Anatomical study and proposed EMG technique for the cervical paraspinal muscles

Paraspinal electromyography has proven to be the most sensitive component of the electrodiagnostic examination for lumbar spinal disorders. However, no standardized, anatomically validated technique has been proposed for the cervical region. This study reviewed the published textbooks on cervical paraspinal anatomy to develop a standardized electromyography technique and scoring system. A library search found 32 anatomy texts published between 2000 and 2021. Of these 11 were unique and appropriate. Most texts described the basic muscle anatomy similarly, but only one cited original research. When the spinous process is defined as the origin, the multifidus and deeper rotatores appear innervated by the posterior primary rami of single cervical roots. However, texts differ in the number of pennae, between two and five, traveling to transverse process regions below. These are crowded into a small area between the spinous processes and transverse processes. Based on this understanding, a proposed cervical paraspinal mapping technique involves skin insertions from 1 to 2 cm lateral to the C5, C7, and T2 spinous processes. The needle samples transversely and deep toward midline, contacts bone, then is withdrawn and redirected to sample medial and caudally to midline to bone, creating two scores of 0–4 at three levels, theoretically resulting in scores of 0–24. This technique must be validated by clinical research to determine the range of normal, reproducibility, and the spectrum of findings in various disorders.


INTRODUCTION
Cervical radiculopathy is a common and important problem frequently seen by electrodiagnosticians, but it is difficult to diagnose because symptoms could be attributed to a large differential diagnosis list. 1 Theoretically, examination of the paraspinal muscles can be useful in diagnosing cervical radiculopathy.3][4] A literature review failed to uncover a standardized, anatomically validated technique or scoring system for evaluating the cervical paraspinal muscles. 5This lack of standardization creates uncertain sensitivity and specificity of the electromyography (EMG) test in diagnosing cervical radiculopathy.
The purpose of this paper is to take the first step in improving the diagnostic power of cervical EMGs by presenting a standardized cervical EMG technique and scoring system based on the current textbook understanding of cervical spinal anatomy and clinical experience.7][8]

METHODS
To understand what is known about the anatomy of the cervical multifidus, available anatomy textbooks were reviewed.Qualifying textbooks were found in a search of a medical school library using the search term "human anatomy."Results were limited to books in English that were available online, published from 2000 to 2021 and had human anatomy as one of the stated topics.
An anatomically valid EMG technique was created by first collating the findings of texts or articles that describe the segmental anatomy of the multifidus muscle.This was mapped to a general construct evolved from a technique proposed for the lumbar spine by MacIntosh and Bogduk and found anatomically valid for lumbar EMG. 2,4This construct involves inserting a needle to the corner of the spinous process and transverse process until contact with bone.

RESULTS
The initial search yielded 32 texts. 9-39Some of these books were not appropriate, were abridged versions, or were duplicated results.[9][10][11][12][13][14][15][16][17] As the focus for the EMG technique was on the anatomy and innervation of the multifidus, we recorded the ways this muscle was described.0][11][12][15][16][17] Only one book provided a primary reference for the information presented. 10hese books described cervical muscles superficial and lateral to the multifidus that, nevertheless, may be sampled by needle EMG.Deep to the superficial trapezius muscle and the intermediate respiratory back muscles are three groups of muscles that all function to support the spine and stabilize the head.Of these, the most superficial muscles are the splenius capitis and the splenius cervicis, which together are in the spinotransversales group and are innervated by posterior rami of the middle and lower cervical nerves.Laterally and deep to the spinotransversales muscles are the erector spinae muscles, which include the iliocostalis, longissimus, and spinalis and are generally the largest integral (deep) back muscles.
Deep to the spinotransversales group of muscles are the transversospinales muscles.These muscles stabilize the spine bilaterally by pulling the head and spine posteriorly.All of them attach inferiorly from a lateral position on or near the transverse process of the vertebrae and travel medially-superiorly to the spinous process, which gives them a steep to very steep angle from spine depending on the number of levels they cross.The transversospinales group includes three muscles, which are the semispinalis, the multifidus, and the rotatores.The semispinalis are the most superficial and superior running only in the thoracic and cervical region and span four to six vertebrae with the most superior members attaching to the base of the skull.The multifidus is deep to the semispinalis and these muscles are present throughout the length of the spine, though several books mention they are most developed in the lumbar region.Deep to the multifidus are the rotatores, which are smaller than the multifidus as they span only one to two vertebrae, but the group also runs the length of the spine with their best development in the thoracic region.
,15,17 Two of nine books (22%) describe the origin as the transverse process for the entire spine and do not describe if that changes depending on what region of the back they are describing.8,14 Eight of nine (89%) books describe the insertion point of the multifidus as the spinous process.[7][8][9]11,12,14,15,17 The one outlier from the other eight books, the 2016 version of Gray's Anatomical Basis of Clinical Anatomy, describes the muscle inversely as moving from the spinous process inferiorly to the transverse process.Interestingly, this is also the only book that provides a primary reference for its claims and has the most detailed description of the multifidus muscle.6 The primary reference used for this text, by MacIntosh and Bogduk from 1991, describes quantitative measures of the muscles on dissection of five cadavers and 21 sets of anteroposterior and lateral radiographs.They found that defining the origin as the spinous process resulted in a more rational understanding of innervation, with all multifidus originating on a spinous process innervated by the nerve root that comes out from under that vertebra.6 A Medline search for articles citing MacIntosh and Bogduk found no subsequent anatomical research on the cervical paraspinal muscles.Of the books that speak to innervation of the multifidus, all nine (100%) agree that innervation comes from the posterior rami of the spinal nerves.[7][8][9][10][11][12][15][16][17] There is variation on how the different books describe the number of levels the multifidus muscle spans.Five of the seven (71%) books that specify the number of levels the multifidus spans say the multifidus spans between two and four levels.7,9,11,12,15 One of these books says the multifidus spans exactly three levels. 13 Gray's 2016,the only book that cites research, states that the multifidus spans between two and five levels.10 T A B L E 1 Findings regarding the multifidus muscles in anatomy textbooks.After reviewing the various descriptions in the textbooks, one of the authors (T.J.M.) was able to inspect dissected cadavers.He found that the area in the cervical spine where the many layers of the multifidus muscle reside was very small.Given a theoretical concern about needles penetrating the dura, he observed that a superiorly directed needle had some potential to do this, whereas direct and inferior insertions seemed much less likely to do so.

DISCUSSION
This study set out to determine an anatomically valid technique for cervical paraspinal EMG.It found a general cohesion between the various texts regarding the descriptions of the paraspinal anatomy but almost no citations of actual research or data to justify the author's descriptions.Nevertheless, a final anatomic schema evolved that allows us to propose a scorable technique.The particular findings and aspects of the methodology should be discussed before proposal of a clinical technique.
Table 1 shows that most anatomy texts agree on some generalities but fail to agree on finer details such as the number of levels the multifidus spans (two to five, two to four, three) or the precise attachment points in the different sections of the spine (transverse process or articular process).Interestingly, the format and wording of information presented is remarkably similar in four of the textbooks, though details differ slightly. 7,11,12,15These books provide little evidence behind their pictures and statements.All but one book fail to provide references with any data. 10Regarding the cervical region there is a sense that the authors followed each others' writing or historical writing without critical thought.Some books of the same brand, but with different editors, give contradictory descriptions in the number of levels the multifidus spanned. 9,10Due to the lack of consistency in details across anatomy textbooks, the current authors relied more heavily on the textbook that was backed by primary research along with viewing the muscle groups directly in the dissection laboratory.A prospective blinded cadaver study has validated the accuracy of anatomical placement as we will describe.(unpublished data).
Based on this research, the authors are confident in an understanding that the transverospinales muscles are the deepest back muscles that stabilize the spine.Importantly, these muscles are segmentally innervated by spinal posterior primary ramus nerves.These nerves attach superiorly at the collocation of the origin of these muscles at the spinous process and then the muscles travel inferiorly-laterally down to the insertion at the transverse or articular process.
In another paper (accepted, PM&R) the authors performed a review of clinical research and reviews on paraspinal muscles for electromyography of cervical radiculopathy. 5Similar to the anatomy texts, the clinical texts and peer-reviewed literature of cervical paraspinal EMG techniques contain contradictory information if any information at all.Almost no publications described a methodology.Anatomical guidelines were nonexistent in most, vague in others.Specific scoring systems were absent in all studies, and no range of findings in asymptomatic persons, or those with disease, were described in any study.The closest exception is a study by Date et al. 40 that describes her technique but judges the entire cervical region as positive if one reproducible train of fibrillations is found. 39o date, the EMG diagnosis of cervical radiculopathy is not considered a standard in the surgical literature.Yet it is the clinical impression of many neurologists and physiatrists that this test has value.To validate or reject this impression, it is important for the test to have valid and reproducible standards.For example, when a reproducible standard for lumbar paraspinal EMG was evaluated, it became the only part of EMG found acceptably valid in the North American Spine Society's spinal stenosis guidelines. 41In addition, standardization decreased observer bias, increased sensitivity of the test, and allowed for a deeper understanding of the pathophysiology of spinal disorders. 3ased on the review of anatomical and clinical research and examination of cadavers, a systematic cervical paraspinal mapping technique was created.This technique takes advantage of the most conservative overlaps cited in textbooks to minimize insertions.Based on the anatomy (but as yet without clinical confirmation) it appears that the insertions at the C5 spinous process would sample C2 3,4 innervated muscles, that C7 bony level insertions would sample C4, C5, C6, and T2 would sample C7, C8, and T1.However, the anatomists disagree on how many levels the muscles travel, and the actual specificity and sensitivity of the needle exam can be determined only clinically.
The technique is as follows: The patient is positioned to relax the neck muscles.Typically, this is while sitting in a chair with forehead resting on a padded table or counter, flexed forward, sometimes with just a  bit of active chin tucking to relax the muscles.Alternatively, the patient can be side-lying with ample pillow support and forward flexed.The C5, C7, and T2 spinous processes are marked.A location 1-2 cm lateral to each is marked (a bit more laterally in persons with more paraspinal adipose).At the T2 location, a 50-mm or longer monopolar needle penetrates the skin and is aimed medially and deep (depth depending on adiposity may be 60-80 , with a goal of contacting the base of the spinous process).Upon entering muscle, insertional activity and spontaneous activity are elicited through 5-mm insertions until the needle contacts bone.Positive waves or fibrillations that last >1 second that are not associated with motor end plate noise or with the periosteum are scored.The scores are 0 (none), 1 (lasts >1 second), 2 (2 different waveforms or insertions), 3 (multiple waveforms), and 4 (waveforms obliterate the baseline at some point).The needle is withdrawn to the surface, directed somewhat inferiorly (20-45 , again depending on adiposity) and this is repeated, obtaining a second 0-4 score.This is repeated at the C7 and the C5 locations.There are finally six scores of 0-4, totaling 0-24 for one side of the neck.
Figure 1 provides a schematic of the technique, Figure 2 a description of the technique, and Figure 3 the scoring template.
Not described in this paper is the authors' use of this technique in clinical cases.Although the cervical paraspinal muscles are more difficult to localize and relax than the lumbar muscles, the proposed technique subjectively appears to be a viable approach to those muscles.No formal data have been gathered, but there is some relationship of findings with the clinical suspicion of radiculopathy.
It is worthwhile to note first that the targeted muscle is small and deep.Figure 4 is transverse section of a magnetic resonance imaging scan, showing that the targeted muscle mass is deep under other muscles and that a specific fascicle of the muscle, potentially innervated by only one posterior primary ramus, may be only a few millimeters thick.EMG of the cervical paraspinals that does not target this small space between the spinous process and the transverse process is likely to miss all of the muscle, and even within that space one cannot be confident that the intended root-specific fascicle has been sampled.This is in contrast to the relatively large lumbar multifidus, which is mostly not overlapped by other muscles.
This proposed technique has a long way to go before it becomes a standard.The next steps include prospective gathering data on a large number of clinical cases to understand the range of findings, study of asymptomatic volunteers to establish norms, then prospective and masked trials in persons with appropriate cervical diseases.The end point may be research to establish or reject the idea that the findings of cervical paraspinal EMG affect actual clinical decision making about things like injections, surgery, novel medication interventions, and prognosis.

CONCLUSION
This study synthesizes knowledge of anatomy and clinical EMG research to propose an anatomically valid technique and scoring system for cervical paraspinal EMG.This technique must be validated by clinical research to determine the range of normal, reproducibility, and the spectrum of findings in various disorders.

A
. Palpate the spinous process of C5, C7, and T2.B. Mark 1-2 cm lateral to the spinous process.C. At each site, insert a needle at 45 degrees to the spinous process and perpendicular to the longitudinal direction of the spinal column.D. With 0.5 cm increments, insert the needle until contact with the vertebrae and record fibrillation and positive waves on scale of 0-4+.0 = no fibrillations >1 second 1 = one fibrillation seen 2 = more than one fibrillation seen 3 = Fibrillations seen frequently 4 = Fibrillations fill the baseline at some point E. Withdraw the needle to surface and then from the same insertion point, redirect the needle at 60 degrees pointing to the inferior and repeat the insertion interval and scoring.(No cranial directed insertions to avoid the spinal cord.)F. Record findings and tabulate total score.

F 3
Cervical paraspinal mapping scoring form.

F I G U R E 4
The depth and small size of the targeted muscle.On this T2-weighted magnetic resonance imaging transverse section across the lower cervical spine, brackets A show the 4.5 mm width of a portion of the multifidus, likely all associated with one spinous process and one nerve root.Brackets B surround the entire transversospinalis group including fascicles innervated from 3 to 5 levels above.Line C shows depth to superficial fat, indicating that the distance to penetrate the intended muscle to bone is 37 mm.
How to cite this article: Haig AJ, McGuire TJ.Anatomical study and proposed EMG technique for the cervical paraspinal muscles.PM&R.2024;16(2):165-173.doi:10.1002/pmrj.13046Thisjournal-based CME activity is designated for 1.0 AMA PRA Category 1 Credit™.Effective January 2024, learners are no longer required to correctly answer a multiple-choice question to receive CME credit.Completion of an evaluation is required, which can be accessed using this link, https://onlinelearning.aapmr.org/.This activity is FREE to AAPM&R members and available to nonmembers for a nominal fee.CME is available for 3 years after publication date.For assistance with claiming CME for this activity, please contact (847) 737-6000.All financial disclosures and CME information related to this article can be found on the Online Learning Portal (https://onlinelearning.aapmr.org/) prior to accessing the activity.