*3.2. The PMSI and RMS of sEMG Activity of Paraspinal Muscles*

The PMSI and RMS values of sEMG activity of paraspinal muscles are shown in Table 2. The PMSI of pre-exercise and post-exercise in the relaxed standing position was over 1, which meant the sEMG activity of the paraspinal muscle on the concave side was lower than that of the convex side. The PMSI significantly reduced from 1.36 to 1.30 after exercise (*p* < 0.05), indicating the sEMG activity symmetry of the paraspinal muscle between the convex and concave side was improved.



\* The RMSconcave and the RMSconvex were significantly different (*p* < 0.05).

#### 3.2.1. The PMSI before, during and after the Schroth Exercise

The PMSI values before, during and after the Schroth exercise are shown in Figure 2. The PMSI during E1 reduced significantly to 0.93 from 1.36 (*p* < 0.05) in the relaxed standing position. It suggested that the sEMG activity of paraspinal muscle at the concave side increased and reached a similar level to that of the convex side, which improved the symmetry of the paraspinal muscles. The PMSI during E1 was closest to 1 among the four exercises, with no significant difference between the RMSconcave and the RMSconvex; thus, it may be regarded as the exercise with the highest symmetry of the sEMG activity among the four exercises. The PMSI during E2 was reduced to 0.75 from 1.36 (*p* < 0.05) in the relaxed standing position. This suggests that the sEMG activity of paraspinal muscle on the concave side increased and reached to the level closer to the convex side, which improved the symmetry of the paraspinal muscle. The RMSconcave and the RMSconvex during E2 did not show significant difference.

The PMSI during E3 increased significantly to 2.56 from 1.36 (*p* < 0.05) in the relaxed standing position. This suggested that the sEMG activity of paraspinal muscle at both sides increased but the convex side increased more, which reduced the symmetry of the paraspinal muscle. The PMSI during E3 was the least close to 1 among the four exercises, with the sEMG activity of convex side being significantly higher than that of the concave

side (*p* < 0.05). Therefore, it may be regarded as the exercise with the lowest symmetry among the four exercises. The PMSI during E4 increased significantly to 1.52 from 1.36 (*p* < 0.05) in the relaxed standing position. This suggests that the sEMG activity of the paraspinal muscle on both sides increased but the convex side increased more, which reduced the symmetry of the paraspinal muscle.

**Figure 2.** The PMSI before, during and after the Schroth exercise (*n* = 9).

3.2.2. The RMS of sEMG Activity before, during and after the Schroth Exercise

The RMS of sEMG before, during, and after the Schroth exercise is shown in Figure 3. The sEMG activity of paraspinal muscle was higher in all Schroth exercises than that in the relaxed standing position before exercise. The RMSconcave significantly increased after exercise (23.51 µV vs. 25.39 µV, *p* < 0.05), while the RMSconvex did not significantly change after exercise (17.28 µV vs. 19.60 µV, *p* > 0.05), which indicates that the exercise induced more sEMG activity of paraspinal muscle change on the concave side of the scoliotic curve.

**Figure 3.** The RMS of sEMG before, during, and after the Schroth exercise (*n* = 9).

The highest RMSconcave (81.79 µV) was observed in E2, which was a symmetric exercise against gravity and induced muscle contraction on both sides. The highest RMSconvex (68.77 µV) was observed in E4, which was an asymmetric exercise, with side bending to the convex side and stretching of the concave side. Upon comparing the magnitude of sEMG activity of paraspinal muscles, this study observed that E4 > E3 (*p* < 0.05) and E2 > E1 (*p* < 0.05) in both the convex and concave side of the scoliotic curve.

#### **4. Discussion**

This study innovatively applied the sEMG to investigate paraspinal muscle activities before, during and after the Schroth exercise in patients with AIS. The findings of this study will provide evidence and support the individualized and case-specific prescription of the Schroth exercise in future clinical practice, which could improve the effectiveness of treatment and improve the quality of life of patients with AIS.

The sEMG activity of the paraspinal muscle on the concave side was found to be lower than that of the convex side during the relaxed standing position in AIS patients. This could be explained by the prolonged stretching of the paraspinal muscles due to deformed vertebrae in AIS patients, which resulted in the asymmetry of muscle fiber types, lengths and locations at bilateral sides [5,19]. The indicated functional imbalance of the paraspinal muscles in AIS suggests that clinicians should prescribe specific exercise to improve muscle balance on both sides according to individual conditions.

This study found that the PMSI of AIS reduced by 4.6% after the Schroth exercise. A previous study also reported the reduced PMSI by 12.0% in the thoracic region and by 7.9% in the lumbar region after the Schroth exercise [6]. Since only patients with single lumbar scoliosis were recruited in the current study, the influence of curve location on the symmetry of paraspinal muscles could be investigated in future studies.

During symmetric exercises, the sEMG activity of the paraspinal muscle was symmetric, while during the relaxed standing position was asymmetric. Chwala et al. [9] also reported higher PMSI during symmetric exercise in comparison with the resting recordings. The possible reason could be that symmetric exercise tried to isolate the muscle contraction between the concave and convex side, and focused more on the atrophied concave side to improve the symmetry of paraspinal muscles [18]. For E1 (on the fours), the sEMG activity on the concave side increased more than the convex side and reached a symmetric sEMG activity on both sides. E1 may be regarded as a symmetric exercise with reduced longitudinal gravity on the spine, which would simultaneously correct the sagittal lordosis and coronal scoliosis of spinal deformity [20]. The symmetric sEMG activity of paraspinal muscles during E1 may be related to both the self-correction of the patients and the spontaneous correction by the postural change. This can also explain the highest RMSconcave in E2 (squatting on the bar), which is a symmetric exercise that was against gravity and induced higher muscle contraction on both sides.

During asymmetric exercises, the sEMG activity of paraspinal muscle on the concave side was lower than that of the convex side. The paraspinal muscle fiber was reported to be weaker on the concave side and stretched on the convex side in the scoliotic spine [21]. The convex side was usually used as the dominant side for daily activities. During asymmetric exercise, side bending created an imbalance load on the spine, requiring greater paraspinal muscle contraction to maintain stability. As a result, an increase in the predominance of the sEMG activity on the convex side was instigated. It could be a sign of an adaptive response to the greater use of the muscles on the convex side in patients with AIS. The highest RMSconvex was observed in E4 (sitting with side bending), which agreed with Chwala et al.' s study [9] who observed the highest sEMG activity of the convex side of paraspinal muscles in an asymmetric exercise, which involved actively stretching the concave side. They also reported that asymmetric exercises demonstrated larger differences in sEMG activity of the paraspinal muscles in comparison with symmetric exercises.

When considering individual patients, two out of nine patients demonstrated lower sEMG activity at the concave side during symmetric exercise, which was opposite to the other subjects. This might be because each patient had different motor habits and variable attempts when performing exercises. The same exercise could result in diverse performance quality and repeatability of the corrective patterns in practice [9]. Therefore, individualized exercise should be recommended based on the specific muscle response and performance quality of patients. This study validated the feasibility of applying sEMG to evaluate the muscle performance during the Schroth exercise, which will provide evidence and contribute to the case-specific training for patients with AIS in clinical practice. It may also be helpful to adopt some ultrasound imaging technologies [22] to study the internal paraspinal muscle contraction pattern during the exercise in AIS patients in the future.

This study has several limitations. This study only involved nine patients with lumbar scoliosis. A larger sample size with diverse types of scoliosis curve needs to be investigated. Unfortunately, due to the COVID-19 pandemic in China, it is extremely difficult to recruit more AIS patients for this study at this time and in the near future. The current study may serve as pilot investigation providing the theoretical foundation and research direction for future studies to further validate the current observations and deepen the knowledge in this field with larger samples. This study has focused on the immediate effect of the exercise on the paraspinal muscles, but lengthier studies will be necessary to confirm the long-term effects of the Schroth exercise on the performance of the paraspinal muscles. It would also be interesting to investigate whether any difference existed in paraspinal muscle activity during the Schroth exercise between adolescents with and without scoliosis. However, due to the limited number of available children/adolescent participants, it has been difficult to recruit the healthy adolescents without scoliosis to perform the Schroth exercise as a control group, especially under the current pandemic situation. Future studies could recruit some healthy children/adolescents without scoliosis to study the difference in paraspinal muscle activity during the Schroth Exercise.
