The current study aimed to characterise the spatial distribution of activation from the thoracic part of the ES in individuals with an incomplete SCI during postural tasks and multidirectional reaching tasks. Our findings demonstrate changes in the spatial distribution of thoracic ES activity during postural and functional reaching tasks in individuals with SCI, compared with age-matched controls. Overall, individuals with SCI showed lower activity of the ES and a more caudally distributed activation during the tasks, compared with the controls. Interestingly, EMG amplitude of ES was the same between reaching forward and returning to the upright posture in individuals with SCI, whilst the EMG amplitude was greater in reaching forward compared to returning to the upright posture in the controls, Additionally, the changes in the y-axis of the centroid within the forward reaching task were less in the individuals with SCI than in the controls. Furthermore, slower reaction time, delayed APAs and shorter reaching distances were observed in individuals with SCI compared with the controls. Taken together, our results suggest altered regional contributions of the thoracic ES to postural adjustments and movements of the trunk after SCI, possibly due to impaired neural drive from the CNS to the ES, affecting the selective activation of the muscles.
Altered spatial distribution of ES activity for postural control after SCI
Research using indwelling EMG electrodes has shown distinct muscle activation between the caudal and cranial regions of the lumbar ES during trunk rotation in healthy adults, suggesting selective control of regions of the ES (29). Studies in clinical populations (e.g., low back pain) have also reported differences in regional activation of the lumbar ES muscles from the healthy controls (17, 19, 30, 31). Here, for the first time, we examined the spatial distribution of the thoracic ES activity in individuals with chronic, incomplete SCI during postural tasks where the role of the ES is to maintain an upright posture. We observed delayed activation of the ES with respect to the AD (i.e., delayed APAs) in individuals with SCI, with the activation being lower and more caudally distributed, compared with the controls. APA is a feedforward control mechanism and is thought to be mediated by the M1 as it occurs in a time window (-100 to + 50ms with respect to initiation of a movement) (32) either prior to a movement or too early to be influenced by afferent input from the periphery (32–34). Previous studies reported an association between corticospinal excitability and timing of APAs in adults with SCI (35) and in older adults (36), suggesting a relationship between corticospinal function and postural control. With this in mind, the changes in regional activation of the ES during APAs observed in individuals with SCI may reflect the impairment of corticospinal function following SCI due to the damage to the spinal cord.
During the external predicted perturbation task, individuals with SCI again activated the caudal part of the ES while the controls recruited the cranial part of the ES when responding to a perturbation to the arm. No difference was found in the EMG amplitudes of the ES or the CoP displacement during APA and CPA between the groups. Again, this may indicate a compensatory postural control strategy by changing regional activation of the ES to maximise postural stability after SCI. Furthermore, we observed a correlation in the CoP trajectory between the time window of APA and CPA in the controls but not in individuals with SCI. In other words, individuals with SCI lost the ability to use the interplay between APA and CPA for postural control (22, 37). APA and CPA are two main mechanisms that the CNS uses to maintain postural stability in humans (33, 38–42). Our findings suggest alterations in the CNS mediating the mechanisms for maintaining postural equilibrium after SCI, as shown in the changes in regional activation of the ES muscle and the lack of the correlation between CoP displacement between the APAs and CPAs windows.
Altered spatial distribution of the ES activity for assisting with movements of the upper extremities after SCI
Different from the postural tasks, the ES is activated to assist with movements made with the upper extremities during forward and lateral reaching (3, 43, 44). Due to impaired control of the trunk, people with SCI show less reaching distance and CoP displacement compared to able-bodied controls (3, 23, 44). This is confirmed in our study; reaching distance was shorter, displacement of the trunk and CoP was less in individuals with SCI than in the controls in both forward and lateral reaching tasks. Interestingly, when comparing activity of the ES between the reaching and returning phases of the forward reaching test, the controls showed greater ES activity during the reaching phase than the returning phase, whereas individuals with SCI showed no difference in ES activity between the two phases, suggesting reduced differential motor control during these movements. Moreover, activity of the ES was limited to a smaller region of the ES within the forward reaching movement in participants with SCI, while it was spread across a larger area of the muscle in the controls, suggesting a localised recruitment of the ES in those with SCI. It is thought that changes in EMG amplitude within the muscle may reflect regional function of the muscle, possibly due to selective motor control from the CNS (45–47). Our results showing RMS amplitudes of the ES correlating with the trajectory of CoP during lateral reaching in participants with SCI but not in the controls support the notion of altered motor control strategies of the ES after SCI and is in keeping with previous work (35). It should be noted that a shift in the regional activation towards the caudal region of the thoracic ES could be a result of a non-specific (generalised) activation over the entire electrode grid that may be more related to muscle weakness of the ES after SCI. Nevertheless, this possibility was controlled by our analysis approach which only considered the electrodes with an RMS amplitude higher than 70% of the maximum RMS amplitude of the entire electrode grid. Given the small activity obtained in the impaired trunk muscles in individuals with SCI, this approach improved the signal-to-ratio, allowing for better detection of differences in the recruitment patter of the ES between individuals with SCI and the controls. Furthermore, the differences in regional activation could also be influenced by architecture of the muscle (e.g., direction of the muscle fibres, innervation zone) and changes in the inter-electrode distance during the movement (48). The distinct distribution of the EMG response in the ES between participants with SCI and the controls in our study could be due to a combination of impaired neural drive from the CNS to the muscles below the injury (35, 49, 50) and changes in the properties of the ES muscles, such as muscle atrophy (51, 52), denervation (53), and changes in signalling transmissions around the neuromuscular junctions (54), that have been reported to occur following SCI.
Functional consequences
Surface bipolar EMG has been extensively used to study trunk muscle control, and muscle synergies of upper-body and back muscles in individuals with chronic SCI (7–9, 35, 55, 56). Emerging evidence has shown the CNS selectively controlling different regions of the ES and the non-uniformity of activation of the ES across spinal segments during trunk movements in healthy adults (29, 57, 58). Hence, it is relevant to assess injury-related changes in the spatial activation of the ES muscles in SCI. HDEMG can provide detailed information on activation patterns of the ES that bipolar EMG is unable to provide. Indeed, we show for the first time that individuals with SCI recruited a different part of the thoracic ES, in comparison to the controls, to maintain postural stability and to assist movements of the upper extremities. The differences in the recruitment patterns between individuals with SCI and the controls were still present even when the kinematic results were similar between groups (i.e., no differences in the CoP displacement between the groups during the RSF and external perturbation tasks). This suggests altered motor control strategies used by individuals with SCI in the postural tasks, as shown in previous studies in wheelchair propulsion (10, 11) and during pivot transfers (12), likely to maximise function and performance. Conversely, in the functional reaching tasks, individuals with SCI reached less and showed different activation patterns compared to their counterparts, indicating muscle impairment and reduced function and performance. These results suggest that the compensatory motor control strategies used by people with SCI may not be sufficient for the reaching tasks, highlighting a potential relevance of task-oriented training for trunk rehabilitation after SCI.
Prior work has applied the spatial-temporal information extracted from HDEMG in automatic recognition of motor tasks of the forearm in people with incomplete SCI (59), with a view to developing assistive technologies for upper-limb function. Research has also shown improved diagnosis of paralysis of hand muscles in individuals with SCI using HDEMG with advanced signal processing techniques (60). Furthermore, surface EMG is a common tool for clinical evaluation of therapeutic outcomes of trunk rehabilitation (61, 62). Given that trunk control plays an important role in activities of daily living and many people living with SCI have impaired trunk control, our findings provide new insights to motor control of the ES muscles in SCI, but also may be relevant to development of assistive technology for postural control, assessment of trunk muscles, and evaluation of therapies for trunk function.
Limitations
Changes in muscle activity measured by surface EMG have been commonly used as an objective measurement for therapy-induced restoration of function after SCI (63–66). Our participants with SCI reached less than the healthy controls and showed less trunk displacement during the reaching tasks. It could be that the different performance between groups played a part in the spatial changes in activity of the ES. Nevertheless, it remains unclear how closely muscle activity relates to function and performance in clinical populations as studies showing a correlation between force and muscle activity employed isometric sustained muscle contractions as an assessment paradigm (59, 67). Moreover, most people with SCI who lack trunk control utilise a posterior pelvic tilt that considerably decreases the thoraco-lumbar curvature to increase the base of support (4, 68). This posture lengthens the ES and may influence spatial distribution of ES activity (69), as seen in our participants with SCI. Note that results here were presented without EMG normalisation, since our previous study reported that with the same technique and the methodology uncorrected EMG amplitudes seem to be more reliable than normalised data (20). Additionally, there were more males in the SCI group than the control group, in-keeping with the prevalence of SCI being higher in males than in females (70). The effect of sex on EMG measurements has been reported during high levels of sustained voluntary contractions (71–74). Nevertheless, prior work reported no sex differences in muscle activity during intermittent low levels of voluntary contractions (75) and our unpublished work did not detect sex differences in the HDEMG-derived parameters in the ES during the reaching tasks. While we are unable to completely rule out the effect of sex in our results, the impact of the imbalance of male and female participants between the groups on our results is likely to be minimum. Moreover, as there were only five participants with a thoracic SCI, we did not perform a subgroup analysis comparing spatial activation of the ES between participants with high-thoracic and low-thoracic SCI. Previous studies using bipolar EMG reported different activity of the trunk muscles between individuals with high- and low-level of SCI during static sitting (7, 9) and posterior transfer (76), albeit the information from the amplitude-based analysis with the bipolar EMG may not be applicable to the spatial activation of the muscles extracted from HDEMG.