Therapeutic Effect of Schroth Method on Plantar Pressure and Related Muscle Imbalance in Children and Adolescent With Idiopathic Scoliosis

Abstract

Objective

To investigate the effects of Schroth method with a 6-month intervention on gait-specific dynamic plantar pressure distribution and functional muscular balance in children/adolescents with idiopathic scoliosis (IS).

Design

This is an observation before-after control study.

Setting

Conducted in the rehabilitation department of a tertiary hospital.

Participants

A total of 21 children/adolescents (N=21) with IS.

Interventions

All participants performed Schroth exercises ≥5 times/wk, each session ≥30 minutes. The training was completed independently at home under the supervision of guardians.

Main Outcome Measures

Plantar pressure parameters and surface electromyogram data of key postural muscles were collected before and after training. Outcome measures: (1) Cobb angle; (2) plantar pressure gait analysis: gait line length, single support line, left-right symmetry, heel-to-forefoot temporal difference, maximum force, maximum pressure, maximum temporal force in foot tripod regions; (3) surface electromyogram parameters: root mean square and mean power frequency, analyzed to assess activation and fatigue of the quadratus lumborum, erector spinae, gluteus medius, and semitendinosus.

Results

After 6-month Schroth intervention, patients with IS showed reduced Cobb angle (P<.05) and decreased convex-concave gait line length difference (P<.05), and improved left-right symmetry (P<.05). Significant decreases were also observed in heel-to-forefoot temporal difference (absolute and relative to total gait time: P<.05) and lumbar extensor/erector spinae root mean square values (P<.05).

Conclusions

The Schroth method not only improves Cobb angle in patients with IS, but also reduces gait line length asymmetry and paraspinal muscles activation imbalance. In addition, it enhances gait left-right symmetry and optimizes dynamic foot pressure distribution.

Idiopathic scoliosis (IS) is the most prevalent form of scoliosis affecting children and adolescents, characterized by a complex and multifactorial etiology,^1,2 Diagnostic is confirmed when a lateral spinal curvature is observed on standing posteroanterior radiographs, with a Cobb angle measuring ≥10°. The global prevalence of IS ranges between 0.47% and 5.20%, and there are significant differences in incidence across different regions.^3,4 The prevalence of scoliosis in adolescents aged 10-18 years in China is 1.2%.¹ The condition demonstrates a female predominance, with man-to-woman ratios ranging from 1:1.5 to 1:3, and is often associated with more rapid progression in women.^5,6 IS manifests as a 3-dimensional spinal disorders, involving lateral curvature—often accompanied by vertebral rotation—because of asymmetric forces acting on the spine. In early stages, increased biomechanical stress is placed on musculature. As the curvature progresses, visible trunk asymmetry develops and leads to spinal cord or nerve root compression, accompanied by chronic pain and psychosocial impacts; thoracic cage distortion can compromise cardiopulmonary function and affect other visceral organs.7, 8, 9 These potential complications underscore its importance as a significant public health issue. Early screening, timely detection, and proactive intervention are critical.

Scoliosis significant alters spinal biomechanics, affecting not only the vertebral column but also the paravertebral muscles, adjacent joints, and overall postural balance.10, 11, 12 These changes can lead to impaired balance, sensory deficits, and reduced spinal mobility.^13,14 Although the precise pathogenesis of IS in children and adolescents remains unclear, abnormal paravertebral muscle activity is a widely accepted contributor.15, 16, 17 Studies have shown that muscles on the convex side of the scoliotic spine are typically thinner and functionally weaker. In addition, the concave side demonstrates prolonged mechanical stress relaxation and a higher Deborah number, reflecting increased viscoelastic creep and contributing to biomechanical asymmetry.^18,19 Biomechanical imbalance between the 2 sides of the spine is considered a key factor in vertebral tilting and rotation, which may drive curve progression.²⁰ Other studies suggest that discrepant growth between vertebral bodies and surrounding muscloligament tissues could also initiate scoliosis.^21,22 Consistent with this, asymmetric paravertebral muscle electromyogram (EMG) activity has been observed in children and adolescents with IS.²³ Interestingly, some reports note reduced EMG amplitude on the concave side during standing,²⁴ whereas others report opposite findings, highlighting the complexity of neuromuscular adaptation. The concept “musculoskeletal balance” has been proposed by some researchers, framing spinal stability as a interplay between a “muscle system” (muscles and soft tissues) and a “bone system” (vertebrae and disc). Asymmetry within the muscle system may disrupt overall spinal equilibrium, thereby promoting scoliosis development. Thus, asymmetrical paravertebral muscle activity appears to be correlated with the onset and progression of IS in children and adolescents.^25,26

Literature reports that IS in children and adolescents can significantly impair walking performance,²⁷ primarily because of pelvic obliquity leading to apparent leg length discrepancy. It is hypothesized that this often manifests as limping, frequent tripping, and difficulty with prolonged ambulation. Haber and Sacco²⁸ noted that reduced step length and altered muscle activation timing in these patients lead to abnormal spatiotemporal gait parameters and decreased walking speed. Catan et al²⁹ observed that individuals with moderate-to-severe scoliosis exhibit abnormal plantar pressure distribution, with significant asymmetry between feet in severe cases. Mahaudens et al³⁰ demonstrated a 7% reduction in step length among patients with scoliosis compared with those without scoliosis. EMG studies using surface electrodes revealed prolonged activation timing during gait in several key muscles, indicating muscular imbalance and increased effort during walking.^30,31 This compensatory mechanism—marked by prolonged key muscle activation, muscular imbalance, and increased walking effort—contributes to earlier fatigue onset during prolonged physical activity. Therefore, biomechanical assessment of balance, gait, and plantar pressure represents a critical component in the clinical evaluation of IS.

Physiotherapeutic scoliosis-specific exercises (PSSE) are individualized exercise regiments tailored to a patient’s specific spinal curvature, with the primary objectives of reducing structural spine disorders and preventing curve progression,³² serving as a key therapeutic option for managing IS in children and adolescents. Among PSSE-based approaches, the Schroth method is regarded as a leading conservative treatments.³² Rooted in sensorimotor and kinesthetic principles, the Schroth therapeutic exercise includes postural correction, breathing retraining, patient education, and home exercises. Through improved trunk alignment, enhanced spinal flexibility, and muscular rebalancing, the Schroth approach aims to stabilize the Cobb angle and alleviate pain.³³ The Schroth method is a highly individualized exercise-based approach for scoliosis management. Internationally recognized and widely practiced, the Schroth method has demonstrated clinical efficacy and scientific validity.³⁴ Nonetheless, further investigation is needed to elucidate its biomechanical mechanisms and long-term effects in patients with scoliosis. We hypothesize that Schroth intervention will improve plantar pressure gait parameters and normalize surface electromyogram (sEMG) (sEMG) profiles of imbalanced muscles in children and adolescents with IS, enhancing their neuromuscular and biomechanical balance.

Currently, integrated studies using both plantar pressure gait analysis and sEMG to evaluate the effects of Schroth corrective gymnastics in IS remain relatively unexplored. This study aims to address this gap by collecting and comparing plantar pressure gait parameters and sEMG data from imbalanced muscle groups before and after a Schroth-based intervention. Through objective biomechanical and neuromuscular analysis, we seek to validate the efficacy and mechanistic basis of the Schroth method, clarify its clinical relevance in scoliosis rehabilitation, and support its evidence-based application in clinical practice.

Methods

General information

This study consecutively recruited 21 patients with IS from the Outpatient Department of the Second Affiliated Hospital of Wenzhou Medical University between September 2023 and December 2024. Of these, 4 were men and 17 were women; 19 patients presented with lumbar scoliosis and 2 with thoracolumbar scoliosis, with patients having other specific scoliosis patterns excluded from the study. Among patients with lumbar scoliosis, 11 presented with left convexity and 8 with right convexity; 2 patients with thoracolumbar scoliosis exhibited right convexity in the thoracic segment and left convexity in the lumbar segment. In this study, the convex side was defined as the main scoliotic side for the comparison of various indicators. Age at initial visit ranged from 6 to16 years, with 15 subjects under 10 years old and 6 aged ≥10 years. According to the inclusion and exclusion criteria, patients with IS with a Cobb angle of 10°-45° were enrolled. None of the participants wore orthotic braces. All participants provided informed consent (approved by the ethics committee. Approval No. 2025-K-174-02) before enrollment, and baseline evaluations were conducted before training initiation. After 6 months after training, all enrolled participants were followed up in the outpatient department, and all indicators were re-evaluated.

Inclusion criteria: (1) meets the diagnostic criteria for IS of the International Society for Spinal Surgery; (2) Cobb angle of 10° and 45°; (3) age 5-18 years, sex not restricted; (4) no obvious lower limbs or foot deformities; (5) no prior scoliosis treatment; and (6) able to understand and follow instructions, with independent ambulation.

Exclusion criteria: (1) non-IS; (2) lower limb length discrepancy; (3) history of neurologic, muscular, skeletal, or psychiatric disorders; (4) history of growth and development disorders or musculoskeletal trauma; (5) pretreatment or during-treatment brace wearers; and (6) inability to cooperate with gait analysis assessment. In this study, no subjects were excluded because of inability to cooperate with gait analysis assessment.

Termination criteria: (1) occurrence of serious adverse events or adverse reactions; (2) nonadherence to medical advice (operationally defined as failure to perform Schroth exercises as scheduled or nonattendance at follow-up visits), subjective unwilling to continue participation, or dropout.

Training methods

Before initiating Schroth corrective gymnastics, patients were educated on their scoliosis curve direction and type. Daily postural adjustments (standing, sitting, walking) were implemented, and physiological curvature correction (including resting sitting posture) was performed based on postural reflex principles. Corrective sitting positions and gaits were selected according to the patient’s scoliosis curve pattern. Given that most patients with IS had weak core strength, physical therapists instruct lumbar-abdominal core strength training on a treatment bed (fig 1A). (The detailed methods are provided in the supplemental appendix S1, available online only at http://www.archives-pmr.org/.)

Fig 1.

(A) Lumbar-abdominal core strength training. (B) Schroth exercises.

The Schroth corrective gymnastics comprises 3 core components (Fig 1B): restoring muscle symmetry for postural alignment; training directed airflow to the body’s concave side via breathing exercises; and maintaining proper daily posture. Its 4 most common exercises—50X, muscle cylinder, prone, and Sailboat exercises, are applicable to various curve types. (The details are provided in the supplemental appendix S1)

Patients were instructed to independently perform home-based core and abdominal strengthening exercises combined with Schroth corrective gymnastics at least 5 times weekly under the supervision of their guardians. Specifically, core and abdominal strengthening exercises were allocated 10 minutes per training session, whereas the Schroth exercise protocol was standardized as follows: each movement was repeated for 20 repetitions per set, with 5 consecutive sets completed per session, corresponding to a total duration of approximately 30 minutes. Strict adherence to this combined exercise regimen was mandated to ensure consistent implementation of home-based training. Patients with access to care should attend regular outpatient follow-up visits for movement protocol adjustments. The training duration for this intervention is 6 months. In this study, 83 subjects were treated. Planned follow-up was once a month, with 21 subjects adhering to the protocol. Visit attendance was documented during outpatient follow-ups (with therapist evaluations and treatment plan adjustments), whereas home exercise adherence was tracked via telephone inquiries with parents. The Schroth therapist holds an International Certification in Schroth 3D Scoliosis Therapy (Germany) and is an International Certified Instructor of Lyon Scoliosis Therapy (France), graduated from the Royal College of Surgeons in Ireland, and serves as a Committee Member of the Pediatric Physical Therapy Group and a Member of the International Society of Clinical Prosthetics and Orthotics.

Detection methods

Evaluation and measurement of Cobb angle

The Cobb angle is a key metric for assessing the severity of scoliosis and is the most widely used angular parameter for monitoring the condition, as treatment aim to delay progression and correct the structural spine disorders. According to the rehabilitation diagnosis and treatment guidelines for adolescent idiopathic scoliosis (2024 Edition), all patients with suspected adolescent idiopathic scoliosis (AIS) should be diagnosed and followed up using standing full-spine anteroposterior and lateral radiographs to evaluate curve location and severity, vertebral development, and growth parameters.³⁵

A Cobb angle ≥10° on anteroposterior radiographs confirms the diagnosis of scoliosis. The measurement procedure is as shown in fig 2A and B, with detailed protocols provided in the supplemental appendix S1. Image measurements were performed by professional radiologists, who were blinded to patients’ exercise treatment status.

Fig 2.

(A) Cobb angle before Schroth exercise. (B) Cobb angle after Schroth exercise.

Plantar pressure gait analysis of convex and concave sides in patients with IS before training

Plantar pressure gait analysis is a technology used to measure and analyze the distribution of pressure under the foot during standing and walking. It is widely applied in clinical medicine, sports science, and rehabilitation. The system operates using capacitive sensors to acquire plantar pressure data, which is transmitted to a computer for analysis. Outputs include dynamic and static pressure distribution maps, gait parameters, and automated reports featuring metrics such as pressure curves, average pressure profiles, single-foot pressure patterns, center of pressure trajectory, and pressure center range. Additional gait parameters include step velocity, velocity variability, total walking time, stance phase duration, swing phase duration, single and double support times, step length, stride length, and step width. Applications range from gait and balance assessment to custom orthotic design and diabetic foot risk evaluation.

In this study, the Zebris FDM-TR40 plantar pressure gait analysis system was employed to collect gait parameters during walking, including step length, gait line length, single support line, left-right symmetry (reflecting lateral deviation of the center of gravity from the midline during standing), heel-to-forefoot time progression, and maximum force, maximum pressure, and timing of maximum force across 3-foot regions. All participants underwent practice trials to acclimatize to the testing procedure. Subjects stood barefoot on the sensor plate with arms relaxed at their sides and feet positioned naturally at shoulder width. They then walked at a self-selected speed for dynamic data acquisition. Because of the relocation of the outpatient department to the Oujiang Estuary Campus of the Second Affiliated Hospital of Wenzhou Medical University during seasonal peaks (winter and summer vacations), where no plantar pressure system was available, data collection was limited. Consequently, only 13 cases were included in this analysis.

All plantar pressure assessments were performed by rehabilitation therapists who had received professional training and held specialized certifications in gait analysis. Furthermore, analysts responsible for data processing and interpretation were blinded to the testing time points (ie, pretraining and post-training) to mitigate potential assessment bias.

Surface EMG detection of the related imbalance muscle groups

The sEMG analysis system acquires neuromuscular bioelectrical signals via electrodes placed on the skin which are then recorded, amplified, processed, and fed back to quantify muscle function. Because of its noninvasiveness, absence of trauma, ease of operation, real-time dynamic monitoring, and ability to measure multiple targets simultaneously, sEMG has gained substantial practical value in clinical medicine, rehabilitation, and sports science in recent years. Moreover, sEMG is currently the only method that enables rehabilitation physicians to assess actual muscle function under dynamic conditions.

In this study, a wireless sEMG system (Model: FreeEMG300^a; Commissioning date: November 27, 2018; Serial No: 04-9F F2) was used to collect sEMG signals from related imbalance muscle groups during walking. Data acquisition and analysis were performed using BTS EMG Analysis software^b. Before testing, the patient’s skin was cleaned and dried. Disposable electrode pads were used to bilateral monitor the quadratus lumborum (QL), erector spinae (ES), gluteus medius (GM), and semitendinosus (ST), with placement strictly adhering to SENIAM guidelines and detailed anatomical landmarks specified for each recorded muscle. For the ES muscles, we placed the EMG electrodes at sites adjacent to the scoliosis curvature. For the remaining muscles included in the study, the electrodes were positioned at their respective midpoints. During testing, the patient walked at a normal pace on open, flat ground for at least 30 seconds to capture sEMG data during gait maintaining consistent movement amplitude throughout the trials and completing a predefined number of valid gait cycles to ensure data reliability. Consistent with SENIAM guidelines, EMG signals were normalized using maximal voluntary contractions: for each recorded muscle, maximal voluntary contraction tasks involved isometric contraction against manual resistance (3-s hold per repetition, 3 repetitions total, with 30-s rest intervals between trials).

The main parameters collected in this study included: (1) root mean square (RMS): a common time-domain analysis metric, defined as the RMS of amplitude over a given period, reflecting the average characteristics of electrical discharge within that interval. RMS values correlate with the level of sEMG amplitude and are generally associated with motor unit recruitment and synchronization of firing patterns. Higher RMS values indicate stronger muscle contraction and reflect the level of muscle activation and signal intensity. (2) mean power frequency (MPF): a frequently used frequency-domain indicator, representing the average frequency over a specific period. Under muscle fatigue conditions, MPF tends to decrease, thus providing insight into spectral characteristics and serving as an indicator of muscle fatigue. This study aimed to collect and analyze RMS and MPF values from sEMG signals of the aforementioned imbalance-related muscle (QL, ES, GM, ST), and to compare the differences before and after treatment.

Given that sEMG signals are considerably influenced by age, height, weight, and daily physiological conditions—resulting in variability between subject participants and within the same subject participant pre- and post-treatment—sEMG data were normalized to a baseline. Specifically, the concave side was set to 100, and the convex side was calculated as: (pretreatment convex side value/pretreatment concave side value) × 100.

Statistical methods

This study used SPSS 27.0^c for statistical analysis. Measurement data are presented as mean±standard ($\overline{\mathrm{x}}\pm s$) deviation. Normality of data distribution was assessed using the K-S test. Normally distributed data were compared using Student t test, whereas not-normally distributed data were analyzed with the Mann-Whitney U test. Pretraining, all indicators (sEMG values after normalization) between the convex and concave sides were compared using a paired-sample t-tests. Pre- versus post-training differences in all indicators were evaluated using the paried t test. P value <.05 was considered statistically significant.

Results

Plantar pressure gait analysis of convex and concave sides in patients with IS before training

Parameters with statistically significant differences were selected as outcome measures to evaluate training efficacy before and after intervention. The analysis revealed several significant asymmetries: the gait line length on the convex side (152.99±8.60mm) was 10.74±2.37 mm longer than that on the concave side (142.25±8.79mm), which may reflect altered weight-bearing distribution and gait alignment associated with scoliosis-related postural imbalance; the heel-to-forefoot loading time on the convex side (0.28±0.03s) was 0.06±0.02 s shorter than the concave side (0.34±0.29s), whereas the ratio of heel-to-forefoot time was 6.38%±2.64% lower on the convex side (33.42%±2.96%) compared with the concave side (39.80%±2.91%), indicating faster load transfer from heel-to-forefoot on the convex side potentially related to compromised muscle control and altered biomechanical efficiency in the scoliotic gait; and the forefoot maximum temporal force value on the convex side (66.68%±1.89%) was 4.21%±1.68% lower than the concave side (70.89%±1.57%), an asymmetry that may reflect uneven sustained forefoot loading during gait that could contribute to long-term foot discomfort or functional adaptation. All the above differences reached statistical significant (P<.05, table 1), and the observed magnitudes of these relative differences are clinically relevant, thereby supporting their utility as sensitive indicators for monitoring training-induced improvements. In contrast, no statistically significant differences were observed between the 2 sides in terms of step length, single support phase duration, maximum pressure in the 3 foot regions, or maximum force in the midfoot and heel (table 1).

Table 1.

Comparison of plantar pressure and gait parameters between convex (main scoliotic side) for the comparison of various indicators and concave side in patients with IS training ($\overline{\mathrm{x}}\pm s$, n=13).

Giat Parameters	Convex Side	Concave Side	Convex-Concave Difference	P Value
Step length (cm)	24.54±2.04	26.23±2.44	−1.69±1.17	.173
Gait line length (mm)	152.99±8.60	142.25±8.79	10.74±2.37	.001
Single support line (mm)	58.56±8.37	53.68±7.58	4.88±3.07	.138
Heel-to-forefoot loading time (s)	0.28±0.03	0.34±0.29	−0.06±0.02	.028
Heel-to-forefoot loading time (%)	33.42±2.96	39.80±2.91	−6.38±2.64	.032
Forefoot maximum force value (N)	227.79±27.82	220.48±26.53	7.31±10.50	.500
Midfoot maximum force value (N)	82.98±6.60	92.90±11.42	−9.92±6.50	.153
Heel maximum force value (N)	181.74±20.68	186.44±20.66	−4.70±6.62	.491
Forefoot maximum pressure value (N/cm)	14.80±1.84	14.42±1.70	0.38±0.80	.639
Midfoot maximum pressure value (N/cm)	8.19±0.57	8.68±0.68	−0.48±0.48	.336
Heel maximum pressure value (N/cm)	17.43±1.93	16.62±1.36	0.82±0.91	.386
Forefoot maximum temporal force value (%)	66.68±1.89	70.89±1.57	−4.21±1.68	.028
Midfoot maximum temporal force value (%)	48.31±2.45	49.67±2.28	−1.36±2.29	.564
Heel maximum temporal force value (%)	25.94±2.38	27.34±1.81	−1.40±2.28	.550

Normally distributed data were analyzed by Student t test, and non-normally distributed data by Mann-Whitney U test; P<.05 was considered statistically significant.

Comparative analysis of sEMG parameters between imbalanced muscle groups on the convex and concave sides in patients with IS before training

The comparison of RMS values revealed that the QL and ES muscles on the convex side exhibited significantly greater RMS values compared with those on the concave side. Specifically, the convex side showed an absolute difference of 76.58±36.78 (standardized difference: 72.90±15.37) and a relative increase of approximately 78.7% in QL RMS values (calculated as [(convex value − concave value)/concave value] × 100%), and an absolute difference of 35.87±10.34 (standardized difference: 48.09±13.80) and a relative increase of approximately 40.5% in ES RMS values compared with the concave side. In contrast, no statistically significant difference were observed in the RMS values of the GM and ST muscles between the convex and concave side (table 2).

Table 2.

Comparison of surface electromyogram (sEMG)-derived RMS amplitude (μV) between imbalanced muscle on convex (main scoliotic side) and concave sides in patients with IS before training ($\overline{\mathrm{x}}\pm s$, n=21).

sEMG Parameters	Convex Side	Concave Side	Convex-Concave Difference	P Value	Standardized Difference	Standardized P Value
QL	173.83±52.00	97.26±21.51	76.58±36.78	.050	72.90±15.37	<.001
Erector spinae (ES)	124.51±15.90	88.65±13.45	35.87±10.34	.002	48.09±13.80	.002
GM	207.82±48.92	268.60±64.39	–60.78±31.63	.069	0.30±19.40	.988
Semitendinosus (ST)	354.54±90.05	510.21±123.41	−155.67±68.81	.035	0.10±18.31	.996

Normally distributed data were analyzed by Student t test, and non-normally distributed data by Mann-Whitney U test; P<.05 was considered statistically significant.

A comparison of MPF between the convex and concave sides in patients with IS revealed significant regional differences in muscle activity, along with notable magnitudes of these differences. In the lumbar region, the MPF of the QL was significantly lower on the convex side compared with the concave side. In the lumbar region, the MPF of the QL was significantly lower on the convex side compared with the concave side (mean difference: −12.54±4.23). The standardized difference for QL MPF was −13.78±5.74, indicating a substantial relative magnitude of this side-to-side discrepancy. In contrast, no significant difference was observed in the MPF of the ES muscles between sides. Within the gluteal region, the GM on the convex side exhibited a significantly higher MPF than that on the concave side, whereas the ST showed no significant different inter-side difference in MPF (table 3).

Table 3.

Comparison of surface electromyogram (sEMG)-derived MPF value (Hz) between convex (main scoliotic side) and concave sides in patients with IS before training ($\overline{\mathrm{x}}\pm s$, n=21).

sEMG Parameters	Convex Side	Concave Side	Convex-Concave Difference	P Value	Standardized Difference	Standardized P Value
QL	63.73±3.72	76.27±3.96	−12.54±4.23	.008	−13.78±5.74	.026
Erector spinae (ES)	76.98±3.77	78.59±4.04	−1.62±3.87	.681	0.98±5.69	.865
GM	59.33±5.79	52.22±4.20	7.11±4.54	.133	15.41±6.99	.039
Semitendinosus (ST)	65.47±6.20	61.05±5.63	4.42±3.18	.179	8.91±5.41	.115

Normally distributed data were analyzed by Student t test, and non-normally distributed data by Mann-Whitney U test; P<.05 was considered statistically significant.

Comparison of Cobb angle before and after training in patients with IS

The Cobb angle of patients with IS showed a statistically significant decrease after training compared with pretraining measurements (table 4, fig 2).

Table 4.

Comparison of Cobb angles before and after training for patients with IS ($\overline{\mathrm{x}}\pm s$, n=21).

Parameter	Before Training	After Training	Before-After Training Difference	P Value
Cobb angle (°)	16.48±1.03	10.28±1.67	6.20±1.37	.00021

Normally distributed data were analyzed by Student t test, and non-normally distributed data by Mann-Whitney U test; P<.05 was considered statistically significant.

In this study, the pretraining Cobb angle had a mean of 16.5°, a maximum of 28.12°, and a minimum of 10.46°, with scoliosis curve types predominantly including thoracic and thoracolumbar curves. After 6 months of Schroth corrective gymnastics training, the maximum reduction in Cobb angle was 22.67°, whereas the minimum change was −0.17°. Of particular note, 2 patients exhibited a shift in curvature direction from one side to the other, with a decrease in magnitude. Because of this change in lateral bending direction, the difference in Cobb angle for these 2 patients was calculated as the sum of the pre- and post-training angles, resulting in the maximum observed difference of 22.67° (table 4, fig 2). Overall, 6 patients in this study exhibited a statistically significant improvement in scoliosis Cobb angle, 15 demonstrated no notable changes, and none presented with disease progression.

Comparison of pre- and post-training plantar pressure gait parameters in patients with IS

After training, the gait-related indicators of patients with IS showed significant improvements in multiple key aspects: the step length difference decreased by an absolute value of 1.92±0.83 cm (from 3.38±0.81cm to 1.46±0.43cm, relative reduction of 56.8%), the gait line length difference between the convex and concave sides was reduced by 9.27±2.15 mm (from 10.74±2.37mm to 1.47±1.50mm, relative reduction of 86.3%), and the left-right asymmetry index decreased by 4.85±1.36 mm (from 9.01±1.66mm to 4.16±1.25mm, relative reduction of 53.8%), which clinically enhances gait trajectory consistency, walking stability; in terms of temporal parameters, the heel-to-forefoot temporal difference decreased by 0.06±0.02 s (from 0.08±0.02s to 0.03±0.03s, relative reduction of 75.0%) and its ratio to total gait time decreased by 5.36%±1.81% (from 9.47%±1.69% to 4.11%±0.76%, relative reduction of 56.6%), reflecting improved coordination between heel strike. In contrast, no significant difference was observed in the maximum force during the forefoot phase before and after training (table 5, fig 3A and B).

Table 5.

Comparison of pre and post-training plantar pressure gait parameters in patients with IS ($\overline{\mathrm{x}}\pm s$, n=13).

Gait Parameters	Before Training	After Training	Before-After Training Difference	P Value
Step length difference (cm)	3.38±0.81	1.46±0.43	1.92±0.83	.039
Gait line length difference (mm)	10.74±2.37	1.47±1.50	9.27±2.15	.001
Single support line difference (mm)	4.88±3.07	−1.43±1.63	6.32±3.44	.091
Left-right symmetry (mm)	9.01±1.66	4.16±1.25	4.85±1.36	.004
Heel-to-forefoot temporal difference (s)	0.08±0.02	0.03±0.03	0.06±0.02	.016
Heel-to-forefoot temporal difference (%)	9.47±1.69	4.11±0.76	5.36±1.81	.012
Forefoot maximum force value difference (N)	30.48±6.10	24.66±4.52	5.82±4.91	.259
Midfoot maximum force value difference (N)	16.55±5.25	21.00±7.55	−4.45±3.37	.212
Heel maximum force value difference (N)	18.22±4.24	12.98±4.28	5.24±5.46	.356
Forefoot maximum pressure value difference (N/cm)	1.98±0.57	1.55±0.42	0.44±0.55	.439
Midfoot maximum pressure value difference (N/cm)	1.19±0.37	1.38±0.42	−0.19±0.34	.587
Heel maximum pressure value difference (N/cm)	2.38±0.63	2.65±1.17	−0.26±0.81	.754
Forefoot maximum temporal force value difference (%)	4.98±1.50	2.95±0.69	2.02±1.28	.141
Midfoot maximum temporal force value difference (%)	6.65±1.31	4.22±1.75	2.44±2.51	.351
Heel maximum temporal force value difference (%)	5.80±1.60	4.46±1.16	1.34±1.56	.407

Normally distributed data were analyzed by Student t test, and non-normally distributed data by Mann-Whitney U test; P<.05 was considered statistically significant.

Fig 3.

(A) Plantar pressure gait parameters of a single participant pre-Schroth corrective gymnastics training. The panel presents comprehensive gait metrics for an individual, encompassing dynamic and static pressure distribution maps, core gait parameters, and automated analytical reports. The reports feature key metrics including pressure curves, average pressure profiles, single-foot pressure patterns, center of pressure trajectory, and pressure center range. Additional gait parameters include step velocity, velocity variability, total walking time, stance phase duration, swing phase duration, single/double support times, step length, stride length, and step width. (B) Plantar pressure gait parameters of a single participant post-Schroth corrective gymnastics training. The panel displays comprehensive gait metrics for an individual (mirroring the parameter set in Panel A). For comparative interpretation between panels A and B, key metrics of focus include gait line length, single support line, left-right symmetry, heel-to-forefoot temporal difference, maximum force, maximum pressure, and maximum temporal force in foot tripod regions.

Comparison of sEMG parameters of related imbalanced muscle groups before and after training in patients with IS

After training, the RMS difference between the convex and concave sides of the QL and ES muscles in patients with IS decreased significantly compared with pretraining values: the baseline RMS value of QL was 72.90±15.37 μV and reduced to 25.22±12.04 μV post-training, with an absolute difference of 47.68±12.47 μV and a relative reduction of approximately 65.4%, whereas ES had a baseline RMS of 47.63±13.91 μV, a post-training value of 15.15±8.58 μV, an absolute difference of 32.48±14.67 μV, and a relative reduction of approximately 68.2%, which is clinically meaningful as it reflects improved muscle activation symmetry of these core stability-related muscles, potentially enhancing lumbar stability. In contrast, no significant differences were observed in RMS values of the GM and ST muscles before and after treatment. Furthermore, the MPF differences between convex and concave sides for the QL, ES, GM, and ST muscles did not show statistically significant changes after training compared with baseline (table 6, fig 4A-D).

Table 6.

Comparison of RMS and MPF difference between convex (main scoliotic side) and concave sides of related imbalanced muscle groups before and after training in patients with IS ($\overline{\mathrm{x}}\pm s$, n=21).

sEMG Parameters	Before Training	After Training	Before-After Training Difference	P Value
RMS (μV)
QL	72.90±15.37	25.22±12.04	47.68±12.47	.001
Erector spinae (ES)	47.63±13.91	15.15±8.58	32.48±14.67	.039
GM	−34.28±28.02	−2.36±18.84	−31.92±20.58	.137
Semitendinosus (ST)	−38.63±29.33	−20.27±12.59	−18.36±19.31	.353
MPF (Hz)
QL	−13.78±5.74	−12.59±6.50	−1.19±5.67	.836
Erector spinae (ES)	0.979±5.69	−5.29±3.52	6.27±4.91	.217
GM	15.42±6.99	7.92±4.63	7.49±6.16	.238
Semitendinosus (ST)	15.41±6.99	6.55±3.78	8.86±4.88	.084

Normally distributed data were analyzed by Student t test, and non-normally distributed data by Mann-Whitney U test; P<.05 was considered statistically significant. Symbols “+” and “−” denote that the measured value is larger on the convex side and concave side of the spine, respectively.

Fig 4.

(A, C) sEMG parameters pre-Schroth corrective gymnastics training; (B, D) sEMG parameters post-Schroth corrective gymnastics training. sEMG signals (μV) of muscles in a single individual during a 30-second lumbar-pelvic functional task. Panels A-D display activities of bilateral quadratus lumborum, erector spinae, gluteus medius, and semitendinosus, respectively. No signal standardization was performed. Key observation: asymmetrical activation of bilateral homologous muscles (eg, disparate amplitudes in quadratus lumborum/erector spinae).

Discussion

The Schroth method has gained broad recognition for its therapeutic benefits in the management of scoliosis. Domestic and international studies have consistently demonstrated that Schroth corrective gymnastics training improves back muscle strength, reduces Cobb angle, slows curve progression, decreases the likelihood of surgery, and improves quality of life.36, 37, 38 Dr Hans Weiss, the grandson of Katharina Schroth, has contributed substantially to this body of evidence through multiple publications supporting these outcomes. In a randomized clinical trial by Kyrkousis et al,³⁶ 12 months of supervised Schroth training in adolescents with AIS significantly reduced the Cobb angle and improved quality of life, with these benefits sustained 6 months after intervention. This rigorous study addresses the research gap in existing literature regarding short-term or unsupervised exercise interventions, providing level I evidence for a potential dose-response relationship between Schroth exercise compliance and clinical outcomes. Complementing this work, interventional studies by Yuan et al³⁹ validated the efficacy of PSSE in mild AIS—extending the evidence base to mild curves that were not covered by Kyrkousis et al’s³⁶ focus on moderate-to-severe AIS (mean Cobb angle 34.04°). Collectively, these 2 studies provide severity-spanning (mild to moderate-severe) and mechanistically supported evidence of PSSE, including Schroth-based approaches. Findings of this study also align with previous research, providing preliminary evidence that Schroth corrective gymnastics training exerts a Cobb angle-reducing effect and thus may have a positive impact on the magnitude of spinal curves in individuals with IS. As a core diagnostic and severity-assessment parameter in scoliosis, the Cobb angle reflects 3-dimensional structural remodeling of the spine. Its changes are closely linked to alterations in plantar pressure, gait parameters, and muscle activity patterns. Our findings may help enhance understanding of IS pathophysiology and provide reference for the development of rehabilitation strategies and progress monitoring.

Plantar pressure gait analysis is an effective method for evaluating gait patterns in patients with scoliosis, offering precise diagnostic support for therapeutic gait training.⁴⁰ Consistent with previous studies on AIS gait biomechanics,⁴¹ our findings demonstrated that children and adolescent patients with IS have longer gait line length on the convex side than the concave side. In addition, the convex side exhibits shorter heel-strike-to-toe-off duration and lower proportion of the total gait cycle, which refines previous observations of shortened stance phases in patients with AIS by explicitly characterizing convex-concave asymmetry.⁴² In addition, the maximum force during forefoot loading phase is also reduced on the convex side. These asymmetric gait characteristics have important clinical implications for IS rehabilitation. Biomechanically, the convex side of the spinal curvature typically manifests muscle elongation and reduced tension, which may directly contribute to increased gait line length of the ipsilateral lower extremity. During the single-limb support phase, diminished muscle tension may elevate the load borne by the convex side limb, necessitating prolonged compensatory efforts to maintain postural stability—a mechanism that underpins the observed shorter stance-phase-related parameters on the convex side. Collectively, such convex-concave disparities in gait metrics induce global locomotor asymmetry and uneven plantar pressure distribution. Notably, the side-to-side difference in gait line length, which aligns with the aforementioned asymmetrical vertebral translation and torsion patterns, represents a hallmark biomechanical alteration in AIS gait. Clinically, these findings can guide personalized rehabilitation strategies (eg, targeted extension of the convex side stance phase) to alleviate trunk imbalance and enhance the precision of therapeutic evaluation and shorter stance phase may reflect locomotor compensation to maintain postural stability under spinal deformity.^28,43

Spinal curvature leads to a more pronounced ipsilateral or contralateral pelvic drop during ambulation. To maintain balance, patients with scoliosis often adopt compensatory strategies such as increased step width to enhance stability; reduced cadence to minimize gait asymmetry; and postural adjustments of the pelvis and trunk to reduce asymmetric loading between the convex and concave sides. This imbalance is also associated with increased unilateral limb loading and disrupted sequencing of body rotation and flexion,⁴⁴ resulting in deviation of the plumb line, lateral trunk inclination, and displacement of the center of gravity. Plantar pressure distribution is often asymmetrical between the 2 sides. In addition, spinal deformities are frequently accompanied by pelvic tilt and rotation, flat feet, high arches, and other lower limb biomechanical abnormalities. Syczewska et al⁴⁵ demonstrated that in patients with thoracolumbar scoliosis, the severity of spinal curvature and the presence of pelvic deformities significantly influence plantar pressure and gait patterns during walking. When pelvic disorder is not significant, reduction of the Cobb angle is associated with more balance muscular activity during gait. This is reflected in reduced differences in gait line length between sides, decreased asymmetry in the heel-to-forefoot progression time, and a smaller discrepancy between this interval and total gait time. The left-right symmetry also decreases, approaching 0. This study further showed that after Schroth-based corrective gymnastics training, reduction of the Cobb angle was accompanied by decreased differences in gait line length between the convex and concave sides, improved left-right symmetry, and narrower discrepancies in heel-to-forefoot time progression and its ratio to total gait time. The left-right symmetry index objectively reflects the degree of gait asymmetry in children and adolescents with IS —a higher value indicates greater asymmetry. A post-treatment reduction in this value suggests improved gait balance between the convex and concave sides. Regarding the single support line, the relatively small sample size may have limited the statistical power to detect potential differences, which could affect the reliability and generalizability of the results. Future studies with larger cohorts are needed to further validate these findings.

Research shows that many patients with IS demonstrate reduced balance capacity and impaired postural control, both statically and dynamically. Progressive spinal disorders in these individuals can lead to asymmetric spinal loading, abnormal center of pressure, increased trunk muscle activity, and muscular imbalance on either side of the spine.^46,47 sEMG data were collected during natural walking at habitual speed on an open flat surface (≥30s per trial) with consistent movement amplitude and a predefined number of valid gait cycles. This ecologically valid task was selected to reflect real-world locomotor function—core to daily lives of patients with IS—ensuring that observed muscle asymmetries are clinically relevant and not confounded by artificial task constraints. EMG asymmetry in affected muscles can be quantified using parameters such as RMS, which reflects muscle activation level and signal amplitude, and MPF, which reflects spectral properties and muscle fatigue. This study identified significant differences in RMS and MRF of the QL muscle among children and adolescents with IS. The convex side showed higher RMS and lower MPF values compared with the concave side. Scoliosis alters spinal biomechanics, typically shortening muscles on the concave side and elongating those on the convex side. This results in differential load distribution: elongated convex side muscles experience higher mechanical stress and are more susceptible to fatigue. Impaired neuromuscular control may also contribute stress and are more susceptible to fatigue. Impaired neuromuscular control may also contribute to asymmetric muscle activation patterns and intensity between sides. Together, these factors account for the electrophysiological differences observed between concave and convex side muscles. In contrast, the ES—another lumbar muscle group—showed differences only in RMS, likely because it functions mainly in isometric contraction to maintain spinal stability during gait and minimize excessive flexion or lateral bending,⁴⁸ reducing its susceptibility to short-term fatigue. No significant differences were observed in the GM or ST between sides. This may be because of: (1) the GM acting synergistically with other hip stabilizers (eg, gluteus minimus and tensor fasciae latae) in abduction and pelvic control, whereas the ST primarily assists in knee flexion and hip extension—functions less compromised in mild scoliosis; (2) limitations in sample size and the predominance of thoracolumbar curves in the cohort. (3) In addition to curve location, there may be a different Schroth posture types presenting with different pelvis blocks effect possibly leading to heterogeneity in pelvis muscle’s function, which would limit the ability to detect differences in response to treatments. Future studies should incorporate larger sample sizes and patients with curves involving diverse spinal regions, while also accounting for Schroth posture types, to further elucidate these muscular asymmetries and their associations with treatment responses. After Schroth corrective gymnastics training in children and adolescents with IS, a reduction in the RMS gap between the QL and ES muscles is observed, indicating decreased asymmetry between the convex and concave sides. However, this study failed to yield a significant effect of Schroth corrective gymnastics on muscle MPF, possibly because of the relatively short intervention period and the younger age of the participants, resulting in less fatigue from compensation mechanisms. Increasing exercise intensity may elicit convex-concave differences in MPF, warranting further investigation.

Study limitations

It is the first prospective study integrating plantar pressure gait analysis with sEMG to evaluate the intervention’s effects on gait mechanics and muscle activity imbalances in this patient population. It also provides a systematic biomechanical analysis of the asymmetrical plantar pressure, gait patterns, and muscular activation resulting from 3-dimensional spinal deformities. However, several limitations warrant consideration. First, 83 patients were initially enrolled, 62 of whom were lost to follow-up. As noted previously, poor compliance to home-based training constituted the primary cause of attrition, with additional cases attributed to academic scheduling conflicts and reduced parental attention to disease progression. This substantial attrition resulted in a relatively limited small final sample size, potentially compromising the study’s statistical power. Future studies should involve larger clinical cohorts to validate these findings. Second, the study population consisted primarily of patients with lumbar-predominant structural scoliosis, thus leaving the potential differential therapeutic effects across various curve types unaddressed. Therefore, more extensive and systematic studies are imperative to establish robust clinical evidence. Third, this study was focused on patients with mild to moderate scoliosis without stratification by functional/structural deformity classification, which limits the generalizability of our findings to patients with severe or well-characterized functional/structural subtypes.

Conclusions

Schroth corrective gymnastics training effectively reduces the Cobb angle and improves the quantified 3-dimensional spinal deformity parameters in children and adolescents with low-to-moderate scoliosis in this study, while concurrently shortening overstretched convex side muscles. During movement after Schroth treatment, muscle activation tends to become more symmetrical bilaterally. However, this study did not demonstrate a significant effect of Schroth corrective gymnastics on muscle MPF.

Suppliers

a. FreeEMG300;

b. BTS EMG Analysis software; BTS Bioengineering (BTS S.p.A.), Italy

c. SPSS, version 27.0; IBM.

Disclosure

The investigators have no financial or nonfinancial disclosures to make in relation to this project.