Thestudy was approved by the local Ethics Committee (Ethics Commission Salzburg/Ethikkommission Land Salzburg; number E-Nr1541) and was designed according to the Declaration of Helsinki. Written informed consent was obtained from all participants prior to the assessments.

Within a study conducted at the Department of Neurology, Paracelsus Medical University Salzburg, Austria, we asked seven patients with cervical or thoracic spinal cord injuries to fill in the Movement Imagery Questionnaire Revised (MIQ-RS; Gregg et al., 2010). As depression is a common comorbidity in patients with SCI we assessed also depressive tendencies using the Beck Depression Inventory II (BDI-II) as a control variable (Hautzinger et al., 2006). The BDI-II questionnaire was completed by all patients at the beginning of the study. The MIQ-RS was given to the patients and they were asked to complete it at a quiet moment between additional experimental sessions that were not part of the presented analysis. These comprised magnetic resonance imaging and electroencephalography, as well as repetitive transcranial magnetic stimulation in some cases.

We assessed 7 men aged between 23 and 70 years (Md = 53) with traumatic SCI. The extent of motor and sensory impairment was classified using the International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI; Kirshblum et al., 2011). A detailed description of the included patient group is given in Table 2. In addition, we asked 20 healthy participants to take part in an online questionnaire containing both the MIQ-RS and the BDI-II. All of these healthy controls were also men and aged between 21 and 54 years (Md = 30). None of them reported any motor or sensory impairments.

The MIQ-RS contains two scales of seven items each, assessing the ability to imagine how the execution of a specific movement would feel like (kinesthetic imagery scale, KIS) and the ability to imagine how a certain movement execution would look like (visual imagery score, VIS; Gregg et al., 2010). Each item consists of a statement describing a particular everyday movement (i.e., pulling a door handle, or grasping a glass). The participants are asked to perform the movement once and afterwards return to their starting positions. Then they are asked to visually or kinesthetically imagine the movement, after which each item is to be rated according to a 7-point Likert scale from 1 = “very hard to see/feel” to 7 = “very easy to see/feel.” Sum scores for both scales range between 7 and 49 with low values indicating increased difficulties to either visually or kinesthetically imagine movements. Since all participants were German native speakers, the MIQ-RS was independently translated into German by two native speaking members of our team and we constructed the final version using matching translations and in case of differences by consensus decision of the study team.

We used the German version of the BDI-II to assess the severity of depressive symptoms (Hautzinger et al., 2006). The BDI-II contains 21-items, each of which is formulated as a question asking the participant to indicate the occurrence of certain moods or types of behavior within the last weeks (i.e., how often one cries, whether one has difficulties concentrating). Each item gives a number of possible answers that are coded from 0 to 3, leading to a sum score between 0 and 63. A higher score indicates more severe depressive symptoms or a stronger tendency toward depression. It has been shown that the BDI reliably assesses depressive tendencies in patients with SCI (Judd et al., 1991) and it is still widely used in clinical and scientific contexts (Kennedy et al., 2016; Barbonetti et al., 2017).

We performed two Wilcoxon-Mann-Whitney tests for non-parametric independent samples to compare the two variables of interest, the VIS-score and the KIS-score, between patients and healthy subjects. Two additional tests were calculated to compare the variables age and BDI-II scores between the two groups. Additionally, we calculated the correlations using Kendall's Tau between VIS and KIS scores and the two variables BDI-II scores and age. We used exact test statistics and corrected the p-values for multiple comparisons using the Holm-Bonferroni method (Holm, 1979). The resulting p-values, being already corrected, were then interpreted using a threshold of p = 0.05. For statistical analysis we used the R-Environment (R Version 3.4.1; R Core Team, 2017), the package “Kendall” (McLeod, 2011), and a slightly updated version of the package “rankFD” (Konietschke et al., 2016).

The main limitation is the small patient group assessed in this study. This is a known problem when it comes to studies in this neurological population. This can also be seen in Table 1, revealing only small sample sizes throughout most investigations, stressing the need for larger studies involving samples from multiple centers. The MIQ-RS not being standardized and designed for this particular patient group is a second limitation. Though, it has been validated in other neurological conditions, such as stroke (Gregg et al., 2010), there are no prior findings regarding validity or reliability in patients with SCI. The fact that subjects are asked to perform a certain movement before imagining it might introduce a confounding error, as patients with SCI are not able to perform all movements. However, there is no neuropsychological assessment of MI ability, that has been validated in SCI patients so far. Larger samples are necessary to assess the reliability of such questionnaires in this specific neurological subgroup. Regarding the MIQ results, one could also argue for an introduced error as healthy controls participated in an online assessment whereas patients were assessed using the pen and paper questionnaire. We assume that there is no difference between the two assessment forms, as both are subjective and equally prone to manipulation by participants.

In this study we investigated seven individuals with SCI who greatly differed concerning their time since injury. Alterations in brain functionality due to SCI-induced plasticity could impact MI abilities (Di Rienzo et al., 2014a). It has been suggested that plasticity occurs shortly after injury (Brasil-Neto et al., 1993; Aguilar et al., 2010). A longitudinal study found that atrophy in the central motor cortex is established as early as after 2 months (Freund et al., 2013). With regard to our sample we have found several indications that plastic changes had already occurred; in a previous study we documented changed or absent motor potentials obtained using electroencephalography during motor imagery and motor execution in the same sample as in this study (Thomschewski et al., 2017), and we found more diffuse and decreased brain activations during MI seen in functional magnetic resonance imaging scans. On this background, the absence of a visible connection between decreased MI abilities and time since injury in our sample seems plausible. We encourage further research that examines fast changes in MI ability in the acute stage of SCI.

It has been reported that the level of lesion as well as completeness of the injury have an impact on MI abilities (Pfurtscheller et al., 2009; Scandola et al., 2017). With regard to completeness of injury our sample was rather homogenous, as all were incomplete with AIS grades C, or D. In addition, all but one patient had cervical lesions and patient 4 with a thoracic lesion did not yield striking differences in experienced MI difficulties. Thus, albeit a small sample size, our patient group's heterogeneity is not likely to have been a confounding factor. However, given this sample, any conclusions drawn from these results should not be extrapolated to female patients or patients with more severe SCI resulting in complete injuries.

As reported, there is a low correlation between depressive tendencies and increased difficulties to perform MI in the patient group. This is not too surprising, as the MIQ-RS asks for the subjective difficulty of MI and depression itself is characterized by difficulties in managing everyday activities, as reflected by the questions posed in the BDI. Also, a negative impact of depression on MI ability has already been reported previously (Bennabi et al., 2014) and might be of special importance in patients with SCI, as depression is a common comorbidity (Elliott and Frank, 1996; Kraft and Dorstyn, 2015; Williams and Murray, 2015). Taking into account the low and nonsignificant correlations in our whole sample between BDI-II and MIQ-RS scores, this finding should be interpreted with caution. Furthermore, it is well known that depression negatively impacts performance in all kinds of tasks (Cohen et al., 1982; Bennabi et al., 2014; Rock et al., 2014), implying a general effect on task performances rather than a specific effect on MI tasks. Nonetheless, we cannot rule out the possibility that in SCI, the depression is closely linked to movement capacities, and thus also to MI.

Our two groups did differ with respect to their age. Statistical analysis also suggested a mild correlation between age and difficulties in MI, though this relationship was not statistically significant after correction for multiple comparisons. This observed tendency does fit to similar reports of an age-related decline in MI (Saimpont et al., 2009; Malouin et al., 2010; Personnier et al., 2010). However, there are also contradicting findings in young individuals showing that MI capabilities increase at least in young ages (Hoyek et al., 2009). Given previous literature and the difference in age between our two groups, age on the one hand might very well contribute to the differences in MI difficulties between our groups. On the other hand, our findings also suggest that age does not solely explain the differences. Despite age as a possible contributing factor, our results suggest that motor impairments are likely to cause difficulties in imagining movement execution.

In summary, we found that our patients with SCI have more difficulties in imagining movements both visually and kinesthetically. This is likely to be at least partially caused by motor impairments and deafferentation, and should be kept in mind when implementing MI in rehabilitation or BCI-applications in this patient group. However, long-term effects of MI training probably help to overcome these experienced troubles, as has been suggested by Gustin et al. (2008). Furthermore, it is important to note, that this does not mean that the ability to perform MI is impaired, as it has been reported that MI vividness in patients with SCI is comparable to that of healthy subjects (Hotz-Boendermaker et al., 2008; Di Rienzo et al., 2014b,c, 2015; Fusco et al., 2016; Roosink et al., 2016). Additionally, we investigated visual and kinesthetic MI and it is possible that MI strategies may change after SCI (Fiori et al., 2014). Depressive tendencies and age might represent further contributing factors of increased difficulties in performing MI. In order to gain more insights into these issues, we suggest to assess subjectively perceived capabilities to perform MI in larger studies. Results obtained from a larger number of patients might give relevant insights on the exact nature of possible problems in MI performances experienced by patients with SCI and in further consequence have important implications on the incorporation of MI in rehabilitation, BCI applications, and neuroprostheses.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.