In this study, we have found that a task testing the limits of stability in SCI individuals who are community ambulators, is mainly characterized by an increase in the distance travelled by the COP when going to its maximal position rather than by a limitation in the absolute distance reached by the COP in the intended direction. Therefore, using COPlength may better characterize performance of individuals with SCI. When the COPmax is greater in individuals with SCI as compared to able-bodied individuals in some direction, it is associated with the COPlength, which further support the use of this measure in a SCI population. To our knowledge, this is the first study to report such a finding in this population. In accordance with our second hypothesis, we also found little association between the quasi-static and the dynamic balance parameters.
Performance of individuals with SCI versus able-bodied individuals on the comfortable limits of stability test
COPlength is calculated from the total excursion of the COP on its way toward its maximal position in an indicated direction. COPlength of individuals with SCI was greater than that of able-bodied individuals in all directions. This increased COP excursion could be interpreted as a decrease in movement precision: whereas COP movement of able-bodied individuals tended to move in a straighter manner, individuals with SCI displayed more extraneous movements on their way to their maximal position, increasing the COPlength. Similar results have been found in other populations presenting neurological deficits such as Parkinson’s disease. Ondo et al. found that individuals with Parkinson’s disease had a significantly increased path sway on the Smart Balance Master limits of stability test as compared to able-bodied individuals, a test comparable to the one performed in our study .
Many factors may be responsible for this decrease in movement precision during the comfortable multidirectional limits of stability test in individuals with SCI. Since a SCI generally alters the integrity of the various sensory tracts travelling within the spinal cord, the somatosensory contribution to postural balance may therefore be decreased, which in turn, could explain the less precise movement of the COP. Clinical evaluation of sensory function revealed residual sensory deficits in most of our participants with SCI. In fact, a varying degree of foot anesthesia induced from various techniques causes a proportional increase in COP motion while standing [26, 27]. Moreover, an increase in visual contribution to postural balance in people with SCI as compared to able-bodied controls does exist . This may represent a compensatory mechanism for a decreased sensory perception in the lower extremity. Future studies could identify to what extent sensory deficits following SCI is a major contributor to observed standing postural balance deficits.
Lower-extremity muscle function is another factor influencing control of standing postural balance in various populations [28–31]. Although our participants with SCI had sufficient lower extremity strength to assume a standing position, a certain deficit in strength was present in the ankle dorsiflexors and plantarflexors, as indicated by the LEMS evaluation. It is known that the location of lower-extremity muscle strength influences balance capability differently. For example, Horlings et al. demonstrated that distal muscle weakness more significantly influences postural stability than proximal muscle weakness . Future studies using a dynamometric evaluation of lower extremities could help identify those muscle groups associated with increased COPlength in each direction.
Contrary to our initial hypothesis, individuals with SCI had similar COPmax compared to able-bodied controls as indicated by the lack of significant differences in all but one direction and on the overall measure given by the COParea. Interestingly, these findings correspond to those of Gauthier et al. who showed that SCI individuals who had partial or full control over their abdominal and lower trunk muscles could bring their COP to a similar distance from their base of support to that of able-bodied individuals when performing a similar postural balance test while sitting . This may seem surprising considering that standing is inherently less stable than sitting [19, 32] and could thus lead to greater differences in performance between impaired and normally functioning individuals. However in our study, individuals with SCI had adequate motor recovery in their trunk muscles to be able to assume a standing position and lean in various directions. This may in part explain the lack of difference seen in COPmax. Yet the significant groups vs. directions interaction indicates that individuals with SCI and able-bodied participants differed on how they performed in various directions. More specifically, individuals with SCI had greater COPmax in lateral directions while able-bodied individuals had greater COPmax in anterior and antero-lateral directions. Complementary studies including a larger sample of participants could help to confirm whether or not a difference in COPmax between the groups exists. Greater COPmax in some directions cannot be attributed to different foot placements since foot placements were standardized using a template. This is also supported by the absence of significant difference between the groups on the COPmax in a posterior, right and left directions.
Ankle plantarflexor muscle groups are known to influence anteroposterior COP excursions  and are especially activated when controlling anterior body displacement with respect to the base of support . Since most of our individuals with SCI had residual distal lower-extremity weakness, we could hypothesize that the lower COPmax in the anterior and anterolateral directions may be explained by this lack of strength. Individuals with SCI could therefore limit anterior COP displacement in order to take into account their decreased ability to control the COP using their ankle plantarflexors when reaching the limits of stability in this direction.
On the other hand, individuals with SCI had greater COPmax in the lateral directions than able-bodied individuals. Body displacements in lateral directions are under the control of hip abductor muscle groups , which did not achieve full recovery in our group of individuals with SCI. A possible explanation for this better performance is that those individuals with SCI who outperformed able-bodied individuals in these directions did it at the expense of precision, as indicated by the COPlength. This hypothesis is partly supported by the significant positive correlation found between the COPmax and the COPlength in the posterior, left and left posterolateral directions where individuals with SCI were found to outperform able-bodied individuals. This indicates that, in these directions, individuals with SCI who could displace their COP further were generally less precise than those presenting more limited COP displacements. Since the comfortable multidirectional limits of stability test challenges the postural control system and attentional resources, some individuals may have favored moving less precisely in the displayed direction in order to achieve a maximal performance on the COPmax. Therefore, repeating this test while imposing no constraint on the movement precision may have yielded different COPmax results.
Association between quasi-static and dynamic postural balance tests
This study also yields little association between dynamic and quasi-static postural balance for individuals with SCI. As mentioned earlier, only four of the 48 possible combinations were found to be statistically significant. This limits the inferences that can be drawn from these associations. These results support a previous study reporting a lack of significant correlation between the static eyes open test and the limits of stability test of the Smart Balance Master, the later sharing some similarities with the comfortable limits of stability test . This is also in line with most studies exploring the association between measures of static and dynamic balance among individuals with stroke or able-bodied individuals [9, 10, 34]. In spite of these remarks, parameters of the COP during quasi-static stance correlated with the comfortable multidirectional limits of stability test in only two directions (left and right anterolateral directions). Further research including more participants with a wider range of balance deficits would be necessary to determine the significance of this specific result.
Our study suffers from a few limitations needing consideration when interpreting these results. Firstly, all our participants with SCI were community ambulators. Many participants with SCI had near normal walking ability, as indicated by the mean natural speed of 1.02 m/s, which is close to the 1.06 m/s value required to be considered as a safe community ambulator . It is thus possible that our group of SCI participants was not representative of an actual population of individuals with traumatic incomplete SCI. These results are therefore not applicable to those individuals who are starting to assume a standing position. This could have limited the possibility to find differences between our two groups. However, although statistical power of the study was in part limited by the small number of participants in each group, effect sizes were at least moderate for the COPlength in the comfortable multidirectional limits of stability test. Therefore, this statistic supports the fact that our sample size was sufficient to find differences between the groups.
As is often done in other studies, COP-based measures were not normalized using the dimension of the base of support or foot length [17, 18]. However, our two groups did not differ in height. Since foot position was standardized and monitored during the study, we therefore presume that this normalization would not have changed our main conclusions. Although participants were told to initiate the movement from the ankle instead of bending their trunk or their hips when leaning, some degree of trunk and hip compensation of varying degrees was present among participants and directions. We thus suspect that the actual performance may be partially related to the strategies used. Adding a kinematic analysis to our protocol may have helped to identify biomechanical markers associated with the difference in performance seen among both groups.
It is possible that the performance of individuals with SCI is potentiated because of the visual feedback provided during this test. In fact, Sayenko et al. have shown that visual feedback can improve standing balance performance in individuals with incomplete SCI . A test performed without visual feedback could have been more reflective of balance capabilities of each participant and could have generated more differences between the groups. Normal dynamic postural balance activities such as walking occur without on-screen visual feedback on actual performance. Therefore, how the results from this study can be generalized to other dynamic balance activities remains to be explored.
Although 15 seconds were given to maximally displace the COP in the indicated direction and come back to the initial position, no actual control on the speed of movement was given. However no difference was seen between the groups in the mean COP speed. Thus this factor could not explain the differences in performance (i.e., COPlength) seen between the groups. We did not analyze the return from the maximal position to the initial position, which may have provided further insight into dynamic balance performance in individuals with SCI in the comfortable limits of stability test. Reliability and minimal detectable change of the COPmax and the COPlength were not assessed. Since the standard error of measurement is unknown for these parameters, this limits the inferences that can be drawn from differences in performance seen between the groups of participants. Lastly, we did not apply a correction for multiple comparisons (e.g., Bonferonni’s) to the correlational analysis. For this reason, the correlation that were significant must be interpreted with caution since a possibility of a type I error exists.