Associations Between Different Exercise Approaches and Brain Structure in Adolescent Idiopathic Scoliosis: An Exploratory Neuroimaging Study

Abstract

Study Design

Randomized Controlled Trial.

Background and Objectives

The Schroth method has emerged as a promising rehabilitation approach for Adolescent Idiopathic Scoliosis (AIS). This study compared the effects of Schroth and conventional exercise programs on spinal curvature, cortical thickness, and white matter pathways using structural MRI and Diffusion Tensor Imaging (DTI).

Methods

Thirty-six individuals with AIS were randomly assigned to a Schroth Exercise Group (SEG) or a Traditional Exercise Group (TEG). A healthy control group (n = 18) was also included. The sample size was calculated to provide 80% power (d = 0.60), and post hoc analysis indicated strong power (>0.99) for the observed treatment effect (d = 3.59). Both exercise groups completed supervised home-based programs for four months. Post-intervention MRI and DTI analyses were performed. The study was registered at ClinicalTrials.gov (NCT06410456).

Results

After adjusting for baseline severity, the SEG showed a significantly greater improvement in Cobb angle compared to the TEG (adjusted mean difference = 8.07°, P < .001). Although the SEG achieved a significantly lower post-treatment Cobb angle compared to the TEG, residual curvature remained significantly different from the healthy controls (P < .001), reflecting the structural nature of scoliosis. Additionally, after adjusting for age and intracranial volume, cortical thickness in sensorimotor regions and tractography metrics of the right corticospinal tract in the SEG were comparable to controls and significantly different from the TEG.

Conclusion

Schroth exercises were associated with brain structural features suggestive of adaptive neuroplastic responses. These exploratory findings highlight potential neurobiological mechanisms and support further longitudinal research.

Keywords

adolescent idiopathic scoliosis (AIS)Schroth exercises cortical thickness neuroplasticity corticospinal tract lemniscus medialis

Introduction

Adolescent Idiopathic Scoliosis (AIS) is a three-dimensional structural deformity of the spine that typically occurs in children and adolescents between the ages of 10 and 18. The condition is diagnosed when the Cobb angle measured in the coronal plane exceeds 10°. Based on the Cobb angle, spinal curvatures are classified as mild (10°–25°), moderate (25°–45°), or severe (>45°).^1-3

In individuals with AIS, not only is there a spinal deformity, but functional impairments such as postural asymmetry, disturbances in proprioceptive sensation, and postural instability are also observed. These symptoms are believed to have neuroanatomical and neurofunctional underpinnings, with structural changes in the brain, brainstem, and cerebellum potentially contributing to the condition.^4,5 Indeed, previous research has reported significant reductions in cortical thickness in certain brain regions of individuals with AIS compared to healthy controls.^3,6 These symptoms may negatively affect biomechanical force distribution and postural balance, ultimately reducing quality of life.⁷

Early diagnosis and intervention in AIS are critical for preventing curve progression and improving quality of life. Treatment strategies include exercise, bracing, and surgical intervention.⁸ The primary goals of exercise therapy are to reduce spinal deformity, restore thoracic functional capacity, improve locomotor ability, and enhance postural appearance.^7,8

An increasingly popular intervention in the management of AIS is the Schroth method, a scoliosis-specific exercise approach. This method aims to provide three-dimensional correction based on the individual’s curve pattern through a combination of sensorimotor, postural, and respiratory exercises.^7,9 In addition to correcting postural asymmetries, the method targets muscle imbalances using techniques such as exteroceptive and proprioceptive stimuli, isometric exercises, and mirror feedback.¹⁰

The literature suggests that the Schroth method is effective in reducing spinal curvature and improving functional outcomes.¹¹ However, its effects on the central nervous system, particularly in terms of cortical structure and neural pathways, remain underexplored. Neuronal plasticity refers to the brain’s ability to undergo structural and functional changes in response to environmental stimuli.^12,13 It is hypothesized that exercise-induced neuroplastic changes may occur in brain regions responsible for motor control and proprioception.

Magnetic Resonance Imaging (MRI) is a non-invasive imaging technology that generates detailed three-dimensional anatomical images using magnetic fields and radiofrequency waves.¹⁴ MRI incorporates various techniques, including volumetric analysis, shape analysis, voxel-based morphometry, cortical thickness measurement, texture analysis, diffusion tensor imaging (DTI), and functional MRI, to evaluate the human brain from both structural and functional perspectives.¹³

The aim of this study was to investigate the effects of the Schroth method on cortical thickness and neural pathways related to motor control and proprioception in individuals diagnosed with AIS, as compared to traditional exercise approaches, using MRI and DTI techniques. The study seeks to provide scientific evidence regarding the potential central nervous system-level effects of the Schroth method.

Materials and methods

Study Design

This study was a randomized controlled trial with parallel groups and blinded assessors. It was conducted in Kayseri between April 1, 2024, and February 1, 2025. The study protocol was approved. by the local ethics committee and carried out in accordance with the principles of the Declaration of Helsinki. Written and verbal informed consent was obtained from all participants and their parents or legal guardians. The study protocol was registered at ClinicalTrials.gov (NCT06410456), (https://register.clinicaltrials.gov/prs/beta/studies/ S000EB3U00000081/recordSummary). The study was registered on March 17, 2024, and the first participant was enrolled on April 1, 2024.

Participants

Participants were recruited among individuals diagnosed with AIS by an orthopedic specialist based on the Lenke classification criteria,¹⁵ and subsequently referred for treatment. Cobb angle measurements were performed by the same physician immediately before and after treatment.

Inclusion criteria:

• Aged between 10 and 18 years

• First-time AIS diagnosis

• Cobb angle between 10° and 25°

• Risser sign between 0 and 4

• Lenke Type 1 or 1A curve

Exclusion criteria:

• Other types of scoliosis

• Spina bifida

• Cardiopulmonary comorbidities

• Use of orthotic braces

• Mental health disorders or contraindications to exercise

• Previous AIS treatment

• Rheumatologic or neuromuscular disorders

• Lack of parental consent

Additionally, age- and gender-matched healthy individuals without scoliosis or any postural disorders were included as a control group for MRI comparisons. Healthy control participants were identified using a snowball sampling method. All individuals underwent clinical screening to exclude scoliosis, neurological, or musculoskeletal disorders. Physical activity levels were assessed through self-report questionnaires to ensure comparability with the patient groups. Controls were scanned using identical MRI acquisition protocols as patients, and all images were subjected to the same preprocessing and quality control procedures.

Intervention

Both the Schroth Exercise Group (SEG) and the Traditional Exercise Group (TEG) followed a supervised home exercise program for four months, five days a week, with each session lasting 60 minutes. Adherence to the exercises was monitored using two complementary methods: (i) daily exercise logs signed by the participant and parent/guardian, and (ii) weekly supervised online sessions (via video conference), during which the therapist verified correct performance and recorded the duration of the exercises. The same Schroth-certified physical therapist (with over 10 years of experience; initials HK) designed and supervised all interventions. Compliance was quantified as the percentage of prescribed sessions completed. When adherence fell below 80%, corrective measures were taken in collaboration with participants and families. Every four weeks, exercises were reviewed in the clinic and modified if necessary. All individual sessions and follow-ups were conducted by the same physiotherapist (HK). Overall adherence was 91.4% ± 6.8 (median [IQR] = 93% [88-96], range = 75-100%), and 85% of participants achieved ≥80% compliance.

Schroth Exercise Group (SEG)

Participants in the SEG followed a home program based on 3D Schroth exercises for approximately one hour, five days per week for four months. The first session covered Schroth fundamentals including rotational breathing, pelvic alignment, spinal elongation, and positioning techniques. Exercises included rotational breathing, shoulder counter-traction (supine, prone, and side-lying), chest twisting, sitting-based variations, weak side stretching, pole-assisted counter-traction, and Schroth gait. Exercises progressed from basic to advanced (e.g., from supine to upright and walking). Initial intensity was 2-5 sets of 4-6 repetitions, gradually increasing to 3-5 sets of 6-10 repetitions. Participants used personalized video guides and maintained exercise logs. Family members were briefed to support adherence.

Traditional Exercise Group (TEG)

The TEG followed a classical home-based exercise program also lasting one hour, five days a week, for four months. The protocol included stretching (targeting tight muscles on the concave side), strengthening (abdominals, back extensors, and convex-side muscles), postural training, and deep breathing exercises. Like the SEG, participants followed video instructions and maintained logs. Exercises progressed from lying to functional postures (e.g., standing, walking), from 1 to 3 sets, and 7-10 to 10-15 repetitions.

Healthy controls were recruited from the community and screened to exclude scoliosis, neurological, or musculoskeletal disorders. Physical activity level was assessed by self-report to ensure comparability with the patient groups. Controls underwent the identical MRI acquisition protocols as patients, and all images were subjected to the same preprocessing and quality control pipeline.

Outcome Measures

The primary outcome of this study was the change in the Cobb angle (°) measured on standing posteroanterior (PA) radiographs from baseline to 4 months. The Cobb angle was selected as the primary endpoint because it is the clinical gold standard for assessing the progression of scoliosis and the effectiveness of treatment.

Secondary outcomes were changes in cortical thickness in areas related to motor and sensorimotor integration. Additionally, tract-specific diffusion tensor imaging (DTI) parameters of the corticospinal tract and medial lemniscus, which are known to play a role in motor control and proprioceptive feedback, were also included.

Brain regions and tracts were chosen a priori based on their established roles in postural control, proprioception, and motor planning. These structures have been consistently implicated in prior scoliosis and motor-control neuroimaging studies (3,16).

Spinal Curvature

The degree of spinal curvature was assessed using the Cobb method on standard standing standing PA radiographs of the spine by a blinded orthopedist. Risser’s sign was also recorded to evaluate skeletal maturity. The Cobb measurements were repeated for both groups by the same blinded orthopedist, and the intraclass correlation coefficient (ICC) for inter-measurement reliability was found to be 0.98, indicating excellent consistency.

MRI Acquisition

MRI scans were performed using a 3T Siemens Magnetom Skyra system. Cortical thickness was assessed using a sagittal T1-weighted MPRAGE sequence (TR = 2300 ms; FOV = 250 mm; TE = 3.4 ms; slice thickness = 1 mm; matrix = 256 × 256).

Diffusion Tensor Imaging (DTI) was performed using a single-shot echo-planar imaging sequence (TR = 4900 ms; FOV = 230 × 230 mm²; TE = 95 ms; slice thickness = 3.5 mm; matrix = 128 × 128; voxel size = 1.8 × 1.8 × 3.5 mm). Twenty diffusion-weighted directions were used at b-values of 0 and 1000 s/mm². A total of 36 axial slices were acquired (∼6 minutes scan time). Neuroimaging (MRI and DTI) was performed only after completion of the four-month intervention period, and therefore no pre-intervention scans were obtained.

Imaging Data Processing

Cortical Thickness Analysis

T1-weighted data (in DICOM format) were converted to NIfTI and analyzed via the volBrain platform (https://volbrain.upv.es). The vol2Brain module was used, incorporating individual age and gender. Results were reported in mm and provided in PDF format (Figure 1).

Figure 1.

Estimation of brain cortical thickness using the vol2Brain software. *Representative patient scans are shown (not standardized templates). The color overlays indicate regional cortical thickness

Tractography Analysis

Tractography of the corticospinal tract and medial lemniscus was performed using DSI Studio (https://dsi-studio.labsolver.org/). Fiber tracking parameters included: threshold = 0.20, angular threshold = 70°, smoothing = 0.50, shortest fiber = 10 mm, longest = 1000 mm. The program’s tract atlas was used for bilateral tracking. Outcomes included fiber count, mean fiber length and fractional anisotropy (FA) which were compared between groups (Figure 2).

Figure 2.

Tractography analysis of the medial lemniscus (LM) with DSI Studio program. *Representative patient tract reconstructions are displayed. *(A): Axial view *(B): Coronal view *L-LM: Left medial lemniscus *R-LM: Right medial lemniscus

All diffusion data were preprocessed prior to tractography using standard pipelines. Motion and eddy-current distortions were corrected with the FSL 6.0.6 eddy tool, including outlier slice replacement and eddy current induced distortion correction. Diffusion tensors were fitted after correction using DSI Studio (build 2024.06), and tractography was subsequently performed with the parameters reported above.

High test-retest reliability has been reported with automatic segmentation using VolBrain/vol2Brain.¹⁷ To verify reproducibility in the current study, a random subset of 30 scans was reprocessed by an independent analyst, and ICC values >0.96 were found for all major regions. Intracranial Volume (ICV) normalization. To account for inter-individual differences in head size, DTI derived global metrics were adjusted for ICV estimated from T1-weighted images via volBrain/vol2Brain.

Previous literature on DTI has also reported moderate to excellent repeatability (ICC 0.80-0.92) for major pathways.^18,19 In the reanalysis of fifteen randomly selected cases, ICC values >0.90 were found for mean FA and MD in the corticospinal tract and medial lemniscus.

Sample Size

Sample size was calculated based on an expected 5° mean difference in Cobb angle between groups, with an effect size of Cohen’s d = 0.60. Power analysis (G*Power 3.1.9.4) determined that 30 participants (15 per group) would provide 80% power at α = 0.05. Accounting for a 15% dropout rate, 18 participants were recruited per group (total n = 36).^20,21 A post hoc power analysis conducted to compare post-treatment Cobb angles revealed that with a sample size of 18 participants per group, the study achieved statistical power greater than 0.99 to detect the observed large effect size (d = 3.59) at an alpha level of .05.

Randomization and Blinding

Participants were randomly assigned (1:1) to either group by an independent statistician using Excel’s random number function. Group allocations were sealed in opaque, sequentially numbered envelopes and opened in front of the participants at enrollment. Group assignment remained concealed from the physician assessors and data analysts. Due to the nature of the intervention, physiotherapist and patient blinding was not feasible; however, participants were instructed not to disclose their group to maintain assessor blinding.

Participants were randomized in a 1:1 ratio to the Schroth or Traditional exercise groups using a computer-generated randomization list (Excel random number generator). Block randomization with a fixed block size of four was applied to ensure balanced group sizes throughout the study. Stratification was performed by sex and baseline Cobb angle severity to minimize confounding effects. Allocation concealment was maintained using sequentially numbered, opaque, sealed envelopes prepared by an independent researcher not involved in enrollment, intervention, or assessment.

Because of the nature of the interventions, therapists and participants were not blinded. To minimize detection bias, outcome assessors for radiographic Cobb angle (the orthopedist) and all neuroimaging analysts were blinded to group allocation and timepoint. MRI/DTI datasets and radiographs were anonymized with study IDs and processed in random order.

Statistical Analysis

All statistical analyses in this study were performed using R software (version 4.4.1) and MedCalc Statistical Software (version 23). Prior to analysis, the normality of distribution for all continuous variables was assessed using the Shapiro-Wilk test, and the homogeneity of variances was evaluated using Levene’s test where appropriate. Descriptive statistics are presented as mean and standard deviation (SD). The significance level (α) was set at 0.05 for all statistical hypothesis tests.

A one-way Analysis of Variance (ANOVA) was used to compare baseline demographic characteristics (age, height, weight) and initial curve severity among the groups to ensure baseline homogeneity.

The change in the degree of curvature was evaluated using two complementary statistical models. First, to examine the interaction between time and treatment, a Mixed Measures Analysis of Variance (ANOVA) was applied. In this model, ‘time’ (Pre-vs Post-treatment) was the within-subjects factor and ‘treatment group’ (Schroth vs Traditional) was the between-subjects factor, with a focus on the time*group interaction. To assess the efficacy of the interventions, the primary confirmatory analysis was conducted using a one-way Analysis of Covariance (ANCOVA) restricted to the two patient groups (SEG and TEG). In this model, post-treatment Cobb angle was the dependent variable, treatment group was the independent variable, and pre-treatment Cobb angle was included as a covariate to adjust for baseline severity. This analysis provided the adjusted between-group difference in change (ΔΔ). Secondary comparisons between the patient groups and healthy controls were performed using a separate one-way ANOVA on post-treatment Cobb angles without covariate adjustment, as healthy controls lack baseline spinal deformity. Post hoc tests were conducted using the Holm-Bonferroni correction.

To compare regional cortical thickness and tractography metrics among the groups, a two-step analytical approach was employed. First, an initial exploratory screening was conducted using a series of one-way ANOVAs for each neuroanatomical variable. Variables that showed a significant unadjusted group difference (P < .05) in this screening were advanced to a second, more rigorous confirmatory step. In this second step, a one-way Analysis of Covariance (ANCOVA) was performed on these selected variables to determine if the group differences remained significant after statistically controlling for the potential confounding effects of age and intracranial volume (ICV), which were included as covariates. For ANCOVA models that confirmed a significant main effect of group, post hoc pairwise comparisons were conducted using the Holm-Bonferroni correction. To control the False Discovery Rate (FDR) across the multiple brain regions confirmed as significant in the ANCOVA step, the resulting P-values for the main group effect were adjusted using the Benjamini-Hochberg procedure.

Finally, to investigate the potential link between brain structure and clinical outcomes, a series of exploratory partial correlation analyses were conducted. Within each treatment group, these analyses examined the association between post-treatment neuroanatomical metrics and the change in Cobb angle (ΔCobb), while controlling for baseline Cobb angle, age, and ICV. Given their hypothesis-generating nature, P-values from these analyses were not adjusted for multiple comparisons.

Results

A total of 52 potential participants were assessed for eligibility, of whom 16 were excluded (10 did not meet inclusion criteria, and 6 declined to participate). The study was ultimately completed with 18 participants in each of the three groups: a healthy Control Group, the Schroth Exercise Group (SEG), and the Traditional Exercise Group (TEG). Although all enrolled participants completed the study (100% retention), the mean adherence to the prescribed exercise sessions was 91.4% ± 6.8 as detailed in the Methods section. A CONSORT flow diagram summarizes participant recruitment, allocation, follow-up, and analysis (Figure 3)

Figure 3.

Study flow chart

As detailed in Table 1, the three groups were homogeneous at baseline. A one-way ANOVA confirmed no statistically significant differences in age (F = 0.119, P = 0.887), height (F = 0.197, P = 0.822), or weight (F = 0.094, P = 0.910). Furthermore, the two treatment groups were well-matched in terms of their clinical presentation, with no significant difference in pre-treatment Cobb angle severity (SEG Mean = 22.11° ± 2.47 vs TEG Mean = 21.89° ± 2.83; F = 0.063, P = 0.803) (Table 2).

Table 1.

Baseline Demographic Characteristics of the Groups

	Count	Control group		Group				P
		Control group		TEG		SEG
		Mean ± SD	CI 95%	Mean ± SD	CI 95%	Mean ± SD	CI 95%
Height (cm)	18	160.44 ± 6.97	(156.98,163.91)	159.11 ± 12.34	(152.97,165.25)	161.00 ± 4.49	(158.77,163.23)	0.822^λ
Weight (kg)	18	49.67 ± 8.12	(45.63,53.70)	49.06 ± 8.62	(44.77,53.34)	48.61 ± 6.19	(45.53,51.69)	0.910^λ
Age (years)	18	15.78 ± 1.44	(15.06,16.49)	16.06 ± 1.89	(15.11,17.00)	15.89 ± 1.81	(14.99,16.79)	0.887^λ
Risser stage	18	2.87 ± 1.27		2.58 ± 0.8		2.72 ± 1.15		0.940^λ
Pre-treatment Cobb angle (major curve) (°)	18	-		21.89 ± 2.83		22.11 ± 2.47		NA

SD: Standard deviation, SEG: Schroth Exercise Group, TEG: Traditional Exercise Group, ^λ: One-way analysis of variance test.

Table 2.

Comparison of Curve Severity and Group × Time Interactions Between Treatment Groups

	Group		Within-group^ϕ (Pre-treatment vs Post-treatment)				Between group (Schroth vs traditional)
	SEG	TEG	Within-group^ϕ (Pre-treatment vs Post-treatment)				Between group (Schroth vs traditional)
	Mean ± SD	Mean ± SD	Group	F	P	η²	Time	F	P	η²
Pre-treatment Cobb angle (major curve)	22.11^a ± 2.47	21.89^a ± 2.83	SEG	221.53	<0.001	0.929	Pre-treatment^ω	0.063	0.803	0.105
Post-treatment Cobb angle (major curve)	15.28^x ± 1.84	23.22^y ± 2.53	TEG	7.556	0.014	0.308	Post-treatment^ω	115.84	<0.001	0.844
Difference	6.83^κ ± 1.95	−1.33^ψ ± 2.06					Difference^δ	149.513	<0.001	0.815

Abbreviations: SD, Standard deviation, SEG, Schroth Exercise Group, TEG, Traditional Exercise Group, ^ω: One-way analysis of variance test, ^δ: Mixed measures analysis of variance test, ^ϕ: Bonferroni test with adjustment for multiple comparisons, η²: Effect size. F: Analysis of variance test statistic, Bold values indicate p<0.05.

The effect of the interventions on spinal curvature was assessed using a complementary two-model approach. First, a Mixed Measures ANOVA was conducted to formally test for an interaction between time and treatment group (Table 2). This analysis revealed a highly significant time × group interaction (F(1, 34) = 149.513, P < .001, $η_{p}^{2}$ = 0.815), confirming that the trajectory of change in Cobb angle over the four-month period was significantly different between the SEG and TEG. While the SEG experienced a substantial and clinically meaningful decrease in curvature (Mean Difference = 6.83°), the TEG, in contrast, showed a slight but statistically significant worsening of their curvature (Mean Difference = −1.33°).

To precisely quantify the treatment efficacy, a primary ANCOVA restricted to the two patient groups was performed, controlling for baseline Cobb angle. This analysis confirmed that the Schroth group had a significantly lower adjusted post-treatment Cobb angle than the Traditional group. The adjusted between-group difference in change (ΔΔ) was −8.07° (95% CI [-9.20, −6.94], P < .001), indicating a very large effect size (Cohen’s d = −4.87) in favor of the Schroth method (Table 3). In secondary comparisons using a separate one-way ANOVA to include the healthy control group, both the SEG (Mean = 15.30°) and the TEG (Mean = 23.22°) remained significantly different from the healthy controls (0° curvature) at post-treatment (P < .001), as expected given the structural nature of the condition.

Table 3.

Analysis of Post-treatment Cobb Angle: Treatment Efficacy and Comparison With Healthy Controls

Comparison	Statistical model	Adjusted/Mean difference (Δ)	95% CI	Effect size (Cohen’s d)	P
Treatment efficacy
SEG vs TEG	ANCOVA^a	−8.07°	[−9.20, −6.94]	−4.87	<.001
Comparison with controls
SEG vs healthy control	ANOVA^b	15.30°	[13.20, 17.40]	-	<.001
TEG vs healthy control	ANOVA^b	23.22°	[20.72, 25.72]	-	<.001

Abbreviations: SEG, Schroth Exercise Group; TEG, Traditional Exercise Group.

^aPrimary efficacy analysis: Performed using ANCOVA on post-treatment Cobb angle, adjusted for baseline Cobb angle, restricted to the two patient groups (SEG and TEG). The 95% confidence interval for the difference is calculated as the mean difference ± 1.96 * SE (0.553).

^bSecondary comparison: Performed using one-way ANOVA on post-treatment Cobb angle including all three groups, without adjustment for baseline values (as healthy controls lack baseline deformity). Mean difference indicates the raw difference from the healthy control mean (0°).

A two-step analytical approach was used to investigate group differences in brain structure. An initial screening with one-way ANOVA identified 11 brain regions with significant unadjusted group differences in cortical thickness. These variables were subsequently entered into a more rigorous confirmatory ANCOVA, controlling for age and intracranial volume (ICV).

As shown in Table 4, after covariate adjustment, all 11 regions maintained statistically significant group differences. For instance, the main effect of group on the right medial frontal cortex was particularly strong (F = 18.39, P < .001,

η_{p}^{2}

= 0.429). A remarkably consistent pattern emerged from the post hoc analyses across all significant regions: the TEG consistently exhibited significantly thinner cortex compared to both the SEG and the healthy controls. Crucially, in all 11 of these sensorimotor-related regions, no significant differences in cortical thickness were found between the SEG and the healthy control group, suggesting a structural profile comparable to that of healthy individuals in the Schroth group post-intervention.

Table 4.

Covariate-Adjusted Comparisons of Regional Cortical Thickness Among Groups

Brain region (mm)	Groups			F	P	$η_{p}^{2}$
	Control group	TEG	SEG
	Mean ± SD	Mean ± SD	Mean ± SD
Right medial frontal	3.93^a ± .17	3.51^b ± .25	3.91^a ± .26	18.39	<.001	0.429
Left medial frontal	3.85^a ± .18	3.56^b ± .28	3.83^a ± .26	7.64	0.001	0.238
Left precentral gyrus	2.37^a ± .29	2.09^b ± .28	2.35^a ± .18	6.43	0.003	0.208
Right parietal	2.67^a ± .28	2.45^b ± .14	2.65^a ± .15	6.14	0.004	0.200
Left parietal	2.64^a ± .30	2.42^b ± .17	2.62^a ± .14	5.29	0.008	0.178
Right postcentral gyrus	1.73^a ± .22	1.52^b ± .16	1.71^a ± .15	7.31	0.002	0.230
Right medial postcentral gyrus	1.60^a ± .33	1.29^b ± .31	1.58^a ± .28	5.66	0.006	0.188
Left medial postcentral gyrus	1.56^a ± .27	1.23^b ± .26	1.54^a ± .26	9.84	<.001	0.287
Right occipital	2.74^a ± .23	2.47^b ± .26	2.73^a ± .22	6.86	0.002	0.219
Left occipital	2.73^a ± .20	2.46^b ± .20	2.71^a ± .22	8.67	<.001	0.261
Left occipital fusiform gyrus	2.78^a ± .28	2.49^b ± .24	2.77^a ± .29	5.89	0.005	0.194

Note. Superscript letters (^a, ^b) in the Mean ± SD columns indicate the results of post hoc pairwise comparisons between groups. Within the same row, means not sharing a common letter are significantly different from each other based on Holm-adjusted post hoc tests (P < .05).

The F, P, and $η_{p}^{2}$ values represent the main effect of the group from ANCOVA, with age and intracranial volume (ICV) included as covariates.

Similarly, covariate-adjusted analyses of brain tractography metrics revealed significant group differences in the microstructure of the right corticospinal tract (Table 5). Specifically, the TEG showed a significantly lower fiber count (F = 4.35, P = .018) and fiber ratio (F = 4.27, P = .019) compared to the other groups. Post hoc tests confirmed that the values for the Schroth group in these metrics were statistically indistinguishable from those of the healthy controls, while being significantly higher than those of the TEG. No other tractography metrics showed significant group differences after statistical adjustment. The results of Tukey HSD and Games-Howell post-hoc analyses performed to determine which groups were different in the regions where significant differences were found are given comprehensively in Figure 4.

Table 5.

Comparative Analysis of Brain Tractography Metrics Among the Groups

		Groups			p^λ
		Control group	TEG	SEG
		Mean ± SD	Mean ± SD	Mean ± SD
Whole-brain tractography metrics	Fiber count	85 373.72 ± 282.37	85 476.22 ± 1448.82	85 618.72 ± 1048.34	0.845
	Mean fiber length (mm)	647.04 ± 19.96	653.68 ± 17.92	659.67 ± 19.42	0.151
	FA	.37 ±.01	.37 ±.02	.36 ±.01	0.866
Left corticospinal tract	Fiber count	5481.17 ± 174.73	4790.39 ± 196.49	5648.33 ± 194.56	0.130
	Mean fiber length (mm)	118.91 ± 6.75	117.90 ± 6.16	120.25 ± 6.56	0.557
	FA	.52 ±.02	.53 ±.03	.53 ±.03	0.331
	Fiber ratio	6.42 ± 1.48	5.61 ± 1.64	6.60 ± 1.53	0.133
Right corticospinal tract	Fiber count	5069.56^a,b ± 175.76	3851.67^a ± 109.62	5545.33^b ± 172.85	<0.001 ( $η_{p}^{2}$ = 0.163)
	Fiber count	5069.56^a,b ± 175.76	3851.67^a ± 109.62	5545.33^b ± 172.85	F = 4.35; P = 0.018; $η_{p}^{2}$ = 0.151
	Mean fiber length (mm)	121.06 ± 5.04	121.21 ± 5.18	122.07 ± 7.23	0.857
	FA	.53 ±.02	.53 ±.02	.54 ±.02	0.331
	Fiber ratio	5.94^a,b ± 1.72	4.51^b ± 1.89	6.48^a ± 2.20	<0.001( $η_{p}^{2}$ = 0.161)
	Fiber ratio	5.94^a,b ± 1.72	4.51^b ± 1.89	6.48^a ± 2.20	F = 4.27; P = 0.0.019; $η_{p}^{2}$ = 0.148
Left medial lemniscus	Fiber count	3061.89 ± 136.97	2493.50 ± 168.04	2816.67 ± 143.71	0.036
	Mean fiber length (mm)	134.06 ± 5.38	131.51 ± 6.08	135.36 ± 6.62	0.162
	FA	.46 ±.03	.47 ±.02	.47 ±.02	0.300
	Fiber ratio	3.58 ± 0.21	2.92 ± 1.02	3.37 ±.70	0.045
Right medial lemniscus	Fiber count	3840.44 ± 178.77	3087.94 ± 203.26	3516.39 ± 187.11	0.042
	Mean fiber length (mm)	129.61 ± 6.46	129.50 ± 4.39	131.77 ± 6.75	0.441
	FA	.49 ±.01	.49 ±.02	.49 ±.02	0.913
	Fiber ratio	4.50 ± 1.02	3.61 ± 1.28	4.11 ± 1.04	0.040

Abbreviations: SD, Standard deviation, SEG, Schroth Exercise Group, TEG, Traditional Exercise Group, ^λ: One-way analysis of variance test, $η^{p}$ : Effect size.

Note: Superscript letters (^a,^b) denote significant post hoc differences based on ANCOVA with Holm-correction. For significant findings, the top line shows the unadjusted ANOVA (P, $η_{p}^{2}$ ), while the bottom shows the primary ANCOVA (F, P, $η_{p}^{2}$ ) adjusted for age and ICV. P-values for the ANCOVA group effect were corrected for multiple comparisons using the Benjamini-Hochberg FDR procedure.

Figure 4.

(A, B) Multiple comparison results of significant factors across different treatment groups

Finally, exploratory partial correlation analyses were conducted to examine whether post-treatment brain structural measures were associated with the magnitude of clinical improvement (ΔCobb) (Table 6). No significant associations were detected in either treatment group (P > .05).

Table 6.

Partial Correlation Analyses Linking Neuroanatomical Measures to Change in Cobb Angle (ΔCobb)

Brain region (mm) and tractography metrics	Group
	SEG		TEG
	r	P	r	P
Right medial frontal	0.194	0.488	−0.151	0.592
Left medial frontal	0.020	0.945	0.083	0.770
Left precentral gyrus	−0.002	0.996	0.124	0.659
Right parietal	−0.131	0.641	−0.139	0.621
Left parietal	−0.256	0.357	−0.026	0.926
Right postcentral gyrus	−0.241	0.387	−0.024	0.933
Right medial postcentral gyrus	−0.419	0.120	−0.123	0.661
Left medial postcentral gyrus	−0.486	0.065	−0.139	0.622
Right occipital	−0.008	0.978	0.086	0.760
Left occipital	−0.053	0.851	−0.059	0.836
Left occipital fusiform gyrus	0.214	0.443	0.392	0.148
Right corticospinal tract (Fiber count)	0.271	0.329	0.015	0.958
Right corticospinal tract (Fiber ratio)	0.270	0.331	0.007	0.980

Discussion

The present study demonstrated that Schroth exercises were more effective than conventional exercises in improving curve severity in individuals with AIS. In addition, beyond musculoskeletal outcomes, our findings suggest neural changes consistent with adaptive plasticity in the central nervous system in the Schroth group.

Notably, increased cortical thickness was observed in several brain regions, including the medial frontal gyrus, precentral gyrus, fusiform gyrus, parietal lobe, postcentral gyrus, occipital lobe, and occipital fusiform gyrus. Alterations were also detected in motor pathways such as the corticospinal tract. These results suggest exploratory associations between scoliosis-specific exercise protocols and both musculoskeletal and neural integration processes; however, further evidence is needed to confirm the durability and causality of such effects. Because imaging was performed only post-intervention, the observed structural differences should be interpreted as associative rather than causal or longitudinal evidence of neuroplasticity. These neuroimaging results should be interpreted as hypothesis-generating findings that support further longitudinal work with pre–post designs and larger samples.

In AIS management, Cobb angle is regarded as the primary indicator of curve progression.²⁰ Exercise-based interventions have emerged as essential components in scoliosis treatment. Prior studies have shown the potential of Schroth exercises to reduce Cobb angle, stabilize curvature, and lower the need for surgical intervention.^21-23 Consistent with this literature, our study also demonstrated a significant reduction in Cobb angle following Schroth intervention.^7,24

Previous neuroimaging studies in individuals with AIS have reported cortical thinning in the frontal, parietal, and occipital lobes.^3,6 One of the most noteworthy exploratory findings was that cortical thickness in the Schroth group trended toward levels seen in healthy controls; however, no formal equivalence testing was performed, so this cannot be interpreted as normalization. Nevertheless, as no formal equivalence testing was performed, these results should not be interpreted as statistical equivalence. Rather, they point to a possible restorative effect that requires confirmation in future research.

The frontal lobe is responsible for cognitive control, motor planning, and executive functioning. In particular, the primary motor cortex located in the precentral gyrus plays a key role in voluntary movement coordination.^25,26 Activation of this area with repetitive motor tasks, such as Schroth exercises, may contribute to improved postural control, which in turn may trigger neuroplasticity in the motor cortex, strengthening postural control mechanisms.¹⁶ While SEG revealed increased fiber counts and fiber ratios in the corticospinal tracts of participants, these measurements are highly sensitive to tractography algorithms, seeding density, and stopping criteria and should therefore be interpreted with caution. Our confirmatory results are primarily based on diffusion tensor metrics (FA), which more directly reflect white matter microstructure.

Similarly, the parietal lobe plays a central role in somatosensory integration and proprioceptive awareness.²⁷ Schroth exercises may enhance central nervous system function and postural control by providing continuous stimulation to proprioceptive receptors, including muscle spindles, joint receptors, and cutaneous mechanoreceptors.^24,28,29 Cortical thickening in this area could reflect adaptive changes in proprioceptive processing. However, no significant correlations were identified between neuroimaging metrics and improvement in Cobb angle, indicating that these structural features cannot yet be interpreted as mechanistic drivers of clinical change.

Visual inputs also contribute to postural regulation, and the use of mirrors during Schroth exercises may enhance visual cortex activation.¹⁰ The increased cortical thickness observed in the occipital lobe may be related to visual-motor coordination³⁰; however, motivational factors, learning effects, and baseline activity levels could also partly explain these changes.

Neuroplasticity is broadly defined as structural and functional brain modifications in response to repetitive motor and sensory inputs.³¹ The repetitive, goal-directed, sensorimotor tasks inherent to Schroth exercises likely support this process. While the patterns observed here are consistent with this concept, causality cannot be established due to the short-term nature of the intervention and lack of longitudinal follow-up.

The literature also indicates a positive relationship between physical activity participation and brain volume and cognitive performance in children.³² Improvements in frontal lobe volume have been associated with enhanced working memory, self-awareness, and social cognition. These findings underscore the need for multidisciplinary approaches in AIS management, addressing both musculoskeletal deformities and cognitive-motor deficits.

Our results further highlight the limitations of conventional treatment methods in reducing spinal curvature in AIS. Consistent with existing literature, scoliosis-specific exercises have been found to produce superior outcomes in curve magnitude, trunk deformity, postural symmetry, and pain compared to general exercise programs.^33-36 This reinforces the importance of incorporating scoliosis-specific protocols into clinical practice.

This study has several limitations that should be acknowledged. First, the sample size was relatively small, which may limit the statistical power and generalizability of the findings. Second, the intervention period lasted only four months, making it difficult to determine the long-term sustainability of the observed musculoskeletal and neuroimaging changes. Third, the inclusion criteria were narrow because only individuals with Lenke Type 1 right thoracic curve and mild to moderate scoliosis (Cobb angle 15-25°) were included. Consequently, the applicability of the findings to individuals with other curve types or more severe curves is limited. Fourth, all interventions and follow-up assessments were performed by a single physical therapist, raising the possibility of therapist-related effects that cannot be dissociated from the treatment itself. Fifth, neuroimaging comparisons were primarily conducted between the post-intervention groups without comprehensive pre-post analyses within each group, limiting conclusions about the course of changes within the group. Considering these factors together, future studies with larger, more diverse samples, longer follow-up periods, multiple providers, and pre-post imaging designs, including within-group analyses and non-inferiority tests, are necessary to determine causal links and to confirm whether these adaptations reflect enduring neuroplastic processes rather than transient or confounding effects.

Clinical Implications

The present findings carry potential clinical implications for the non-operative management of AIS. Structural changes observed in motor and somatosensory regions, along with improved integrity of key white matter pathways, accompanied reductions in Cobb angle following Schroth-based rehabilitation. Although direct correlations between neural findings and clinical improvement were not demonstrated, these adaptations may reflect sensorimotor engagement associated with scoliosis-specific exercises rather than mechanistic determinants of treatment response. Such neural characteristics may serve as exploratory biomarkers that inform treatment monitoring and support individualized, early-stage intervention strategies. Identifying neural signatures associated with favorable response to conservative rehabilitation may help optimize non-operative management and potentially reduce surgical need in select patients. However, longitudinal studies incorporating pre–post imaging are required to validate these associations and to determine whether they represent enduring adaptive processes with clinical utility.

Our findings demonstrate that Schroth exercises produced significantly greater short-term improvements in Cobb angle compared with traditional exercises in adolescents with mild AIS (Lenke type 1) over a 4-month period. While this highlights the clinical value of Schroth-based interventions, the durability and generalizability of these results are limited by the study’s focus on mild curves, a single Lenke type, and treatment delivered by a single therapist. Neuroimaging findings should be interpreted as convergent, associative signals consistent with sensorimotor engagement rather than causal evidence of neuroplasticity. Future studies employing pre- and post-treatment longitudinal imaging are required to establish neuroplastic mechanisms underlying clinical change. Additionally, the non-significant difference between the Schroth group and healthy controls should not be interpreted as statistical equivalence, as no formal equivalence testing (e.g., TOST) was performed.

Conclusion

In conclusion, Schroth exercises resulted in greater improvement in Cobb angle compared with traditional exercises in adolescents with mild AIS. Post-intervention cortical and white matter characteristics in the Schroth group were comparable to those of healthy peers, although no formal equivalence testing was performed. These findings suggest that scoliosis-specific rehabilitation may be associated with structural adaptations in sensorimotor pathways. However, these neural features should be interpreted as exploratory and associative rather than causal evidence of neuroplastic change. Future studies incorporating pre–post imaging and larger sample sizes are required to determine whether these structural patterns reflect enduring adaptive processes with clinical relevance for non-operative management of AIS.

Footnotes

ORCID iDs

Ahmet Payas

Mustafa Arık

Ethical Considerations

The Kayseri City Training and Research Hospital Clinical Research Ethics Committee granted approval for this study (31.10.2023/number: 932). Kayseri/Turkiye.

Consent to Participate

Written informed consent was obtained from each patient. The study was conducted in accordance with the principles of the Declaration of Helsinki.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Appendix

References

Dimitrijević

Šćepanović

Jevtić

, et al. Application of the schroth method in the treatment of idiopathic scoliosis: a systematic review and meta-analysis. Int J Environ Res Public Health. 2022;19(24):16730. doi:10.3390/ijerph192416730

Cheng

Castelein

Chu

, et al. Adolescent idiopathic scoliosis. Nat Rev Dis Primers. 2015;1:15030. doi:10.1038/nrdp.2015.30

Payas

Kocaman

Yıldırım

Batın

. Prediction of adolescent idiopathic scoliosis with machine learning algorithms using brain volumetric measurements. JOR Spine. 2024;7(3):e1355. doi:10.1002/jsp2.1355

Payas

Çiçek

Ekinci

, et al. Tractography analysis results of the trigeminus nerve, which contains fibers responsible for proprioception sensation and motor control in adolescent idiopathic scoliosis. Eur Spine J. 2024;33(12):4702-4709. doi:10.1007/s00586-024-08524-y

Peng

Wang

Qiu

Zhang

Zhuang

. Research progress on the etiology and pathogenesis of adolescent idiopathic scoliosis. Chin Med J. 2020;133(4):483-493. doi:10.1097/CM9.0000000000000652

Wang

Shi

Chu

Burwell

Cheng

Ahuja

. Abnormal cerebral cortical thinning pattern in adolescent girls with idiopathic scoliosis. Neuroimage. 2012;59(2):935-942. doi:10.1016/j.neuroimage.2011.07.097

Ceballos Laita

Tejedor Cubillo

Mingo Gómez

Jiménez Del Barrio

. Effects of corrective, therapeutic exercise techniques on adolescent idiopathic scoliosis. A systematic review. Efectos de las técnicas de ejercicio terapéutico correctivo en la escoliosis idiopática del adolescente. Revisión sistemática. Arch Argent Pediatr. 2018;116(4):e582-e589. doi:10.5546/aap.2018.eng.e582

Dimitrijević

Viduka

Šćepanović

, et al. Effects of Schroth method and core stabilization exercises on idiopathic scoliosis: a systematic review and meta-analysis. Eur Spine J. 2022;31(12):3500-3511. doi:10.1007/s00586-022-07407-4

Tan

Moe

Vaithinathan

Wong

. Curve progression in idiopathic scoliosis: follow-up study to skeletal maturity. Spine (Phila Pa 1976). 2009;34(7):697-700. doi:10.1097/BRS.0b013e31819c9431

10.

You

, et al. Effectiveness of physiotherapeutic scoliosis-specific exercises on 3-Dimensional spinal deformities in patients with adolescent idiopathic scoliosis: a systematic review and meta-analysis. Arch Phys Med Rehabil. 2024;105(12):2375-2389. doi:10.1016/j.apmr.2024.04.011

11.

Kocaman

Bek

Kaya

Büyükturan

Yetiş

Büyükturan

. The effectiveness of two different exercise approaches in adolescent idiopathic scoliosis: a single-blind, randomized-controlled trial. PLoS One. 2021;16(4):e0249492. doi:10.1371/journal.pone.0249492

12.

Legault

Litcofsky

. Neuroplasticity as a function of second language learning: anatomical changes in the human brain. Cortex. 2014;58:301-324. doi:10.1016/j.cortex.2014.05.001

13.

Kurtoğlu

Payas

Düz

, et al. Analysis of changes in brain morphological structure of taekwondo athletes by diffusion tensor imaging. J Chem Neuroanat. 2023;129:102250. doi:10.1016/j.jchemneu.2023.102250

14.

Qiu

Mori

Miller

. Diffusion tensor imaging for understanding brain development in early life. Annu Rev Psychol. 2015;66:853-876. doi:10.1146/annurev-psych-010814-015340

15.

Slattery

Verma

. Classifications in brief: the lenke classification for adolescent idiopathic scoliosis. Clin Orthop Relat Res. 2018;476(11):2271-2276. doi:10.1097/CORR.0000000000000405

16.

Payas

Batin

Kurtoğlu

, et al.

Is the integration problem in the sensoriomotor system the cause of adolescent idiopathic scoliosis?

J Pediatr Orthop. 2023;43(2):e111-e119. doi:10.1097/BPO.0000000000002300

17.

Koussis

Toulas

Glotsos

Lamprou

Kehagias

Lavdas

. Reliability of automated brain volumetric analysis: a test by comparing NeuroQuant and volBrain software. Brain Behav. 2023;13(12):e3320. doi:10.1002/brb3.3320

18.

Vetter

Köhler

Hepp

, et al. Diffusion tensor imaging: influence of segmentation on fiber tracking in the supraspinatus muscle-An inter-operator reliability analysis. PLoS One. 2023;18(9):e0286280. doi:10.1371/journal.pone.0286280

19.

Hakulinen

Brander

Ryymin

, et al. Repeatability and variation of region-of-interest methods using quantitative diffusion tensor MR imaging of the brain. BMC Med Imaging. 2012;12:30. doi:10.1186/1471-2342-12-30

20.

Schreiber

Parent

Moez

, et al. The effect of Schroth exercises added to the standard of care on the quality of life and muscle endurance in adolescents with idiopathic scoliosis-an assessor and statistician blinded randomized controlled trial: “SOSORT 2015 Award Winner”. Scoliosis. 2015;10:24. doi:10.1186/s13013-015-0048-5

21.

Dimitrijević

Rašković

Popović

, et al. Treatment of adolescent idiopathic scoliosis with the conservative schroth method: a randomized controlled trial. Healthcare. 2025;13(6):688. doi:10.3390/healthcare13060688

22.

Horne

Flannery

Usman

. Adolescent idiopathic scoliosis: diagnosis and management. Am Fam Physician. 2014;89(3):193-198.

23.

Otman

Kose

Yakut

. The efficacy of schroth`s 3-dimensional exercise therapy in the treatment of adolescent idiopathic scoliosis in Turkey. Neurosciences (Riyadh). 2005;10(4):277-283.

24.

Park

Jeon

Park

. Effects of the Schroth exercise on idiopathic scoliosis: a meta-analysis. Eur J Phys Rehabil Med. 2018;54(3):440-449. doi:10.23736/S1973-9087.17.04461-6

25.

Kuru

Yeldan

Dereli

Özdinçler

Dikici

Çolak

. The efficacy of three-dimensional Schroth exercises in adolescent idiopathic scoliosis: a randomised controlled clinical trial. Clin Rehabil. 2016;30(2):181-190. doi:10.1177/0269215515575745

26.

Lemon

. Descending pathways in motor control. Annu Rev Neurosci. 2008;31:195-218. doi:10.1146/annurev.neuro.31.060407.125547

27.

Jang

. The role of the corticospinal tract in motor recovery in patients with a stroke: a review. NeuroRehabilitation. 2009;24(3):285-290. doi:10.3233/NRE-2009-0480

28.

Whitlock

. Posterior parietal cortex. Curr Biol. 2017;27(14):R691-R695. doi:10.1016/j.cub.2017.06.007

29.

Akuthota

Ferreiro

Moore

Fredericson

. Core stability exercise principles. Curr Sports Med Rep. 2008;7(1):39-44. doi:10.1097/01.CSMR.0000308663.13278.69

30.

Larni

Mohsenifar

Ghandhari

Salehi

. The effectiveness of schroth exercises added to the brace on the postural control of adolescents with idiopathic scoliosis: case series. Ann Med Surg. 2022;84:104893. doi:10.1016/j.amsu.2022.104893

31.

Batın

Uçar

Payas

, et al.

Can adolescent idiopathic scoliosis be associated with visual impairment?

J Clin Pract Res. 2023;45(4):385-391.

32.

Taube

Gruber

Gollhofer

. Spinal and supraspinal adaptations associated with balance training and their functional relevance. Acta Physiol (Oxf). 2008;193(2):101-116. doi:10.1111/j.1748-1716.2008.01850.x

33.

Chaddock-Heyman

Erickson

Holtrop

, et al. Aerobic fitness is associated with greater white matter integrity in children. Front Hum Neurosci. 2014;8:584. doi:10.3389/fnhum.2014.00584

34.

Muthukrishnan

Shenoy

Jaspal

Nellikunja

Fernandes

. The differential effects of core stabilization exercise regime and conventional physiotherapy regime on postural control parameters during perturbation in patients with movement and control impairment chronic low back pain. Sports Med Arthrosc Rehabil Ther Technol. 2010;2:13. doi:10.1186/1758-2555-2-13

35.

Gür

Ayhan

Yakut

. The effectiveness of core stabilization exercise in adolescent idiopathic scoliosis: a randomized controlled trial. Prosthet Orthot Int. 2017;41(3):303-310. doi:10.1177/0309364616664151

36.

Seleviciene

Cesnaviciute

Strukcinskiene

Marcinowicz

Strazdiene

Genowska

. Physiotherapeutic scoliosis-specific exercise methodologies used for conservative treatment of adolescent idiopathic scoliosis, and their effectiveness: an extended literature review of current research and practice. Int J Environ Res Public Health. 2022;19(15):9240. doi:10.3390/ijerph19159240