Rest and Following Cranio-Cervical Flexion (CCF), With Mean Differences and 95% Confidence Intervals (CI) and P Values for Comparisons of T2 Shifts Between Control Participants and Patients With Chronic Whiplash-Associated Disorder (WAD)
Mean T2 shifts plotted by muscle and group are shown in Figure 2. In the overall statistical model for T2 shift, there was a significant main effect for muscle (F=3.906, P=.033) but not for group (F=2.855, P=.101). The muscle × group interaction effect was significant (F=3.618, P=.041). There was no effect of participant's sex (P=.404).
T2 shifts in milliseconds (mean±SD) of the longus capitis (Lca), longus colli (Lco), and sternocleidomastoid (SCM) muscles for the group with chronic whiplash-associated disorders (WAD) and the control group. *P<.05.
In the control group, there was a significantly higher T2 shift in the Lca compared with the SCM (P=.010), whereas no differences among muscles were found in the WAD group (Fig. 2). The patients with WAD demonstrated a trend for lower T2 shifts in both the Lco (P=.061) and the Lca (P=.060) compared with the control group. Although the SCM showed higher T2 shifts, this difference was not significant (P=.291).
There were no significant correlations between the observed T2 shifts and NDI, NRS, and time since injury (r=−.319 to .562). Although there was a significant difference in CCF time between the control and WAD groups (P≤.001), no correlation was found between CCF time and observed T2 shifts in all muscles (r=−.244 to .132).
Discussion
The results of this study indicate that mfMRI demonstrated a difference in muscle recruitment between the Lca and SCM during CCF in the control group, but failed to demonstrate a changed activity pattern in the WAD group compared with the control group. In the control group, the Lca displayed a significant higher T2 increase than the Lco and SCM. This finding is in accordance with previous studies that demonstrated that the CCF method is more specific to the anatomical action of the deep cervical flexor muscles and less specific to the anatomical action of the superficial cervical flexors.20,21 However, in these studies, the Lca and Lco were taken together as the deep cervical flexors, and no distinction was made between the muscles.14,19,25 As the results demonstrate that the Lca is more activated than the Lco during the CCF, this study highlights that the action of these muscles can be distinguished using mfMRI.
When comparing the WAD and control groups, the results lacked significance, although the patients with WAD demonstrated a trend for lower T2 shifts in both the Lco and Lca. To the best of our knowledge, Falla et al16 are the only researchers who also investigated the deep cervical flexors during CCF between patients with chronic neck pain and asymptomatic controls by use of surface EMG. They demonstrated a trend for lower deep cervical flexor activity at all stages of the CCF in participants with neck pain, but this difference was statistically significant only at the higher stages of the test (28 and 30 mm Hg). As the pressure level in the present study was fixed at 26 mm Hg, it is plausible and in accordance with the data of Falla et al16 that only a trend for lower muscle activation was found. It could be questioned why 26 mm Hg was used in this study, as Falla et al16 previously had demonstrated significant differences only at the 2 highest levels of the test. The reason for choosing this level was based on the results of Jull et al,10 who demonstrated large shortfalls in pressure in the latter 2 stages of the test, indicating that many of the participants would fail to execute these stages of the test.
Although the SCM showed higher T2 shifts in the WAD group, this difference was not significant compared with the control group (P=.291). Different studies have demonstrated higher measures of EMG signal amplitude in the SCM in patients with WAD, as well as in patients with cervicogenic headache and idiopathic neck pain and office workers with neck pain, compared with control participants who were healthy.2,8,10,16,31–33 Except for the study of Falla et al,16 who only found a trend for greater SCM activity for participants with mild neck pain (NDI=12.4±9.5), a common finding in these studies was a significant increase in EMG amplitude of the SCM for the final 2 stages of the test (28 and 30 mm Hg).2,8,10,31,33 The results for the other stages of the test differed among the previous studies. Jull et al33 found significantly greater SCM activity in individuals with cervicogenic headache for the final 3 stages of the test (P<.001), and Johnston et al32 also found a significant difference at the 24-mm Hg level. Sterling et al31 only found higher levels of EMG activity with CCF in individuals with WAD and persistent moderate to severe symptoms (NDI=37.1±8), whereas there was no difference between individuals with mild pain and disability (NDI=16.7±6) and a control group.
Based on these results, we hypothesize that the impairment in CCF is not dependent on the etiology of neck pain, but rather on the severity of the neck pain. As the mean (±SD) NDI score of the patients with WAD in this study was 17.2±7.4, which is quite similar to the patient characteristics in the studies by Falla et al16 and Sterling et al,31 it could be argued that this reflects a mild pain and disability group, resulting in no significant differences between groups. Further research in patients with WAD with moderate to severe symptoms is needed to evaluate whether mfMRI can distinguish between patients with neck pain and controls.
There are different hypotheses as to why there is a differential muscle recruitment in patients with WAD. The current evidence suggests that an altered control strategy may be induced by pain, which, in turn may contribute to muscle overload or disuse.34 These changes in motor control may lead to additional adaptations at the muscle level. Elliott et al35 recently demonstrated a significantly greater muscle fat infiltration and cross-sectional area in the anterior neck muscles, especially in the deeper Lca and Lco, in participants with chronic WAD compared with a control group. It is assumed that larger intracellular fat stores would increase the T2 values, because fat has a high T2 value relative to muscle.36 However, the current study could not demonstrate a significant difference in T2 values at rest between the control and WAD groups.
The present results must be viewed within the limitations of the study. The male/female ratio was not similar in both groups, which may have affected the results. However, there was no effect of participant's sex (P=.404), which is in accordance with previous studies.10,31
As mfMRI is very sensitive to the intensity of the exercise, a fixed pressure level (26 mm Hg) with an isometric hold until exhaustion was used for all participants.28,37 This is not a real reflection of clinical practice, in which the baseline assessment is documented as the pressure level that the patient can achieve and hold for 2 to 3 seconds with the correct CCF action, with minimal superficial muscle activity and in the absence of any other substitution strategy.12 This pressure level differs between patients and people who are healthy; individuals who are asymptomatic normally attain 26 to 28 mm Hg, and the baseline assessment in patients with neck pain is only 22 to 24 mm Hg.
Higher T2 values were found in this study compared with previous studies.25,38–40 The most acceptable reason for this discrepancy is the fact that different imaging sequences were used, which may have given a systematic deviation of 10% to 20%. The current T2 values, however, are comparable to the values found by Dickx and colleagues,24,41 who evaluated the lumbar muscles with the same apparatus and the same imaging sequences.
Besides the discussion of methodological limitations, a perspective on the state of the art of mfMRI is warranted. It is important to recognize that mfMRI may not be seen as a superior method compared with EMG but rather as a complementary evaluation technique. The underlying mechanisms and physiological representation differ between the 2 methods. Electromyography monitors electrical activity of activated muscles, whereas the mfMRI method is based on registering changes in metabolic activity and water content.
Both MRI and EMG measurement techniques have intrinsic advantages and disadvantages. Electromyography has the advantage of evaluating real time, which permits the investigation of timing of muscles and changes over time. However, due to the problem of cross-talk, it often is difficult to obtain an EMG signal representing isolated activity of the target muscle.26 In addition, deep muscles are not accessible with surface electrodes.42 Another technical problem is the variability in myoelectrical signal attributed to subcutaneous tissue and electrode type and placement. The mfMRI technique has the advantage, especially in the case of the spinal muscles, of easily evaluating deep muscles, adjacent muscles, and even overlying muscles without the problem of cross-talk.
However, some careful considerations about the mfMRI technique need to be made. First, mfMRI is a postexercise evaluation of muscle recruitment, which enables only evaluation of spatial characteristics, whereas no temporal characteristics such as changes in timing of muscle activity can be assessed. Second, mfMRI is limited to resistance exercises. Although a linear relationship has been demonstrated among exercise intensity, EMG activity, and T2 times, the lowest activity threshold to induce a significant shift in signal intensity is still not known.27,37,43 However, most studies agree that changes in T2 times can be detected with as few as 2 repetitions of a task.25,44,45 The validity of this method has been demonstrated in different muscles, but studies are lacking defining the sensitivity of this method in the assessment of cervical muscles.27,37,43,46,47 As the muscles evaluated in this study are rather small compared with larger muscles of the limbs, which have been investigated in previous studies using mfMRI, the small muscle size may partly explain the wide variability observed in the T2 values and the lack of significant results.
Finally, it is clear that changes in T2 values with exercise are multifactorial, including fiber type distribution, differences in regional perfusion, and aerobic capacity.44 The time course and the persistence of the changes in T2 times recently were modeled by Damon and Gore.48 However, the precise mechanism of how individual physiological and biochemical variables contribute to the changes in T2 remains unclear and warrants further research.
As a consequence of the advantages and disadvantages of both techniques, it is clear that mfMRI and EMG can be used independently to assess muscle activity, depending on the purpose of research. However, combining both techniques provides additional information and adds value to answering research questions.
Conclusions
The results of this study confirm the findings of earlier experiments on cervical muscle recruitment by using a novel technique of mfMRI to evaluate muscle function and dysfunction of the cervical flexors. The mfMRI technique demonstrated differences in muscle recruitment among the Lco, Lca, and SCM during CCF in the control group, but failed to demonstrate a changed activity pattern in the WAD group compared with the control group. The mild symptoms in the WAD group and the wide variability in T2 values may explain the lack of significance.
Footnotes
-
Dr Cagnie, Mr Peeters, Dr Achten, and Dr Danneels provided concept/idea/research design. Dr Cagnie provided writing. Dr Cagnie, Ms Dolphens, and Mr Peeters provided data collection and analysis. Dr Cagnie, Dr Cambier, and Dr Danneels provided project management. Ms Dolphens provided participants. Dr Achten, Dr Cambier, and Dr Danneels provided facilities/equipment. Dr Achten and Dr Cambier provided institutional liaisons. Dr Cagnie, Dr Achten, Dr Cambier, and Dr Danneels provided consultation (including review of manuscript before submission). The authors thank the staff of the Ghent Institute for Functional and Metabolic Imaging for their technical assistance and Nick Vandekeybus, Martijn Schouten, and Robert Dirks for their assistance with data collection.
-
This study was approved by the local ethics committees (Ghent University).
-
This study was supported by the Research Foundation–Flanders (FWO).
-
↵* Chattanooga Group, 4717 Adams Rd, Hixson, TN 37343.
-
↵† Siemens AG, Erlangen, Germany.
-
↵§ Office of Research Services, National Institutes of Health, 31 Center Dr, MSC 2324, Building 31, Room 4B30, Bethesda, MD 20892-2324.
-
↵‡ SPSS Inc, 233 S Wacker Dr, Chicago, IL 60606.
- Received October 27, 2009.
- Accepted April 12, 2010.
- © 2010 American Physical Therapy Association
References
Issue highlights
