Inadequate knowledge in the recruitment patterns of abdominal muscles in individuals with spastic-type cerebral palsy (STCP).
To determine whether there is any difference between the neuromuscular activity (activation pattern) of the abdominal muscles in children with STCP and those of their typically developing (TD) peers.
The NORAXAN® electromyography (EMG) was used to monitor the neuromuscular activity in abdominal muscles of individuals with STCP (
EMG frequencies were recorded during rest and during active states and compared using repeated measures ANOVA. Spearman’s rank order correlation was used to explore relationships between age, body mass index and abdominal muscle activity. With the exception of the rectus abdominis (RA) muscle, the pattern of neuromuscular activity in children with STCP differs significantly from that of their TD peers. Three of the muscles – external oblique abdominis (EO), internal oblique abdominis (IO) and RA – in both groups showed significant changes (
The findings from this study suggest that the RA could be targeted during rehabilitation regimens; however, the force generated by this muscle may not be sufficient for the maintenance of trunk stability without optimal support from the EO and IO muscles.
In order to maintain the necessary equilibrium of stability and mobility of the spine, recruitment of trunk muscles has to be adequately organised and coordinated (Tucker & Hodges
In children with cerebral palsy (CP), hypertonia and muscle weakness are reported to be the common impairments contributing in varying degrees to motor dysfunction seen in these children (Rosenbaum et al.
There are, however, limited studies that aim to describe recruitment strategies of the trunk muscles in CP, but these are predominantly inferred from observational analysis (Tucker & Hodges
Our study, therefore, aimed to investigate and compare the neuromuscular activity of the anterior trunk muscles (trunk stabilisers) in children with STCP with their age-matched TD peers. Measurements at rest and during activity were recorded using surface electromyography (sEMG). We hypothesised that selected EMG measurements would reflect the degree of contractile response in the abdominal muscles in individuals with STCP and that this would differ from that recorded in children with TD. Relationships between age and body mass index (BMI) and abdominal muscle activity were also explored.
A descriptive correlational study design was used in that the EMG recordings of abdominal muscle activity in children with STCP and those with age-matched TD peers were compared. Ethical approval was obtained from the Human Research Ethics Committee of the University of Cape Town (HREC REF: 490/2011). Permission to recruit the learners for the study was obtained from The Western Cape Department of Education for the participating schools in the Cape Town Metropole (REF: 2011202-0028). Recruitment was done in three schools for children with special education needs and two randomly selected public mainstream schools located within the vicinity of the special schools. All children who complied with the inclusion criteria and whose parents or legal guardian gave written informed consent were invited to participate.
A final group of 63 children aged between 7 and 16 years with STCP was compared to a group of 82 TD individuals. The two groups were not matched for any demographic or anthropometric variables. Differences were explored and, where appropriate, factored into the statistical analysis. For both groups, any child was excluded if he or she had any surgical procedure involving the anterior abdominal wall in the last 6 months before the start of the study. Children with STCP had to be ambulant (i.e. level I-IV according the GMFCS [Palisano et al.
The NORAXON® sEMG was used in this study as EMG is considered a more objective method for describing muscle activation patterns. sEMG has been shown to be a valid and reliable method for detecting muscle activity in CP, and the limitations on the use of surface electrodes have been well described by Hlustik et al. (
Each recording channel on the equipment is designed to detect activity from one muscle site and is composed of two active electrodes and a reference one. Surface EMG uses disc-coated silver-silver chloride electrodes that can detect algebraic sums of voltages associated with muscle action potential within their pick zones. The electrodes are about 0.5–1.0 cm in their wider diameter. The difference in electrical charges between each active electrode and the reference one provides input to a differential amplifier with impedance of a common mode rejection ratio (CMMR) between 90 and 140 dB. There are filters as part of the internal design of this NORAXON® sEMG equipment, which allow frequencies related to a muscle activity but reject those frequencies that are associated with electromagnetic noise. According to the literature, the evaluation of EMG by means of amplitude is influenced by many factors: anthropometrics, precise electrode placement and overlying tissues in the body (Moreau, Holthaus & Marlow
The study aimed to record EMG at rest and during active state. Limb movement is reported to be accompanied by a concomitant contraction of the abdominal muscles (Vasseljen et al.
Prior to the collection of the data, participants were trained to reach paced activities within 1 min with regard to tasks to be accomplished during active or contracted phases (head up with chin tucked in and lower limb movements). For the latter, the more affected limb was identified by a neurodevelopmental therapist for the participant in the STCP group. In the supine position on a plinth with the arms resting along the sides of the body, the children were asked to flex the hip and knee and bend up as far as possible. For the head and neck movements, participants were instructed to tuck in their chins and then lift their heads to their chests. All of these movements were performed under the supervision of their physiotherapist.
Testing was officially initiated (usually a day or two after training) when participants appeared to be confident and consistent with executing the tasks. All the silver-silver chloride electrodes came prepared with a creamed electrode gel that makes them self-adhesive. However, electrode placements were reinforced with adhesive sellotape. The electrode sites were gently abraded with fine grain sand paper and cleaned with isopropyl alcohol. The placement of the electrodes was in accordance with standard procedures, (Hermens et al.
For the RA, the active electrodes were placed 3 cm above the umbilicus and 3 cm away from the median plane always on the right side but if unilaterally affected as in the case of right hemiplegia, then placement of this electrode was done on the left side. The ground electrode was placed at the level of the 12th thoracic spinous process (T12) for those participants with spastic diplegia and quadriplegia, whereas this electrode was placed on the patella of the non-affected limb for participants from the hemiplegic and TD groups. For the EO EMG readings, the active electrodes were placed midway between anterior superior iliac spine (ASIS) and the lowest point on the subcostal angle in the mid-axillary line always on the right side, but if unilaterally affected, then on the left. Alternatively, the more lateral electrode was placed at the most lateral point in the mid-axillary line on the trans-umbilical plane on the affected side. The EMG recordings for IO were taken with active electrodes at a point midway between the ASIS and pubic tubercle, about 2 cm above the inguinal ligament in the mid-clavicular line always on the right side, but if unilaterally affected, then on the left. In the case of children from the TD groups, the active electrodes were placed at these anatomical sites as described for the above but always on the right side.
The sEMG were sampled at 1000 Hz and saved on a personal computer for further analysis with a custom made DELL/NORAXAM programme. Based on a trial test, patterns of sEMG activity in the participants were found to be variable across 10 trials, and therefore, the tasks were limited to a maximum of 5 after which the mean scores of 3 consistent patterns were recorded as the sEMG score for a particular muscle. The raw sEMG signals were first amplified 300 times by pre-amplified electrodes by default and then 4 times with the computer analysis. The signals were then filtered with 10–1000 Hz band per filter. All data were then filtered with a second low-pass filter at 16 Hz. A muscle onset of activity was defined as the point when sEMG recording exceeded the baseline by two standard deviations for greater than 25 ms and the software marks this point as the EMG traces. These were then visually inspected by the principal investigator (PI) and the research assistant independently ensured that subsequent outbursts or values/traces obtained could be compared to determine inter-tester (between PI and research assistant) and intra-tester reliability (within day activity of PI) (see
Comparison of demographic data between the two groups.
Variables | Mean STCP | SD STCP | Mean TD | SD TD | |
---|---|---|---|---|---|
Age (years) | 11.9 | 2.92 | 11.0 | 3.04 | 0.087 |
Height (cm) | 139.2 | 16.04 | 143.3 | 17.13 | 0.213 |
Weight (kg) | 39.7 | 10.28 | 38.7 | 12.38 | 0.457 |
BMI (kgm−2) | 20.1 | 2.16 | 18.4 | 2.62 | < 0.001 |
SD, standard deviation; TD, typically developing; STCP, spastic-type cerebral palsy; BMI, body mass index.
Descriptive statistics for electromyograph measurement (raw scores) during resting and active stages for both groups of participants (STCP:
Diagnosis | Muscle type and state | Mean EMG (Hz) | SD | Diff | SD of Diff | 95% CI of Diff | |
---|---|---|---|---|---|---|---|
STCP | EO R | 85.0 | 2.3 | 0.001 | |||
EO Ac | 123.1 | 2.5 | 38.1 | 2.1 | 37.6–38.6 | ||
TD | EO R | 11.8 | 1.6 | ||||
EO Ac | 107.3 | 1.8 | 95.5 | 1.5 | 95.1–98.0 | ||
STCP | IO R | 89.9 | 4.0 | 0.001 | |||
IO Ac | 127.5 | 3.8 | 37.7 | 2.8 | 37.0–38.4 | ||
TD | IO R | 11.6 | 1.6 | ||||
IO Ac | 110.8 | 2.0 | 99.2 | 1.7 | 98.6–99.8 | ||
STCP | RA R | 12.2 | 1.4 | 0.074 | |||
RA Ac | 97.3 | 1.8 | 85.1 | 1.2 | 84.8–85.4 | ||
TD | RA R | 11.4 | 1.4 | ||||
RA Ac | 97.0 | 1.8 | 85.6 | 1.3 | 85.3–88.3 |
EMG, electromyography; R, resting state; Ac, active state; SD, standard deviation; Diff, differences in EMG scores between active and resting states; CI, confidence interval; STCP, spastic-type cerebral palsy; TD, typically developing; RA, rectus abdominis; EO, external oblique abdominis.
Spearman’s correlation between the electromyograph activity of the abdominal muscles during the resting and active stages as well as between age and body mass index of participants.
Variables | STCP – 63 |
TD – 82 |
||
---|---|---|---|---|
Spearman | Spearman | |||
Resting stage – Age versus EMG | ||||
EO R EMG (Hz) | −0.71 | 0.001 | −0.84 | 0.001 |
IO R EMG (Hz) | −0.73 | 0.001 | −0.88 | 0.001 |
RA R EMG (Hz) | −0.61 | 0.001 | −0.89 | 0.001 |
Active stage – Age versus EMG | ||||
EO Ac EMG (Hz) | −0.26 | 0.036 | −0.76 | 0.001 |
IO Ac EMG (Hz) | −0.32 | 0.012 | −0.74 | 0.001 |
RA Ac EMG (Hz) | −0.64 | 0.001 | −0.81 | 0.001 |
Resting stage – BMI versus EMG | ||||
BMI and EO R EMG (Hz) | −0.04 | 0.764 | −0.36 | 0.001 |
BMI and IO R EMG (Hz) | −0.06 | 0.653 | −0.36 | 0.001 |
BMI and RA R EMG (Hz) | 0.00 | 0.974 | −0.31 | 0.004 |
Active stage – BMI versus EMG | ||||
BMI and EO Ac EMG (Hz) | −0.11 | 0.400 | −0.27 | 0.015 |
BMI and IO Ac EMG (Hz) | −0.03 | 0.793 | −0.34 | 0.002 |
BMI and RA Ac EMG (Hz) | −0.07 | 0.578 | −0.39 | 0.001 |
R, resting state; Ac, active state; STCP, spastic-type cerebral palsy; TD, typically developing; RA, rectus abdominis; EO, external oblique abdominis; EMG, electromyography; Hz, Hertz; BMI, body mass index.
STATISTICA software package, version 11 (2012), was used to analyse the data. Descriptive statistics were presented. The Shapiro-Wilk test indicated that most data sets were normally distributed. Spearman’s rank order correlation was calculated between age, BMI and muscle activity. Although age was shown to consistently predict EMG scores, results using data normalised for age did not differ from analysis using raw data and are therefore not presented as such in this article. Repeated measures ANOVA were used to compare resting and active state sEMG scores both within and between groups. The level of significance for all statistical tests was set at 0.05.
The results shown in
During the active state, the sEMG scores were again higher in the STCP group compared to the TD group for the IO and EO muscles. Again no difference was seen for the RA muscle between the two groups.
For all three muscle groups, there was a significant increase in EMG activity from resting to active states with
Relationships between age, BMI and muscle activity were explored and are shown in
Graphical representation of changes in rectus abdominis thickness from resting to active state against categorised ages in years for both groups (
This study has shown that similarities and differences exist in the EMG activity recorded for the abdominal muscles in children with STCP and their age-matched TD peers. Although the RA of the children with STCP showed similar patterns of recruitment to those of TD children, activity in the other two abdominal muscles recorded in this study differed significantly both at rest and during contraction. Despite the demographic data showing that the individuals in the STCP group were heavier than their age-matched TD counterparts (refer to
The sEMG scores recorded during the resting state were above absolute zero for both groups of children. With respect to the individuals with STCP, our results showed that these scores were significantly elevated, suggestive of hypertonia (Rosenbaum et al.
This study appears to suggest that during the active stage, EO and IO muscles in children with STCP showed an inability to synchronise the rate of activation of all the already activated motor units. The small difference found between the two groups in terms of changes in EMG activity during the resting and active stages is most likely because of the elevated resting stage measurements seen in the STCP group. The changes in EMG activity between the active and resting stages in the IO and EO muscles could be explained on the basis of these muscles being in a state of fatigue prior to the demand for work output (active stage) which would negatively affect their ability to (recruit fibres) contract fully. Our findings therefore are in support of other studies that report the existence of a correlation between over-contraction and fatigue in muscles and low or poor EMG activity (Dario, Merletti & Enoka
Some other studies have shown that the magnitude of the differences in neuromuscular activity between the resting and active stages, measured in Hertz units (change in EMG frequency), is indicative of the strength of a given muscle (Chapman et al.
The EMG frequency is reported to correspond to recruitment patterns but is not necessarily related to performance levels of a muscle (Gough & Shortland
In contrast to EO and IO abdominal muscles, the RA muscle in the STCP group showed electrical activity typical of TD individuals. It suffices to conclude that if high EMG scores at rest signify spasticity or increased tone or muscle weakness, then neither of these characterised the RA muscle in individuals with STCP. The RA in the STCP group as well as the muscles in the TD group appear to be more relaxed with some resting tone and therefore able to activate sufficient motor units during activity. Low muscle tone during the resting stage indicates healthy muscle activity (Damiano & Moreau
Alternatively, the difference in the amount of change in EMG activity between the oblique and RA muscles in the STCP group with regard to the provision of trunk stability may possibly be explained using the concept of agonistic and antagonistic behaviours of the skeletal muscles. It is possible that, with the alteration of the neuromuscular system as associated with STCP, co-contraction takes place between the RA and the remaining abdominal muscles, similar to muscles acting as agonists and antagonists (Damiano et al.
The TrA muscle is also considered as an important trunk muscle alongside the IO and, therefore, the electrical activity from the TrA muscle would have been useful in establishing its role with respect to trunk stability and maintenance of posture. However, owing to the use of sEMG in this study, any inference of TrA has been excluded. Participants could not be restricted from physical activity prior to testing sessions which may have influenced the resting state sEMG measurements.
The EO and IO muscles in the STCP group have higher sEMG scores at rest than their TD peers, suggesting that these muscles are actively contracting even at rest. The RA muscle in individuals with STCP recorded sEMG scores comparable to those of TD individuals. The sEMG results for the RA muscle in individuals with STCP suggest that the RA muscle in individuals with STCP can contract optimally, therefore appear unaffected or least affected by the condition of STCP. With the exception of the RA muscle, the neuromuscular activities (recruitment patterns) of individuals with STCP differ significantly from those of TD individuals.
There is a need for further research as to the underlying causes of the increased tone in the EO and IO muscles at rest as reflected in the elevated sEMG scores (above the zero score mark), especially in the STCP group: whether neurological damage or simply a physiological adaptation was beyond this study and therefore requires further investigation, especially why the RA muscle appears to be unaffected. This study has implication for clinical practice. In terms of therapy, it would appear necessary to attempt to reduce the tone of the EO and IO muscles before attempting to recruit and strengthen these oblique muscles during functional activities. This may be either direct or in cases where the over recruitment is a compensatory strategy, balance between agonists and antagonists should be sought in order to reduce over reliance on any one group for stability. It is also necessary to attempt to selectively train and strengthen the EO and IO muscles to reduce the over reliance on the RA muscle resulting in a rigid trunk. It may be worth looking into the mobility of the pelvis and how this influences recruitment of the RA muscle.
The authors thank the staff and learners of all the schools from which participants were recruited for this study. They are grateful to the technical staff of the Division of Clinical Anatomy & Biological Anthropology of the Department of Human Biology, Faculty of Health Sciences, University of Cape Town for the transportation of the equipment to and from the data sampling sites. Finally, they acknowledge the financial inputs of both faculty and management of the postgraduate units of the University of Cape Town, Faculty of Health Sciences and the University of Ghana Medical School.
The authors declare that they have no financial or personal relationships that may have inappropriately influenced them in writing this article.
The conception of the research idea was J.J., a physiotherapist. Accessibility to participants and equipment was overseen by M.U., also a physiotherapist. Recruitment of the participants and collection of data were carried out by S.K.A., anatomist. Data analysis was done by J.J. and S.K.A. G.L., anatomist and medical educationist, was in charge of supervision, drafting and editing of the manuscript. All four authors shared equal responsibilities in the final preparation of the manuscript.