Surface electromyography activity in the upper limbs of patients following surgery for compressive cervical myelopathy
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.170071
Source of Support: None, Conflict of Interest: None
Background: Surface electromyography (EMG) is a noninvasive, accurate method to measure electrical activity produced in muscles.
Keywords: Cervical; electromyography; myelopathy; spasticity
Spasticity is a state of increased tone in muscles resulting from hyperactivity of the stretch reflexes. It has been defined as a motor disorder characterized by a velocity-dependent increase in the tonic stretch reflex, as one component of the upper motor neuron (UMN) syndrome. Patients with compressive cervical myelopathy classically present with the triad of spasticity, weakness, and sensory deficits in the upper and lower limbs. Appropriate surgical relief of cord compression can improve outcomes in all three domains. The Nurick grading and the Modified Japanese Orthopaedic Association (mJOA) scoring system are typically used to assess functional status in these patients. However, these are subjective tests and suffer from inter-rater variability. Similarly, the commonly used Modified Ashworth Scale (MAS), a measure of spasticity, has been criticized for its lack of objectivity being relatively inconsistent and ineffective in detecting improvement in spasticity when compared to neurophysiological testing with electromyography (EMG).,, The more recently introduced Modified Modified Ashworth Scale (MMAS), which replaces the grade "1+" with "2," has shown a good intra- and inter-rater reliability.,
Surface EMG recordings are not only noninvasive but also have a large pick-up area and are ideal for assessing gross muscle function. They are well-suited for understanding the temporal relationship between EMG activity and muscle contraction dynamics, and to a limited extent, the magnitude of muscle contraction force. Objective identification of changes in upper limb spasticity in patients with compressive cervical myelopathy used biomechanical methods employing a dynamometer to document improvement in elbow spasticity. In this study, we wished to observe changes in surface EMG activity before and after surgery by recording from the antagonistic muscles of the forearm that are responsible for pronation and supination. The secondary aim was to correlate the changes observed in surface EMG activity before and after surgery with the MMAS.
Thirty-one nonconsecutive patients with cervical compressive myelopathy and spasticity (MMAS score ≥1) were chosen for the study; of these, 28 (90.3%) were male and 3 (9.7%) were female patients. Their ages ranged from 25 to 72 years (mean 54.1 ± 10.1 years). Patients with Nurick grade 5 (since transfer to and from the neurophysiology laboratory was difficult), those with lower motor neuron findings on clinical examination, and those with joint deformities, contractures or thrombophlebitis of the upper limbs were excluded from the study. All patients underwent a detailed pre- and postoperative neurological evaluation. Their functional status was assessed by the Nurick grade and mJOA score by one of the authors (SB). The MMAS score was determined for the subjective assessment of spasticity. EMG testing was performed in the patients preoperatively and 1 week after undergoing surgery by another author (MM). The study was approved and supported by grants from the Institutional Review Board Min no. 6698, dated November 19, 2008. Patients were enrolled after obtaining an informed consent.
Normative electromyography data
Normative electromyography (EMG) data were collected from 31 age-related subjects (25 males and 6 females) with a mean age of 53.4 years (range 29–62 years) by administering the test as described below.
Description of the apparatus and the test
The patients were assessed by the MMAS score, and the more symptomatic limb was chosen for EMG recordings. Patients were seated comfortably in a chair with the elbow supported in a holder attached to the chair. The skin of the forearm and arm was prepared with alcohol (70%) to decrease impedance. A two-channel recording was done using surface electrodes (Discoid Ag-AgCl gel) of 10 mm diameter which were applied with an inter-electrode distance of 10 mm (this was to ensure that the electrodes record from the same group of muscles when the forearm is rotated) as follows: For the biceps brachii (representing forearm supinators), at 1/3rd of the length from the cubital fossa to the acromion process, and for the pronator teres (representing forearm pronators), 3 cm distal to the midpoint of the line connecting the medial epicondyle of the humerus and the biceps tendon. The ground electrode was placed between the two channel recording electrodes, against the medial epicondyle. The surface EMG activity was measured from these two muscles while pronation-supination was performed actively by the patient and then passively by the investigator who created a pronation-supination movement by rotating the device that held the arm. While clinically assessing these patients, increasing velocities were used till a resistance or "catch" was felt, and a comparable range of velocity was used by the investigator during the surface EMG activity. The following EMG data were collected from the pronators and supinators during active pronation alternating with supination followed by passive movement in which the patient's forearm was pronated and supinated alternately by the clinician by holding an extended disc.
Electromyography signals were amplified (×666) using a high precision instrument amplifier and passed through a band filter (15 Hz–1.5 KHz) to reduce movement artifacts and other noise influences as per standards. The range of movement was encoded using an optical disc encoder with a resolution of 8°. The analog signal of EMG activity was fed to a personal computer through a data acquisition card (National Instruments) at 1000 samples with a resolution of 14-bit conversion. To record the EMG and position signal, a custom graphical user interface was designed. Similarly, to review the recorded signal, another interface was made using Labview 8 (National Instruments). The neutral position taken for the test was the fully supinated forearm. A direction sensor gave information about the direction of movement and a position encoder provided information on the start and end of each movement. Each pulse of the encoder was 8°. Therefore, the range of movement was calculated from the number of pulses, and the velocity was derived from the time taken for the complete range of motion. Three trials were performed, and the average value was taken for analysis. The built-in review program gave the plot of RMS EMG activity, raw EMG activity, range of movement, and angular velocity. [Figure 1] illustrates the working of the apparatus and [Figure 2] is a picture of the custom made device used in the study. [Figure 3] shows the surface EMG activity during alternate active pronation and supination in a normal subject whereas [Figure 4]a and [Figure 4]b show the preoperative and postoperative traces, respectively, of a 46-year old lady with cervical spondylotic myelopathy (CSM) who underwent a two-level anterior cervical discectomy. Her MMAS score and mJOA grade remained unchanged after surgery although there was a definite subjective improvement in her spasticity.
Grouping of postoperative patients
For further comparison of clinical and EMG data, we excluded 10 patients with a preoperative MMAS scores of 1 since this was considered to be almost normal. The remaining 21 patients were categorized into three groups based on the subjective feeling of improvement of their spasticity after surgery: No improvement, mild improvement, and good improvement.
SPSS (Version 16.0. Chicago, SPSS Inc.) was utilized to analyze the data. Descriptive statistics such as mean or median were used for continuous variables and frequency distributions for categorical variables. The independent sample Student's t-test was used to calculate the significance of the association for continuous variables between study and control groups, and paired t-test to compare pre- and post-surgical assessments in the study group. We used the Pearson correlation coefficient to assess the significance of associations. Mean and median were calculated based on the distribution of parameters. A P < 0.05 was taken as statistically significant and <0.001 as highly significant.
The most common etiology of cervical compressive myelopathy in our study was CSM (n = 17) while other etiologies were inter-vertebral disc prolapse (n = 11) and craniovertebral junction anomaly (n = 3) [Table 1]. The duration of symptoms ranged from <6 months to 5 years. Most patients underwent an anterior cervical discectomy, laminectomy or corpectomy (central or oblique). Three patients underwent instrumented fusion for craniovertebral junction anomalies. Postoperatively, there was a significant improvement in MMAS, Nurick grade, and upper limb mJOA (ulJOA) score as shown in [Table 2].
Surface electromyography recordings in controls and patients
The baseline EMG activity (±standard deviation [SD]) in the pronators and supinators was less in the controls than in the patient population. In the pronators, it was 0.031 ± 0.011 mV in controls and 0.069 ± 0.025 mV in the patients (P < 0.001). The mean baseline activity in the supinators of the controls was 0.015 ± 0.007 mV and in the patients, it was 0.076 ± 0.078 mV (P < 0.001).
The EMG activity generated during active and passive pronation and supination in preoperative patients was significantly greater than that of controls in both agonist and antagonist muscles as shown in [Table 3]. The agonist activity in pronators and supinators during active movement was greater by 24% and 40%, respectively. The co-activation noted during active pronation-supination was also significantly more in patients compared to controls by 84% in both the pronators and supinators. Similarly, during passive movement, it can be seen that the co-activation of the muscles being stretched was significantly higher in patients than in controls—approximately 43% for pronators and 48% for supinators. EMG recording done 1 week postoperatively showed a reduction in baseline activity in the pronators and supinators by 21% and 36%, respectively. There was a decrease in co-activation of the pronators during active supination by almost 62%. Similarly, co-activation of the supinators during active pronation reduced by around 33% (P < 0.05). During passive movement as well, there was a decrease in co-activation of the pronators during supination by approximately 23%. There was also a decrease in activity of the supinators during pronation by almost 35% (P < 0.05) [Table 4].
All patients reported a subjective improvement, either mild (n = 20) or good (n = 11) of their symptoms, postoperatively. Patients who reported good improvement after surgery showed a significant decrease in the pronator activity during active (P < 0.05) and passive (P < 0.05) supination; in other words, a reduction in co-contraction. However, supinator activity during pronation did not differ significantly between these two groups.
Among the 21 patients who had preoperative MMAS scores >1, 15 improved in their postoperative MMAS scores, while in 6 patients, the MMAS scores remained the same. [Table 5] shows the preoperative and postoperative mean RMS EMG amplitude of these patients in the pronators and supinators during active pronation and supination. We found a significant reduction of co-contraction in the pronators during supination in all the patients, including those whose MMAS scores remained the same postoperatively. During passive movement, there was no significant decrease in co-contraction in both groups.
Range and velocity of movement
We also compared the mean range of movement (in degrees) and mean velocity of movement (degrees per se cond) between patients and controls during active pronation and supination. Patients had a mean range (±SD) of 123.90° ± 28.14° and a mean velocity of 396.29° ± 149.17°/s during pronation, which was significantly lower than in controls, who had a mean range of 139.16° ± 13.95° and a mean velocity of 487.68° ± 91.91°/s (P = 0.009). During active supination, it was seen that only the range of movement was significantly lower among patients (139.68° ± 14.89° vs. 125.16° ± 29.86°, P = 0.018). However, when the range and velocity of pronation-supination were compared between patients preoperatively and postoperatively, we found no significant improvement.
The signs and symptoms of the UMN syndrome are classically divided into "positive" and "negative." The term "spasticity" has become synonymous with the positive features. However, in reality, spasticity is only one of the several "positive" phenomena, which occur as part of the UMN syndrome. In the spinal model of spasticity, as seen in our patients, partial spinal lesions usually involve the lateral corticospinal tract and dorsal reticulospinal tract. Damage to the corticospinal tract leads to weakness while the loss of inhibitory influences from the dorsal reticulospinal tract, leaves the effect of the facilitatory medial reticulospinal tract and vestibulospinal tract unopposed, giving rise to spasticity. In this situation, there is often significant spasticity, with tone being greatest in the antigravity muscles. In severe or complete cord lesions, the supraspinal influence on the cord is completely abolished. The hypertonicity is sometimes less marked in such patients due to the fact that the descending excitatory systems are no longer acting unopposed., Another closely related entity, which usually contributes to the clinical assessment of spasticity, is pathological co-contraction of muscles. It arises due to the failure of reciprocal inhibition, which was originally described by Sherrington in 1906. The other positive signs include hyperactive tendon reflexes, clonus, and flexor spasms. The negative signs are paresis and loss of dexterity.
In 2003, Engsberg et al., published their study on a single patient who underwent an anterior cervical discectomy for CSM. Spasticity was assessed at the ankle and elbow using a Kin-Com isokinetic dynamometer. Spasticity was estimated based on the work required by the dynamometer to move the passive joint throughout its range of motion at various speeds. They found a significant decrease in elbow flexor spasticity after surgery. This study laid the foundation for the use of biomechanical methods to measure changes in function and impairments associated with surgical intervention. In a study by Van der Salm et al., of the 9 patients with complete spinal cord injury, the authors developed a method of assessing spasticity in which the whole range of motion at a wide variation of speeds was applied. They developed an apparatus that included a footplate which allowed complete plantar and dorsiflexion of the ankle joint. EMG electrodes were placed on the soleus muscle, and dorsal and plantar flexion movements were manually applied by the examiner at varying velocities. The resistance to passive movement was measured as torque using a calibrated strain gauge; ankle angle was determined using a potentiometer at the axis of rotation; and, angular velocity was determined with a gyroscope fitted on the footplate. EMG of the soleus muscle was measured using surface electrodes. They found that the RMS values of the EMG responses increased significantly at increasing stretch velocities. They also demonstrated several advantages of manually performing the testing as opposed to a motorized device; the setup would be less complex, less expensive, and more applicable in a clinical setting. The basis of our apparatus and the inclusion of the measurement of range of movement and angular velocity stemmed from this study.
The Ashworth scale  and the MAS  are commonly used methods to clinically assess the degree of spasticity. The recently introduced MMAS has been shown to have a good inter- and intra-rater reliability ,, but not superior to the MAS in the same regard. The Ashworth scale suffers from clustering of most patients within the middle grades. It offers ease of measurement but may lack temporal reproducibility. Sköld et al., were the first to directly compare simultaneously performed clinical and neurophysiologic tests. They obtained EMG recordings in the knee flexors and extensors during flexion and extension in 15 men with motor-complete tetraplegia, and the degree of correlation between the MAS and EMG recordings was analyzed. They concluded that all included EMG parameters significantly correlated with simultaneous Ashworth measurements of spastic muscle contraction. On the other hand, a similar study  done more recently on 10 poststroke patients with upper limb spasticity before, 30 days after, and 180 days after, botulinum toxin type A injection suggested that the MAS may not be a sensitive enough measure of spasticity and, therefore, an incomplete tool to measure spasticity. They based this on their finding of a poor concordance between the MAS scores and EMG activity and proposed that this could be due to the different constructs of the two evaluations: MAS is a measure of resistance to passive movement, while EMG quantifies levels of muscle activity and is a direct measure of spasticity. The limitations of their study were the small sample size and an absence of a control group, which we sought to overcome.
In our study, the EMG activity among patients with cervical compressive myelopathy during active pronation and supination was found to be significantly higher than in controls. This is due to their muscles being weaker and hence the need to recruit more motor units to complete similar tasks. Similarly, during passive movement, it can be seen that the co-activation of the antagonists was significantly more in patients than in controls. After surgery, the improvement was seen in patients, both in terms of decrease in baseline EMG activity as well as in co-contraction. This improvement after surgery is presumed to be due to the descending excitatory systems, namely the pyramidal and extrapyramidal tracts, resuming their function.
Co-contraction or co-activation refers to the simultaneous contraction of both agonist and antagonist muscles. It is dysfunctional when it is inappropriate or excessive and impairs the agonist muscle's function, making the agonist appear weaker than it is. In healthy subjects, the activity of antagonistic muscles is controlled by central modulation of the transmission in the inhibitory pathways that link the muscles. It is a likely possibility that a deficient control of these inhibitory mechanisms is, at least partly, the basis of the inappropriate antagonistic co-contraction in spastic patients. The pathophysiological substrate of co-contraction is impairment of Ia reciprocal inhibition in the spinal cord., The reduction of co-contraction in the pronators during supination was statistically significant, including in those whose MMAS scores remained the same postoperatively (P < 0.05). The subjective improvement as felt by the patients was also considered as an important factor in the analysis, and it was observed that the more improvement the subjects felt, the better was the change in co-contraction seen in EMG. Despite all the recruited patients stating that they felt a subjective improvement in hand function, this was reflected as an improvement in the MMAS scores of only 15 patients. On the other hand, all these patients showed an improvement in the EMG criteria implying that the latter was more sensitive in detecting the subtle changes. Our own findings from a previous publication  indicate that patients report an immediate improvement in upper and lower limb motor function after surgery. They describe this subjective improvement as a "decrease in tightness" or a feeling that the limbs are "free." This is not always accompanied by an objective improvement in muscle strength as assessed by routine bedside clinical examination. It confirmed that there was improvement in hand function as assessed by rapid opening and closing of the hand and it was speculated that this was a result of an improvement in spasticity. In the current study, our findings seem to corroborate this hypothesis and are substantiated further by the fact that the mean (upper limb JOA) ulJOA scores improved in the cohort. However, further research is required to correlate EMG testing with clinical outcomes as assessed by other tests such as the Jebsen-Taylor hand function test or the nine-hole peg test.
We faced intrinsic difficulties in measuring EMG signals such as ensuring that the placement of electrodes and skin impedance were identical in all patients. We minimized error by adhering to the SENIAM recommendations on sensor locations. A limitation of this study is that inter- and intra-rater reliability of this method was not tested, which needs to be explored in future publications. Furthermore, it would have been ideal if the surface EMG activity of the lower limbs was also recorded. However, we found that the construction of a similar device for the lower limbs was more cumbersome and, hence, restricted the study to the upper limbs.
We found that, by using EMG, we were able to detect a significant decrease in both baseline electrical activity as well as in co-contraction of the pronators and supinators of the forearm in patients with cervical compressive myelopathy following decompressive surgery. Furthermore, while the MMAS and mJOA scoring system are good clinical measures of spasticity and function, EMG is helpful in the early detection of improvement in spasticity in patients with no clinical improvement in the immediate postoperative period though it remains uncertain whether such patients have a better long-term prognosis. We believe the introduction of a patient and user-friendly portable device to perform EMG testing at the bedside will greatly augment monitoring of the rehabilitation process in these patients.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3], [Figure 4]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]