Cervical Proprioception Impairment in Neck Pain-Pathophysiology, Clinical Evaluation, and Management: A Narrative Review

Studies have shown that the spindles of the cervical muscles are the major proprioceptors of the neck and not the joint capsules [4, 16]. The spindle density of the cervical muscles is much higher than that of the shoulder and thigh muscles [17, 18]. In general, high spindle density in small muscles is associated with fine motor tasks [19]. The deep cervical muscles in humans have shown high spindle content, particularly in small suboccipital muscles [18‐22]. Complex integrative mechanisms involving head–eye coordination may require complex proprioceptive inputs from the neck muscles, which may account for their high spindle content [18].

Mechanoreceptors, Ruffini corpuscles, Pacinian corpuscles, and Golgi tendon organs have been reported in human cervical facet joints [23, 24] and discs [25, 26]. As with the peripheral joints [27, 28], there are only a small number of mechanoreceptors in the lumbar and cervical facet joints [23, 24]. There was no significant difference in the distribution of receptors between the upper and lower cervical facet capsules [24]. As suggested by McLain, these findings may indicate that the role of these receptors in cervical proprioception is more limited [29]. Similarly, there are only a few proprioceptors in the cervical discs [25, 26]. However, because the cervical disc is located on the axis of cervical motion, they are in a favorable position to detect subtle changes in the direction or position of the cervical spine [25].

The structures of cervical spine present degenerative changes with age, but the spindle characteristics such as the spindle distribution, morphology, and density in the cervical intrinsic muscles (longus colli and multifidus) remain unchanged [19]. So far, there is still a lack of comparative studies on the distribution of mechanoreceptors in normal and osteoarthritic cervical facet joints. Cervical facet joints and peripheral joints (such as hip and knee joints) are synovial joints and are anatomically similar. Moraes et al. [28] noted that the morphology of mechanoreceptors in the hip joint capsule did not seem to differ between healthy people and patients with osteoarthritis. Overall, the density of mechanoreceptors in the joint capsule is low, but in the joint capsule of osteoarthritis patients it is lower. The density of mechanoreceptors in patients with osteoarthritis is 0.044 per mm², and in healthy people is 0.053 per mm² [27, 28]. Recently, a comparative study by Yang et al. [25] found that Ruffini corpuscles were obviously increased in number and deeply ingrown into inner annulus fibrosus and even into nucleus pulposus in the diseased cervical discs from cervical spondylosis patients with dizziness in comparison with the discs from patients without dizziness and control discs. The densities of Ruffini corpuscles in anterior outer annulus fibrosus are 1.02, 0.6, and 0.33 per mm² in patients with dizziness, patients without dizziness, and normal controls, respectively [25]. In addition, the authors [30] also found that a large number of Ruffini corpuscles and substance P-positive free nerve fibers grew in the degenerative cervical discs of patients with chronic neck pain and dizziness compared with the normal control group. These patients only showed degenerative changes of the cervical disc on imaging, without cervical disc herniation or nerve root compression. The studies suggest the degenerated cervical discs play a key role in impaired cervical proprioception.

Proprioceptive information is transmitted to the central nervous system through an encoding across populations of afferent receptors, the ensemble coding, rather than the discrete units from individual receptors [31‐33]. In physiological conditions, cervical proprioceptive information from cervical muscle spindles and mechanoreceptors of cervical discs and facet joints is integrated and transmitted to the central nervous system to control head position, head orientation, and full-body posture [29, 34, 35]. Any dysfunction of these cervical sensory organs or asymmetry of afferent inputs can lead to a sensory mismatch between abnormal information (e.g., from degenerative discs) and normal information (e.g., from normal spindles) [35, 36].

It has been reported that patients with nonspecific neck pain have motor control impairments, but the specific deficits of the underlying regulatory system remain unclear [37]. Experimental muscle pain caused by injection of chemicals such as hypertonic saline provides a way to assess changes in motor control due to changes in afferent feedback. Experimental muscle pain often results in inhibition of the activation of painful muscles. The inhibition of muscle activity induced by pain has been observed to decrease the surface electromyographic (EMG) signal amplitude [38, 39] and motor unit discharge rate [40, 41]. Since the force does not change when muscle activity decreases [38, 42], compensation mechanisms should appear in both painful and non-painful conditions to allow similar motor outputs. Muscle pain affects motor control strategies through central mechanisms [43, 44], including in the painful muscles [38] as well as in synergistic and antagonistic muscles [45]. In repetitive dynamic tasks, the nociceptive inputs from the upper trapezius by injection of hypertonic saline induces the reorganization of the coordinated activities of the three subdivisions of the trapezius muscle [46]. This is consistent with the pain adaptation model [47], and except for the decrease in painful muscle activity, the activity of the ipsilateral and contralateral non-painful muscle subdivisions increases [46].

Cervical spine is a complex biomechanical system consisting of innumerable degrees of freedom motion of each joint and at least 20 pairs of muscles, many of which perform similar functions [48]. The decrease in muscle activity due to neck pain can be compensated, for example, by reducing the activity of antagonist muscles or increasing the activity of synergistic muscles. An experimental study has shown that the local excitation of the noxious afferents in cervical muscles has an inhibitory effect on the painful muscles, which are offset at the level of the muscle group involved in the task by complex motor strategic reorganization. This change in muscle coordination is task-dependent, so motor output remains constant in the state of pain [48].

Acute neck pain induced by hypertonic saline injection at C2/3 level in the splenius capitis muscle on one side caused cervical proprioceptive disturbance with side-specific changes [49]. The clinical implications of this finding are that neck pain itself has a clear role in proprioception and neck sensorimotor control, and subsequently influence orientation. The main function of pain is to prevent further tissue damage. This corresponds to a decrease in muscle activity of the painful muscles, whose activity moves from deeper pained muscles to superficial muscles. It is likely that in neck pain, this transfer may interfere with normal proprioception [49].

A number of studies have indicated that patients with chronic neck pain may be associated with alterations in cervical motor behavior (timing and activation) [50‐52], a decrease in cross-sectional area of cervical muscles [53] as well as muscular functional deficiencies in strength, endurance, precision and acuity, and range of motion [54]. Neck pain may cause maladaptive strategies, change the neck muscle coordination, and reduce the specificity of neck muscle activation, for instance, through reduced activation of the deep segmental muscles and increased activation of the superficial muscles [48, 54]. As mentioned above, muscle spindles densely packed in the deep neck muscles are the main source of proprioception afferents in the neck. These structural and functional changes in the cervical deep and superficial muscles can change the discharge of muscle spindles, which affects the afferent input and leads to alterations in proprioception [13, 55]. The results suggest that the complexity and multifaceted nature of neck muscle impairment exists in people with chronic neck pain. In addition, the effect of pain on many levels of the nervous system can alter the sensitivity of muscle spindles and change the representation and regulation of the cortex to the cervical afferent input [43, 56, 57].

Most chronic idiopathic neck pain is thought to arise from degenerative cervical discs and facet joints [3]. The degenerative changes in cervical discs and facet joints are always related to inflammation [58, 59]. In the inflammatory environment, the discharge characteristics of the mechanoreceptors will be excessively active, thereby producing erroneous sensory signals [25, 30].

The mechanoreceptors in the discs and facet joint capsules can monitor and control activity of paraspinal muscles. Electrical stimulation of mechanoreceptors in lumbar discs and capsules seems to induce reflex contraction of lumbar muscles. Similarly, mechanical stimulation of mechanoreceptors in the lumbar disc and facet capsule can excite the surrounding muscles [60]. Thus, the function of the cervical motor control is closely related to the cervical proprioceptive inputs from the cervical discs and facet joints, which are responsible for the optimal recruitment of cervical muscles. In the process of maintaining body balance, there are both coordination and competition among the balance organs [36, 61]. In pathological conditions, such as cervical disc degeneration or facet osteoarthritis, the erroneous proprioceptive input distorts the direct linear interaction between neck proprioception and vestibular information, resulting in subjective body orientation and spatial psychological representation, which is manifested as dizziness or subjective unstable perception [25, 62].

Neck pain itself may interfere with afferent signals from the proprioceptors of the neck, leading to erroneous proprioceptive information. The system that conducts pain carries activity to motor neurons in the same spinal segment [63]. Therefore, increased pain increases muscle tension. Electrical stimulation of group III muscle afferents and intramuscular injection of hypertonic saline can significantly change the activity of leg muscle γ-motoneurons in animals [44, 64]. The algesic chemical injection of the masseter muscle preferentially affects the amplitude sensitivity of jaw muscle spindle afferents, which is consistent with the effect on static γ-motoneurons [65]. The results suggest that pain-induced modulation of spindle afferent responses is mediated by small-diameter muscle afferents and that this modulation depends in part on the transmission of nociceptive information from the trigeminal subnucleus caudalis onto trigeminal γ-motoneurons [65]. According to the pathophysiological model [66], the metabolites produced by muscle contraction activate the muscle afferent nerves of groups III and IV, which in turn activates γ-motoneurons, supplying intrafusal fibers in both homomorphous and heteromorphic muscles, thereby increasing spindle stretching sensitivity and reflex-mediated muscle stiffness. In addition, the concentration of metabolites such as potassium, lactic acid, and arachidonic acid may directly sensitize different proprioceptors, leading to erroneous proprioceptive signals. Furthermore, the erroneous proprioceptive input can trigger the increased and prolonged reflex activation of neck muscles, which may lead to neck pain over time, thus forming a vicious circle [60].

The vestibular, visual, and cervical proprioceptive systems have unique central and reflex connections and integration [4]. As mentioned before, the cervical sensorimotor control system in patients with idiopathic neck pain has been affected due to impaired proprioception. To date, eight neck sensorimotor control tests have been reported to evaluate patients with idiopathic neck pain: joint position error (JPE) [67], postural sway [67, 68], subjective visual vertical [69], head tilt response [70], The Fly [71], smooth pursuit neck torsion [72, 73], head steadiness [74], and rod-and-frame test (a summary of eight neck sensorimotor control tests is shown in Table 1) [14, 75, 76]. Although some tests may involve different subsystems (such as oculomotor system and vestibular system), all tests measure sensorimotor control in the neck, and the most commonly used is JPE test [77]. However, it is difficult to compare test results due to the different tasks required or the techniques used to quantify the measurements. In addition, it is not clear whether different measurement methods are equally reliable and effective, and which test is preferred [7]. A recent study [78] compared seven tests (all except the rod-and-frame test) in 50 patients with chronic idiopathic neck pain and found that different tests measured different components of sensorimotor control. The study implies that all seven tests are independent of each other. It is not possible to recommend a test battery for clinical practice.

Cervical joint position sense (JPS) is a major component of proprioception, and mainly reflects the ascending input (afferent) of cervical muscle, disc, capsule, and ligament receptors [79]. Abnormal cervical afferent input leads to an impaired cervical JPS, which is measured as cervical JPE [68]. Several methods exist to study cervical JPE, and the most commonly used is active motion angle reconstruction test and requires subjects to relocate a neutral head position or a predefined target head position without visual assistance selected by the researchers (Fig. 3). In the study of head position sense measurement, the measured variable is the difference between the initially determined reference point position (neutral or target position) and the position produced when the subjects try to match the target position. This difference is called JPE and the angle is in degrees (°) [80]. The test has been widely used to distinguish patients with chronic neck pain from the healthy control group [11, 12, 79].

Revel et al. [67] firstly reported that patients with chronic neck pain showed poorer ability (6.11°) to restore the original position of the head after actively rotating the head to the maximum extent compared with healthy subjects (3.50°). They suggested changing the proprioception of the neck to account for their findings. Alahmari et al. [81] compared the JPE of 42 patients with chronic neck pain and 42 age-matched healthy people, and found that the patients had a greater error in all movement directions tested (p < 0.001). Similarly, other studies have revealed a significant greater error in patients with chronic idiopathic neck pain when compared with asymptomatic controls, despite the variability in methods used in these studies [68, 82‐87]. Cervical proprioceptive errors in the patients with cervical spondylosis were also shown to be significantly larger than those of the healthy control group, indicating that the cervical proprioception in the cervical spondylosis was impaired [55]. A systematic review including 14 studies suggested that JPE was significantly higher in the neck pain group than in the control group [79]. De Zoete et al. [77] in a systematic review and meta-analysis revealed a significant difference between idiopathic neck pain and healthy groups in JPE testing, indicating that this test may be clinically useful in assessing sensorimotor control. Similarly, another systematic review and meta-analysis conducted by Stanton et al. [11] aimed to synthesize and critically evaluate the available evidence for proprioceptive dysfunction in patients with chronic idiopathic neck pain by comparing the JPE of asymptomatic controls. Pooled estimates showed that people with chronic idiopathic neck pain had moderately impaired cervical JPS during head-to-neutral repositioning tests, compared with asymptomatic controls.

It is difficult to quantify pain and sensorimotor control. Although pain can be described and graded subjectively, sensorimotor control is more difficult to understand. In JPE testing, since the neutral neck position is a commonly used position, the memory head position may be recalled during the test [87]. Recall of the head position in a memory may involve more than one sensory component because memory may not only rely on proprioception. Therefore, the validity of this test in evaluating sensorimotor control may be limited. In studies using a preset target in the transverse plane, recall of the head position was less likely because the head was in a less common position on the neck. The head repositioning test makes indirect evaluation of proprioception possible [88]. However, the results vary greatly. Although most studies have reported reduced sensorimotor control in patients with neck pain, other studies [88, 89] have reported that patients with subclinical neck pain have fewer relocation errors than controls. Rix et al. [86] and Palmgren et al. [90] found that people with chronic, nontraumatic neck pain were no less accurate in head repositioning than healthy controls for all movement directions except flexion. In addition, other studies have reported that there is no measurable effect on cervical proprioception in the case of cervical pain [91‐93]. In general, these findings contradict the theory that the change of position sense may be due to the nociceptive input, or the pain experienced in the position-matching task provides additional feedback to improve the accuracy of position matching [29]. These contradictory results suggest the multifaceted nature of neck pain and the different mechanisms in sensorimotor control. For example, muscle fatigue and/or muscle tension and central nervous system regulation may change sensorimotor control [55]. Pain itself may interfere with muscle activity or adaptation of exercise strategies [49]. Cervical proprioception errors were significantly and positively correlated with the intensity of neck pain, indicating that the increase in pain intensity will impair proprioception [55, 94]. Different pain time courses such as acute, subacute, or persistent pain may have different effects on sensorimotor control [92]. Persistent pain may cause morphological changes in muscle components [53, 54]. In addition, an increase of age was thought to be related to the increase of cervical JPE. In elderly subjects, neck muscle strength decline, sedentary lifestyle, memory decline, and cognitive or motor impairment may result in damage to the target reproduction test [80, 81]. However, the literature on the relationship between age and JPE is contradictory [81, 83], which suggests that there may be other factors beyond age that affect the cervical proprioception [95]. Therefore, many causes may affect the proprioception and sensorimotor control of cervical spine.

After a rehabilitation program based on eye–head coupling exercises aimed to improve proprioception in the neck, Revel et al. [12] found that the ability to reposition the head after rotation was significantly improved, and neck pain was significantly reduced. There was no significant change in JPE in the control group. Jull et al. [13] achieved similar results in a group of patients with chronic neck pain using proprioceptive training. A randomized clinical trial evaluated the impact of balance training on JPS in patients with chronic neck pain, and found that joint repositioning accuracy was improved and pain was reduced in the intervention group, while no effect was observed in the control group [97]. Another randomized control trial revealed that proprioceptive training with a gaze direction recognition exercise combined with conventional physical therapy was more effective than conventional physical therapy for patients with chronic neck pain in improving neck disability and balance [98]. In a double-blind, randomized controlled trial, Saadat et al. [99] demonstrated that sensorimotor training combined with traditional physical therapy exercises could be more effective than traditional exercise alone in improving JPS, endurance, dynamic balance, and walking speed in patients with chronic neck pain (Table 2).

Deep muscles are important for maintaining neck posture. The support of the cervical segments mainly depends on the muscular sleeve formed by the longus colli muscle anteriorly and the semispinalis cervicis and cervical multifidus muscles posteriorly [100]. In order to stabilize the cervical segments, the deep cervical muscle activity should cooperate with the superficial muscle activity [101, 102]. Recent studies have found that the activation of deep flexors, the longus colli and longus capitis, is impaired in patients with neck pain [50, 51]. There is evidence that the strength and endurance of the craniocervical and cervical flexors decrease in the patients with neck pain, and therefore rehabilitation is required [103]. The muscle impairment determined by the craniocervical flexion test seems to be the common feature of various neck pain. These observations prompted the use of the craniocervical flexion test (Fig. 3) to retrain the deep cervical flexor muscles in the motor relearning program for patients with neck pain [104]. Craniocervical flexion training increased deep cervical flexor muscle EMG amplitude and decreased sternocleidomastoid and anterior scalene muscle EMG amplitude across all stages of the craniocervical flexion test [103]. In clinical exercise of craniocervical flexion, retraining the deep cervical flexors can improve the function of cervical proprioception, maintain neutral posture, increase the activation and endurance of the deep cervical flexors, and reduce the neck symptoms [13, 100, 105, 106].

The deep cervical extensor muscles are anatomically capable of coordinating cervical segmental movement with the deep cervical flexors [107]. The cervical extensor muscles are considered equally important in the recovery of patients with neck pain [108]. In patients with neck pain, the activity of superficial cervical extensors is enhanced and delayed offset (relaxation) after the activity [107, 109]. Compared with healthy controls, patients with neck pain showed structural and functional changes in the deep cervical extensor muscles [110‐117], emphasizing the importance of exercise in improving performance. Specific exercises of the deep neck extensors have not been extensively studied. A muscle functional magnetic resonance imaging study [118] in a group of healthy volunteers showed that both the deep and superficial extensors were activated below the C2 level when an isometric neutral head/neck extension was performed at 20% of the maximum voluntary force. Compared with the exercise in the neutral position, the same exercise in 15° of craniocervical extension increased the activity of the semispinalis capitis muscle, but not the deep extensor muscles below the C2 level. This is to be expected, because multifidus and semispinalis cervicis are the highest adherent to C2 and are therefore less affected by craniocervical extension. The results provided some preliminary insights into the impact of craniocervical orientation in response to different deep and superficial cervical extensors during performing cervical extensor exercises. O 'Leary et al. [119] then used the same method to compare the cervical extensor response patterns in patients with chronic neck pain with healthy controls, suggesting some alteration in the differential activation in cervical extensors in patients with chronic neck pain. High-load exercise is also recommended to increase muscle strength and endurance [107]. Twelve weeks of specific cervical resistance training resulted in a 34% increase in head-extension strength in healthy subjects [120]. Cross-sectional area increased by nearly 25% in splenius capitis, semispinalis capitis, semispinalis cervicis, and the multifidus, showing important hypertrophic effects [121]. A study using intramuscular EMG investigated the activity of the deep semispinalis cervicis and the superficial splenius capitis muscle at two spinal levels (C2 and C5) in healthy volunteers, and displayed the activation of the semispinalis cervicis relative to the splenius capitis when manual resistance was applied in extension over the vertebral arches (C1 and C4) [122]. A similar study in a group of women with chronic neck pain showed that localized resistance selectively activated the semispinalis cervicis muscle [123].