The results of this paper suggest that during the performance of the hollowing maneuver the CLBP group demonstrated “clinically meaningful” greater RA muscle activation levels (lower IO:RA ratios) compared to matched control groups indicating a bias toward RA when performing an abdominal hollowing exercise. As mentioned above, note that the term “clinically meaningful” indicates the observed difference is of sufficient magnitude and consistency to have at least 75% likelihood of having a substantial (meaningful) impact in a clinical setting. This indicates a spinal or supraspinal response to cervical orientation altered the activation patterns of the anterior trunk musculature.
As far as the authors are aware, this is the first study to assess the influence of altered cervical and limb position on the hollowing maneuver. These positions were intended to mimic the positions of PR. The novel finding in the present study was that activation levels were substantially affected by altered body position. This implies that in certain body positions CLBP patients illustrated an altered motor pattern when attempting to selectively activate their IO, and in other body positions had a motor pattern that clinically resembled the healthy population.
Effects of cervical and limb positions on motor patterns
While there are different abdominal musculature activation patterns between CLBP patients and healthy populations [
3,
4,
16,
43], it is consistent that there is an alteration in the mechanism of how the central nervous system controls the spine [
7,
44,
45]. One of the manifestations is a variable pattern of muscle substitution [
9,
44]. There are different activation patterns of the abdominal musculature during activities such as postural tasks [
46], gait [
47], trunk rotation [
48], holding a load [
49], orthopaedic tests [
50,
51] and specific exercises [
4]. Alterations in muscle activity are considered to be deficiencies in the coordination and control of the abdominal musculature and may result in a less stable spine during movement [
52].
The main objective of the abdominal hollowing maneuver is to bias activation of the TrA and IO while minimizing activity of RA. The present study illustrated that it was clinically likely, compared to controls, for the CLBP group to have higher activation levels of the left RA in the supine (cervical neutral), ATNR left and right, cervical rotation to the right and cervical extension positions while performing the hollowing maneuver. Similarly, the CLBP group had substantially higher levels of the right RA in the cervical neutral and cervical flexion positions. This difference illustrates that when performing the hollowing maneuver in these positions the CLBP group required greater RA activation to change the pressure cuff 10 mmHg. It is thought that muscle substitution occurs because the RA attempts to compensate for the deficient IO in the CLBP group [
16]. However, there is no previous research demonstrating that an alteration of body position changes the ratio of IO:RA during the hollowing maneuver.
The substantially increased level of left RA activity seen in the CLBP group was not illustrated on the right RA to the same extent. The CLBP group was clinically likely to have higher levels of left RA activity, compared to the control, in the cervical flexion, cervical rotation to the right, ATNR left/right, and cervical neutral positions. The CLBP group was clinically likely to have higher activation of the right RA, compared to the control, in only the cervical flexion and cervical neutral positions only. There are three possible explanations for this result. 1) There were a predominant number of participants with right-sided pain and thus increased contractions on the left side may have been predominant to help brace or stabilize the area. 2) Only certain PR were assessed therefore there could have been other PR or neurological soft signs (i.e. frontal release signs, clumsiness, motor incoordination, difficulty with motor sequencing) present that were not accounted for. 3) Finally the physiotherapist always stood on the right of the participant during instruction of the AHM, which may have affected the individual’s focus for ipsilateral motor recruitment.
The present study did not show any clinically meaningful differences between the two groups for IO activation. This would suggest that there is not a deficiency in the IO magnitude of the CLBP group compared to the control. It is postulated that to attain the 10 mmHg pressure change of the biofeedback cuff must require higher levels of IO activation. Therefore, there may be a ceiling effect for IO for the hollowing maneuver performed in this study. These results agree with O’Sullivan et al. [
4,
16] who used a similar protocol. Similarly, O’Sullivan et al. [
4,
16] did not illustrate a significant difference in activation levels of the IO between a CLBP group and healthy control. However, when a ratio of RA:IO was compared there were significant differences between groups, which indicated an altered motor program.
The CLBP group had similar activation levels of the left RA with the cervical rotation to the left and extension positions. Interestingly, these activation levels were altered by changes of cervical orientation and not the extremities. This indicates that the changes in activation were unlikely due to structural changes in the position of the muscle but more likely at the spinal and/or supraspinal level.
Other researchers have investigated the influence of altering limb position on aspects of motor control in humans. Cervical positions are known to alter the accuracy of upper limb movement in healthy people [
53‐
55] and elbow joint position error in subjects with whiplash associated disorders [
56]. The head may be used as a reference for the performance of upper limb movements and the altered proprioception of the neck may introduce error in the mechanism of central control of movement. CLBP subjects also have altered proprioception [
57]. It is unknown if the trunk uses the cervical spine in a similar manner to the upper limb, however it could be possible that a similar mechanism is involved in the altered motor patterns observed here. Investigations on humans have found an influence of altering neck position on the motor system. Deutsch et al. [
58] reported that head repositioning may affect the strength of the upper limb through the influence of the tonic neck reflexes. LePellec and Maton [
59] concluded that the tonic neck reflexes can have a small influence on high force production with elbow flexion.
In normal function, neck receptor influences on muscle tone are involved with complex postural responses, which also reflect sensory information from visual, vestibular, proprioceptive and somatic sources [
30,
60]. These will interact strongly to each other and will likely have variable individual influences. Turning the head to one side accentuates the extensor tone of the limbs on that side, with flexion on the contralateral side [
60]. Given the different levels of integration of PR in adults, the tone changes with repositioning of the neck can be variable and complicated.
CLBP may be associated with an overall reduction of CNS inhibition. Baliki et al. [
61] theorized that the increased activation of medial pre-frontal cortex (PFC) in individuals with CLBP, compared to controls, might be due to a disruption of the mutual inhibitory interactions. A decrease in inhibition could help explain why motor cortical maps increase in volume with CLBP [
62,
63]. If CLBP causes a decrease in inhibitory interneurons then the altered activation pattern of the anterior trunk muscles of the CLBP group during the hollowing maneuver compared to the control may be due to over activity of the CNS. In terms of muscle substitution, it is generally thought that the RA increases its activation level to make up for a deficient ability to activate IO [
16]. While theoretically and functionally this makes sense, recent research on CLBP and brain morphology and activity poses an alternate explanation. With a decrease in gray matter volume and density it is mainly a loss of inhibition that results [
26]. Therefore the increased activation levels of RA may be due to an inability to inhibit this activation when attempting to perform the hollowing maneuver. This explanation would support the results in the present study as our CLBP subjects exhibited substantially higher normalized levels of both left and right RA compared to a matched control in a variety of different positions but had similar levels of IO activation. This indicates that while both groups were able to activate IO to a similar extent, however the control group was substantially better at activating IO while minimizing activity in RA.
Primitive reflexes with CLBP
While reduced inhibition helps explain muscle substitution, it does not clarify why the present study showed that altering cervical orientation can substantially affect RA activation. Subgroups of subjects with CLBP have been found to have significantly higher levels of PR than other groups [
28]. PR are brainstem mediated movement patterns which are inhibited by areas in the frontal lobe [
64]. PR typically start to be inhibited at six months [
21] and their presence is used to assess CNS integrity [
23]. While it is unknown whether the PR in this current population has resurfaced, as it does with normal advanced aging, or if they have been present throughout the subjects’ life it can possibly indicate CNS disruption.
Age is an unlikely reason for the resurfaced PR as the population employed had an average age of 45 and PR re-emergence is usually not seen until the sixth decade [
65]. Similarly if PR are resurfacing it would agree with the theories that there is an overall reduction in inhibition associated with reduced grey matter and CLBP [
26,
61]. If PR are present in an individual with CLBP it may be possible that the altered motor strategy is due to a reduction in supraspinal inhibition.
There was an expectation that the muscle activation patterns of the CLBP group would be more similar to the control during performance of the hollowing maneuver when placed in a position mimicking either ATNR or the supine extension reflexes (e.g. Moro reflex, tonic labyrinthine reflex {TLR}). Placing our CLBP subjects in the ATNR position with altered position of the extremities did not substantially affect performance compared to controls. However, cervical rotation to the left with the hollowing maneuver by CLBP, had activation of left RA similar to controls. Likewise, cervical extension illustrated similar levels of RA activation between groups. How cervical orientation affects trunk muscle activation patterns in this study can only be speculative. In this study, the subjects’ body and limb orientations were not performed passively. They actively placed themselves in these positions. Hence it may be possible that the process of consciously placing the limb in a position resembling the PR is what provided inhibition.
Perhaps by inhibiting this reflex with altered head and limb orientations, it is reopening latent inhibitory synaptic pathways in the frontal lobe. This may in turn access other inhibitory pathways allowing the CLBP patients to activate IO while also inhibiting activation of RA during the hollowing maneuver. Wand et al. [
66] also came to the conclusion that widespread disinhibition may be a fundamental issue with CLBP and that treatment paradigms that elicit intracortical inhibition should be explored.
Limitations
This study poses new insight into both muscular activation patterns of CLBP patients as well as how altering cervical orientation can affect these activation patterns. However, the results must be considered within the limitations of the study. In this study, only surface EMG electrodes were used. At the site of IO, there will be recordings from TrA since it lies directly beneath this point. McGill et al. [
67] reported that surface electrodes adequately represent the EMG amplitude of the deep abdominal muscles (i.e. TrA and IO) within a 15% RMS difference. Ng et al. [
68] indicated that electrodes placed medial to the ASIS would receive competing signals from the EO and TrA with the IO. Based on these findings, the EMG signals obtained from this abdominal location are described in the present study as the IO, which would be assumed to include EMG information from both the TrA and IO. However this limitation should not affect the interpretation of the results in this study for three reasons. 1) Anatomically it has been shown that the lower fibers of both IO and TrA have similar orientation and attachments [
69]. 2) Likewise it has been proposed that they have similar synergistic functions in ipsilateral rotation and sacroiliac joint closure [
70]. 3) Finally it has been shown that the hollowing maneuver is performed by the combined activity of IO and TrA [
71]. Because of the similarity in function and anatomy, these two muscles have been recorded together with surface electrodes in a number of studies from this laboratory and their EMG activity have been differentiated from other neighbouring muscles such as the RA and EO [
34‐
37]. A further limitation of the study was that CLBP patients without PR were not included in the investigation. In addition, palpation for the presence of activation of TrA may also have been helpful but due to the number of researcher responsibilities during the experiment (e.g. monitoring EMG, PBU, subject’s pushing of feet, performance of hollowing maneuver and others), it was not possible to add this additional measure.
The expression of our results may not be familiar to all readers. We purposefully omitted p-values and discussion of statistical significance. While both p-values and 95% confidence limits can be used to infer statistical significance (see last paragraph of results section), 95% confidence limits are much more information-rich to the clinician. The p-value only represents the probability of the mean response to a treatment not being zero. Concluding that a mean was “unlikely to be zero” is not a clinically useful conclusion because it does not express the variability in responses in clinically meaningful units. However, expressing results using a mean in conjunction with upper and lower confidence limits allows clinicians to easily interpret the likely effects they can expect from an intervention.
Local and global abdominal hollowing
It is clear from the descriptions used in the literature for laboratory research and clinical trials on abdominal hollowing that there are different versions of the exercise in use [
12,
72,
73]. In general, there are two types of abdominal hollowing. The first aims to bias TrA over IO, EO and RA. This may be considered “local abdominal hollowing”. During the hollowing maneuver described here and by O’Sullivan et al. [
4,
16], it should be noted that the change in the pressure biofeedback unit from 40 mmHg to 50 mmHg does not occur with TrA alone. This 10 mmHg change requires higher levels of superficial muscle activity. Here, the goal of the exercise is to bias activation of IO while minimizing activity of RA. This may be considered “global abdominal hollowing”. The former exercise is designed for translation control of individual spinal segments, while the latter is designed for movement control [
4,
16]. It should be noted that the ‘global’ abdominal hollowing may be performed as a progression of the ‘local’ abdominal hollowing or independently of it. Invariably, there will be some element of translation control with the ‘global’ abdominal hollowing even if it is performed independently of the ‘local’ abdominal hollowing. This requires further clarification, but is beyond the scope of this paper. In this study, ‘global’ abdominal hollowing was taught as a progression from ‘local’ abdominal hollowing.
The change in the PBU from 40-50 mmHg may also consist of a contribution from a co-contraction with lumbar multifidus since this is known to co-contract with TrA [
74]. Given that the change in the PBU depends on a change in the orientation of the posterior abdominal wall, and this may be at least be partially dependent upon intra-abdominal pressure, it may be permissible that coordination of the whole deep muscle system or cylinder (TrA, diaphragm, pelvic floor, psoas major, deep lumbar multifidus) contributes to this change [
72]. Although other studies have looked at abdominal hollowing [
75,
76], only O’Sullivan et al. [
4,
16] have looked at ‘global’ abdominal hollowing, used EMG and as a monitoring tool and standardized the amount of abdominal hollowing with a PBU.