Clinical and Biomechanical Effects of AI-Supervised Exercise Telerehabilitation on Patients with Nonspecific Chronic Low Back Pain: A Single-arm Clinical Trial

Clinical and Biomechanical Effects of AI-Supervised Exercise Telerehabilitation on Patients with Nonspecific Chronic Low Back Pain: A Single-arm Clinical Trial | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Clinical and Biomechanical Effects of AI-Supervised Exercise Telerehabilitation on Patients with Nonspecific Chronic Low Back Pain: A Single-arm Clinical Trial Pengcheng Lu, Zhuodong Zhang, Jingjing Hu, Qing Zeng, Tao Fan, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7344603/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Nonspecific chronic low back pain (NSCLBP) is a major cause of disability worldwide, and the long-term efficacy of traditional treatments is limited. Although highly recommended, core stabilization exercises are often limited by adherence and access to supervised programs. Recent advances in artificial intelligence (AI) have created new opportunities for remote, real-time supervision of exercise. This study aims to evaluate the clinical and biomechanical effects of a remote AI-supervised virtual exercise program (RAVEP) on patients with NSCLBP. Methods Fourteen patients with NSCLBP underwent a four-week RAVEP intervention, which was delivered via a smartphone application that integrated human key-point recognition and real-time feedback. The program incorporated breathing, core strengthening, and McKenzie exercises. The effects before and after the intervention were assessed using ultrasound imaging of the lumbar multifidus (LM) and transversus abdominis (TrA), gait analysis using 3D motion capture and surface electromyography, and musculoskeletal modeling using a full-body lumbar spine model in OpenSim. The model was scaled using patient-specific anthropometric data, and the muscle properties were calibrated using ultrasound-derived muscle thickness to assess personalized lumbar kinematics, intervertebral loads, and muscle forces and activations. In addition, the evaluation of pain and clinical function was conducted using the Visual Analog Scale (VAS), Oswestry Disability Index (ODI), Roland-Morris Disability Questionnaire (RMDQ), and Timed Up and Go (TUG) tests. Results Following the RAVEP, patients with NSCLBP showed significant improvements in both pain (VAS) and clinical function (ODI, RMDQ, TUG) (p < 0.01). Ultrasound assessment revealed increased LM and TrA thickness, as well as contraction ratios after the intervention. The results of the OpenSim simulation indicated significant increases in LM force (p < 0.05), and decreased compressive and torsional loads at L4-L5 and L5-S1 during gait. Kinematic results showed more symmetrical lumbar movement. Increased activation of deep core muscles was found; however, the differences in activation did not reach statistical significance. Discussion Our study showed that the RAVEP could significantly improve the pain, functioning, and core muscle performance in NSCLBP patients. AI-supervised real-time feedback in the RAVEP helped patients to perform their home exercises properly, which improved lower back stability and lessened stress on the spine. Personized musculoskeletal modeling and simulation provide a new understanding of how virtual rehabilitation works. This study demonstrates that RAVEP may serve as an effective and accessible strategy for the management of NSCLBP. Nonspecific chronic low back pain AI-supervised virtual exercise Opensim Ultrasound imaging Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Low back pain (LBP) represents a major public health concern and is one of the leading causes of disability worldwide [ 1 ]. It affects approximately one in every thirteen individuals, according to 2020 reports indicating personal experiences with the condition [ 2 , 3 ]. We categorize LBP into specific and nonspecific types, with approximately 80 to 90% of cases being nonspecific. Non-specific chronic low back pain (NSCLBP) refers to discomfort persisting for over three months, not attributable to any identifiable disease or abnormal structural issue [ 5 ]. This condition has been shown to have a significant impact on individuals suffering from persistent pain, characterized by a rigid lower back that restricts mobility, impairs balance, and diminishes proprioception. The inadequate consideration of biomechanical factors in lumbar and lower-extremity muscle pathology in NSCLBP may hinder the development of effective interventions [ 7 ]. In current clinical practice, NSCLBP is often a diagnosis of exclusion due to the lack of precise etiological identification [ 5 ]. This ambiguity has led to controversies regarding the relative roles of biomechanical dysfunction, central sensitization, and psychological influences in pain persistence. Muscle pathology may be the initial alteration to spinal biomechanics and nociceptive input [ 6 ]. Prior research indicates that individuals with NSCLBP exhibit significantly reduced thickness changes in the transversus abdominis (TrA) and lumbar multifidus (LM) compared to asymptomatic individuals during voluntary tasks [ 8 , 9 ]. Additionally, intramuscular electromyography revealed a delayed anticipatory activation of TrA/LM fibers during trunk loading or limb movement [ 10 ]. These studies focus solely on local lumbar biomechanics, neglecting the influence of movements in other body regions on the lumbar spine, leading to an incomplete analysis of motion. The lower limb plays a critical role in spinal function and stability. The stability of the spine is affected by the lower limb joints, particularly the hip, which influences posture and movement patterns in patients with NSCLBP [ 11 – 14 ]. A study examined the potential benefits of incorporating hip strengthening exercises into conventional rehabilitation therapy for enhancing pain relief and reducing disability in individuals with low back pain [ 15 ]. Individuals experiencing acute low back pain exhibit altered coordination between the hip and spine during balance-related tasks, such as sitting on an unstable surface [ 15 ]. The altered coordination may influence the balance performance of the overall lower extremity, suggesting the necessity for targeted interventions to improve coordination and stability of the lumbar region and the overall lower extremity. Recent advancements in the full-body lumbar spine (FBLS) model of OpenSim provide a detailed approach to musculoskeletal modeling [ 17 ]. This helps us better understand the effects of various exercises on the lumbar region and the entire kinetic chain, which potentially improve the treatment strategies for patients with NSCLBP. Our previous study utilized the FBLS model to analyze the lumbar biomechanical changes during walking in patients with NSCLBP compared to healthy individuals, revealing a significant reduction in lumbar rotation angles, moments, muscle forces, and activation among the patient group . In the management of NSCLBP, analgesics are commonly used; however, their effectiveness is often restricted relative to placebo, and they may induce adverse effects [ 19 ]. Furthermore, other modalities such as superficial heat and cold, massage, acupuncture, and ultrasound may provide short-term relief; however, the evidence supporting their long-term effectiveness remains insufficient [ 20 ]. Exercise therapies are commonly recommended for NSCLBP patients and have shown moderate improvements in pain relief and functional capacity [ 20 ]. Recent studies have found that core strength training interventions effectively alleviated symptoms related to lumbar disc herniation, which may be due to optimized load distribution and reduced stress on the lumbar spine [ 21 ]. Improved strength and activation of deep core muscles may concurrently restore the physiological curvature of the lumbar spine, leading to an optimal load-bearing posture [ 22 ]. However, the effectiveness of exercise depends upon individual patient characteristics and preferences [ 22 ] indicating that a personalized approach may provide the best outcome. Achieving lasting therapeutic effects in CNSLBP requires consistent and ongoing supervised exercise interventions to ensure optimal results The accurate execution of exercise movements is crucial, as improper technique can reduce effectiveness and increase the risk of secondary injuries [ 24 ] Supervised exercise programs are advocated for NSCLBP patients due to their ability to ensure proper execution, facilitate individualized modifications, and offer immediate feedback on movement accuracy [ 23 ] Hospital-based supervised exercises encounter several barriers, including transportation difficulties, scheduling conflicts, financial limitations, and shortages of healthcare personnel [ 25 ] The challenges lead to an overloaded rehabilitation system, limiting access to essential services, especially for older adults [ 26 – 28 ]. Virtual rehabilitation offers exercise programs directly to patients’ homes, potentially reducing barriers to attendance and program completion [ 25 – 28 ]. Virtual exercises can yield clinical outcomes similar to in-clinic supervised exercise programs by effectively decreasing pain and improving lumbar function through core muscle strengthening. Video-based guidance faces challenges in delivering real-time motion corrections [ 27 , 29 – 31 ] Recent advancements in artificial intelligence (AI) and real-time motion capture technology[ 25 ] have facilitated the development of AI-supervised exercise training that employs a standard household tablet’s camera for dynamic tracking of patients’ movements. Our study uses the human key-point recognition technology to monitor the trajectory of exercises. It provides immediate feedback for posture correction and replicates the benefits of supervised in-clinic training, such as enhanced posture stability and injury prevention. The integration of AI into exercise programs to automate different training phases has the potential to improve the quality of home-based care. Therefore, this study explores whether early-stage personalized training that targets the abnormal biomechanics of the lumbar spine and lower limbs can effectively improve patients’ symptoms and abnormal biomechanics. This research aims to evaluate the efficacy of a remote AI-supervised virtual exercise program (RAVEP) in improving clinical outcomes and biomechanical metrics in individuals with NSCLBP. In addition to conventional scaling using anthropometric data, this study intended to adopt ultrasound imaging-derived muscle thickness to calibrate the hill-type muscle parameters of each patient in the FBLS model of OpenSim. Therefore, this personized FBLS model can accurately analysis of alterations in lumbar and lower limb kinematics, lumbar moments and loads, as well as muscle activation patterns during walking after the intervention. The study has the potential to improve our understanding of lumbar spine biomechanics following a virtually supervised exercise program and provides important theoretical insights for the remote management of NSCLBP. Methods Participants From July to December 2023, fourteen patients with NSCLBP (4 males, 10 females; mean age 25.71 ± 6.79 years; weight 56.60 ± 10.04 kg; height 162.64 ± 5.13 cm; BMI 21.32 ± 3.13) were recruited from Zhujiang Hospital, Southern Medical University. The Ethics Committee of Zhujiang Hospital, Southern Medical University approved the study (Approval Number: 2023-KY-017), and it was registered with the Chinese Clinical Trial Registry (Registration Number: ChiCTR2300073185). All participants provided written informed consent before participating in the study, which included consent for the publication of any potentially identifiable images or data. Inclusion criteria adhered to the diagnostic guidelines for NSCLBP by the American College of Physicians and the American Pain Society [ 32 ] Fourteen patients were enrolled based on the following criteria: (1) Participants provided written informed consent and voluntarily engaged in the study; (2) had a clinical diagnosis of NSCLBP lasting over three months with a Visual Analog Scale (VAS) score exceeding thirty; (3) reported no history of back trauma, lumbar fractures, lumbar spondylolisthesis, ankylosing spondylitis, spinal deformities, spinal tuberculosis, spinal tumors, sacroiliac joint dysfunction, lumbar degenerative changes, lumbar disc herniation, spinal stenosis, osteoporosis, or other conditions causing low back pain, and no history of lumbar or abdominal surgery in the past two years; (4) had not utilized nonsteroidal anti-inflammatory drugs (NSAIDs) or other medications in the past six months; and (5) were aged between 18 and 65 years old. The exclusion criteria were as follows: (1) inability to walk independently or the presence of an abnormal gait; (2) diagnosis of severe systemic diseases (e.g., acute infection, malignant tumors, cardiovascular disease, acute cerebrovascular events, mental disorders, significant visual or auditory impairments, or other conditions considered inappropriate for vigorous exercise); (3) pregnancy. The sample size was calculated utilizing the pwr.t.test function in R (version 4.3.2, released on October 31, 2023). Previous studies of remote intervention therapy indicated an assumed effect size of Hedge’s g = 0.88, with a statistical power of 1 − β = 0.81, and a significance level of α = 0.05. A paired-sample, two-tailed t-test revealed a requirement of at least 12 participants. The final sample size was adjusted to 14 participants to account for a 10% dropout rate. Interventions This study used a remote AI-supervised virtual exercise program (RAVEP) as an intervention. The program was delivered via the WeChat mini-program “RAVEP” developed by Beijing Yinshan Future Health Technology Co., Ltd. Participants accessed the program from their home or any other preferred location using a smartphone or tablet. The mini-program employs AI-driven image recognition to detect 32 critical body points, such as the head, torso, and limbs. This makes it enables precise tracking of exercise movements with continuous monitoring of the motion trajectories of the head, torso, and limbs. It checks for accuracy by the analysis of the angle and length ratios among key points. This mini-program has two main sections: “Assessment” and “Training”. Participants can access the “Assessment” section via a QR code. This allows them to complete a self-assessment that collects their medical history, NRS scores, ODI responses, and Roland-Morris questionnaire results. Participants automatically receive an evaluation report along with a tailored exercise prescription. The exercise selection in the “Training” section follows established protocols and aligns with the 2021 American Physical Therapy Association (APTA) Low Back Pain Guidelines [ 33 ] A standardized exercise prescription was implemented for all participants, encompassing a weekly regimen of breathing techniques, core muscle strengthening, flexibility training, and McKenzie exercises according to the abnormal biomechanics of each patient according to the results of OpenSim analysis. Prior to each session, participants engaged in a preparatory meeting online video meeting, where therapists provided detailed explanations of the forthcoming exercises and addressed essential topics relevant to NSCLBP education, such as posture, pain management, etiology, and lumbar anatomy. The intensity and difficulty of exercises progressively escalate from week 1 to week 4. Participants first identify a suitable location, access the RREP mini-program, select their assigned exercise course, and assume a position on a yoga mat, either standing or lying down, while placing their smartphone horizontally on a stand approximately 1.5 meters away. The program delivers audio instructions and displays instructional exercise videos, while an AI recognition frame in the top-right corner of the screen concurrently monitors motion in real time, offering guidance to ensure proper form (Fig. 1 ). Following each exercise, the program delivered an immediate performance rating based on key-point coincidence ratios, categorized as “Pass,” “Good,” “Excellent,” or “Outstanding” to promote patient adherence. At the end of each session, participants completed the rate of perceived exertion scale and received a summary report detailing their completion rates, thus enabling the evaluation of exercise efficacy. During the second to fourth weeks, participants conducted a follow-up assessment via the mini-program, leading to adjustments in their exercise prescriptions for those indicating discomfort. Evaluation Ultrasound measurement The Terason uSmart 3300 ultrasound instrument was utilized to measure the thickness of the TrA and LM muscles in both resting and contracting states [ 34 ] During the TrA measurement, patients were positioned supine with hips flexed to approximately 135 degrees and knees flexed to about 90 degrees. Participants were instructed to breathe in a calm manner. A linear probe was positioned above the iliac crest, anterior to the axillary line, to assess muscle thickness at the end of quiet inspiration, thereby inhibiting TrA activation and enabling the acquisition of the relaxed-TrA image. Participants were directed to fully exhale and progressively contract their abdomen as much as possible to obtain the activated-TrA image prior to the image acquisition process. Muscle thickness refers to the measurement between the hyperechoic fascial layers of the muscle. During the LM measurement, patients were positioned prone with a thin pillow placed under the abdomen to ensure neutral alignment of the lumbar spine. The contraction image was obtained by muscle activation; the participants lifted the contralateral arm off the bed and maintained it at 120 degrees of shoulder abduction and 90 degrees of elbow flexion. The LM muscle thickness was measured with a curved probe, defined as the distance between the dorsal edge of the L4-L5 zygapophyseal joint and the thoracolumbar hyperechoic fascia, which separates the muscle from the subcutaneous fat, excluding the fascia itself. All participants underwent ultrasound examinations and muscle thickness assessments conducted by the same experienced physician both before and after the intervention. Measurements were performed three times on each side, and the mean value was calculated for data analysis. Motion capture and electromyography measurement Prior to the test, participants received instructions to wear tight-fitting clothing and were provided with guidance on performing the walking tests in the designated area. Participants first assumed an anatomical position, characterized by an upright posture, facing forward, eyes level, feet together with toes pointing forward, and arms resting at the sides with palms forward, to collect static data. Participants were then instructed to walk across the force plates at a self-comfortable pace. After each pass, participants turned and returned to the starting point at a comfortable pace, repeating this process five times. A successful trial was characterized by the simultaneous contact of both feet with the two force plates during a gait cycle. Kinematic data were recorded utilizing a 6-camera infrared 3D motion capture system (BTS SMART-DX EVO2) at a sampling frequency of 100 Hz. A total of forty-nine reflective markers, each with a diameter of 14.0 mm, were affixed to the bodies of the participants. Both static and dynamic tests demonstrated an error margin of less than 0.1 mm within a volume of 4 m × 3 m × 3 m (L × W × H), comparable to other commercial biomechanics systems. Ground reaction forces (GRF) were recorded using two force plates (BTS P6000) at a sampling frequency of 1,000 Hz. Surface electromyography (EMG) data were obtained from the bilateral erector spinae and iliocostalis muscles using the BTS FREEEMG 300 system. The sampling frequency was 1,000 Hz, and the electrodes were placed according to SENIAM guidelines. The 3D kinematic, GRF, and surface EMG data were calibrated and synchronized within the BTS system to ensure that the data were collected at the same time. Musculoskeletal modeling and simulation Gait data were imported into OpenSim software (version 4.1) for kinematic and dynamic analysis using the Full Body Lumbar Spine (FBLS) model developed by Raabe and Chaudhari [ 35 , 36 ] The model was scaled to correspond with each subject’s height, weight, and body proportions. Reflective marker trajectories were imported into OpenSim, and inverse kinematics were utilized to compute motion data. The model’s kinematics were modified using the reduced residuals algorithm to improve dynamic consistency with the measured GRF and joint moments. The process produced an adjusted model and new sets of inverse kinematics and inverse dynamics results. In the FBLS model of OpenSim, the LM is modeled as a circular cross-sectional, nonlinear spring structure, with muscle thickness represented by the cross-sectional diameter. Muscle thickness was assessed using the cross-sectional diameter of the LM model, and alterations in resting cross-sectional area were evaluated before and after the intervention. The model demonstrates a relationship between muscle cross-sectional area and maximum isometric contraction force. The post-intervention model was personalized by adjusting the maximum isometric contraction force of the muscles based on the ratio of each subject’s muscle cross-sectional diameter, measured through ultrasound measurements conducted before and after the intervention [ 37 ] Static Optimization was subsequently performed to assign net joint moments, derived from inverse dynamics, to specific muscle fibers on a frame-by-frame basis. This process generated muscle forces and activation levels throughout the motion. Muscle activation occurred when the calculated activation exceeded 5% [ 38 ] The GRF data, adjusted muscle force, and kinematic data were input into the joint reaction module to calculate the compressive loads and shear forces at the L3-L4 and L4-L5, and L5-S1 segments. This study adjusted the scaling factor for the maximum isometric contraction force in the OpenSim muscle model according to alterations in resting muscle area, determined by the ratio of post-intervention to pre-intervention muscle area measured via musculoskeletal ultrasound. Statistical analysis Statistical analyses were conducted using IBM SPSS (version 25). The Chi-square test was employed to examine the gender variable. The Shapiro-Wilk test was used to check the normality for continuous variables, whereas the Levene test was used to evaluate homogeneity of variance. A one-way repeated measures analysis of variance (ANOVA) was performed to evaluate variations in muscle thickness, VAS, flexion-extension and lateral bending peak moments, and muscle peak forces before and after the intervention. The mean ± standard deviation was used to report all outcomes that were continuous variables. A significance level (α) of 0.05 was set for all statistical comparisons. Results Pain and functional assessment The study findings demonstrated substantial decreases in Visual Analogue Scale (VAS) scores for worst pain experienced in the past week, average pain in the past week, and current pain levels after the intervention (P<0.01). Furthermore, significant decreases were observed in the Oswestry Disability Index (ODI) and Roland-Morris Disability Questionnaire (RMDQ) scores (P <0.01), as well as in Timed Up and Go (TUG) test times (P <0.01) following the intervention. Table 1 presents the detailed questionnaire results for NSCLBP participants before and after the intervention. Table 1 Questionnaire before and after intervention (X ± S) Note: The symbol “α” denotes the use of non-parametric tests, while paired t-tests are used for other statistical analyses. Variable Pre-intervention (n = 14) Post-intervention(n = 14) p worst VAS score in the past week(mm) 50 ± 13 29 ± 13 <0.01 the average VAS score recorded over the past week α (mm) 39 ± 08 2.0 ± 12 <0.01 the current VAS score α (mm) 34 ± 05 14 ± 13 <0.01 ODI α 7.8 ± 3.1 5.3 ± 2.7 <0.01 RMDQ α 5.1 ± 4.0 2.0 ± 1.8 <0.01 TUG (s) 10.4 ± 1.1 8.5 ± 0.7 <0.01 Muscle thickness, muscle force, and activation Ultrasound measurements (Table 2 ) demonstrated that four weeks of RAVEP training significantly increased the resting thickness of the bilateral LM and TrA in participants with NSCLBP (p 0.05). The contracted thickness of the bilateral TrA and the right LM exhibited a significant increase (p<0.01). The contraction ratio of the bilateral LM and the left TrA showed a statistically significant improvement (p 0.05). Table 2 Muscle thickness was measured both before and after the intervention(X ± S) pre-intervention (n = 14) post-intervention (n = 14) P The resting thickness of the left TrA (mm) 3.3 ± 0.9 3.7 ± 0.8 <0.01 The resting thickness of the right TrA (mm) α 3.2 ± 0.7 3.7 ± 0.8 <0.01 The resting thickness of the left LM (mm) α 26.2 ± 4.3 27.1 ± 4.3 <0.01 The resting thickness of the right LM (mm) 25.4 ± 3.5 27.3 ± 3.6 <0.01 The contracted thickness of the left TrA (mm) α 4.7 ± 1.1 5.3 ± 1.1 <0.01 The contracted thickness of the right TrA (mm) 4.7 ± 1.0 5.4 ± 1.1 <0.01 The contracted thickness of the left LM (mm) α 34.5 ± 4.3 38.4 ± 5.0 0.14 The contracted thickness of the right LM (mm) 34.1 ± 4.2 39.9 ± 4.9 <0.01 The contraction ratio of the left TrA (%) 43.8 ± 17.2 46.0 ± 14.9 0.732 The contraction ratio of the right TrA (%) 48.7 ± 19.0 48.8 ± 20.2 0.956 The contraction ratio of the left LM (%) 32.7 ± 12.0 42.7 ± 11.7 0.018 The contraction ratio of the right LM (%) 34.8 ± 12.0 46.7 ± 11.7 <0.01 Note: The symbol “α” denotes the use of non-parametric tests, while paired t-tests are used for other statistical analyses. The variation in muscle force during walking was similar in both the pre- and post-intervention phases. The OpenSim static optimization model revealed significant increases in the maximum muscle forces of the bilateral LM following the intervention (p0.05); however, curve-fitting analysis revealed that during 60%-80% of the gait cycle, the force of the left external oblique abdominal muscles experienced a slight decrease (Fig. 2 B), while the forces of the left internal oblique abdominal muscles (Fig. 2 D) and right external oblique abdominal muscles (Fig. 2 A) showed slight increases. The simulated muscle activation patterns (Figs. 3 I and 3 J) closely aligned with the observed trends in muscle force. Analysis of muscle activation (Fig. 3 A–H) demonstrated increased activation of the bilateral LM, right internal oblique, and right external oblique muscles during 60–80% of the gait cycle. Conversely, the activation of the left external oblique and internal oblique muscles exhibited a slight decrease during the same period. However, these alterations in muscle activation did not reach statistical significance (P > 0.05). Kinematics Patients with NSCLBP exhibited consistent alterations in lumbar flexion, lateral flexion, and rotation angles while walking, both before and after the intervention (Figs. 4 A-C). Curve-fitting analysis indicated a reduction in the sagittal plane motion angle of the lumbar spine in patients with NSCLBP after the intervention. The maximum lumbar sagittal plane motion angle decreased from − 0.23 ± 0.22° pre-intervention to -0.15 ± 0.14° post-intervention; however, this difference was not statistically significant (p > 0.05). Figure 4 A illustrates the lumbar motion in the sagittal plane during walking, with fitted curve analysis indicating a reduction in motion among NSCLBP patients post-intervention. Figures 4 B and 4 C depict lumbar motion in the coronal and axial planes, respectively, before and after the intervention. Figure 4 D–F presents the alterations in hip, knee, and ankle joint angles during walking prior to and following the intervention. Figure 2 D details hip joint motion during walking under both pre- and post-intervention conditions. The peak sagittal plane knee flexion angle in patients with NSCLBP decreased from 69.42 ± 4.13° before the intervention to 68.54 ± 4.89° after the intervention (p > 0.05). Figure 2 E illustrates the detailed results of knee joint motion during walking. After the intervention, participants with NSCLBP demonstrated decreased dorsiflexion and plantarflexion angles at the ankle. The peak ankle joint range of motion declined from 16.95 ± 5.62° prior to the intervention to 15.15 ± 4.76°following the intervention, as shown in Fig. 2 F (p > 0.05). Lumbar intervertebral loading Figure 5 presents the simulation curves depicting changes in the intervertebral compression force, sagittal shear force, and twisting force at the L3-L4, L4-L5, and L5-S1 segments during a single gait cycle, both prior to and following the intervention. Curve-fitting analysis indicated that the peak shear forces at the L3-4, L4-5, and L5-S1 segments in NSCLBP participants were 0.77 ± 0.29, 0.90 ± 0.31, and 1.07 ± 0.36 times body weight, respectively, before the intervention, and 0.78 ± 0.22, 0.83 ± 0.21, and 1.02 ± 0.29 times body weight, respectively, after the intervention (p > 0.05). The shear force at the L3-4 segment exhibited a minor increase, whereas reductions were observed at the L4-5 and L5-S1 segments. Figures 5 A, 5 D, and 5 G illustrate the alterations in lumbar shear forces during walking before and after the intervention. Curve-fitting analysis revealed that the peak compressive forces at the L3-4, L4-5, and L5-S1 segments in NSCLBP participants were 3.33 ± 0.59, 3.49 ± 0.64, and 3.70 ± 0.68 times body weight, respectively, prior to the intervention, and decreased to 3.18 ± 0.33, 3.32 ± 0.23, and 3.50 ± 0.25 times body weight, respectively, following the intervention (p > 0.05). Reductions in compressive forces were observed across all three lumbar segments. Figures 5 B, 5 E, and 5 H present the detailed changes in lumbar compressive forces during walking before and after the intervention. Curve-fitting analysis indicated that the peak torsional moments at the L3-4, L4-5, and L5-S1 segments in NSCLBP participants were 1.12 ± 0.22, 1.24 ± 0.23, and 1.30 ± 0.24 times body weight, respectively, before the intervention, and subsequently decreased to 1.04 ± 0.14, 1.16 ± 0.16, and 1.23 ± 0.18 times body weight, respectively, after the intervention (p > 0.05). All three lumbar segments showed decreases in torsional moments. Figures 5 C, 5 F, and 5 I depict the specific alterations in lumbar torsional moments during walking, both pre- and post-intervention. However, these differences did not achieve statistical significance (p > 0.05). Discussion This study evaluates the efficacy of RAVEP training in patients with NSCLBP by analyzing clinical outcomes and biomechanical factors, such as kinematics, dynamics, lumbar load, and muscle force. A biomechanical analysis of pre- and post-intervention data indicated significant improvement in dynamics, muscle force, and overall clinical performance. Clinical performance Post-RAVEP training demonstrated a significant improvement in pain and functional outcome for NSCLBP patients. Functional tests revealed significant decreases in patients’ weekly worst, average, and current pain levels, as well as overall pain alleviation. The ODI and RMDQ scores decreased dramatically, indicating that NSCLBP had a smaller impact on daily activities. Furthermore, shorter TUG test durations indicated better lower limb mobility after the intervention. These findings suggest that RAVEP training significantly reduces pain and improves functional performance in patients with NSCLBP. These findings align with prior research showing that both isometric and isotonic core stabilization exercises significantly alleviate pain and disability in patients with NSCLBP compared to no intervention [ 39 ] Just as the recent guidelines also recommended the various exercise therapies for NSCLBP patients, which have shown moderate improvements in pain relief and functional capacity [ 20 ]. This study then showed that NSCLBP patients exhibited significant bilateral increases in LM thickness and TrA thickness in both resting and contracted states, as well as a marked increase in contraction rate in bilateral LM after the RAVEP training. Similar results were also seen in a comparative study that reported a notable increase in the thickness of the TrA and LM, along with enhanced contractile capacity, after an eight-week core training program in healthy people [ 40 ] The results indicated that extended core training leads to significant muscle hypertrophy in the core region, thereby increasing muscle cross-sectional area and enhancing force production. Core exercise significantly improves the biomechanical structure of the lumbar spine through this adaptation. This enhancement probably contributes to the noted advancements in lumbar biomechanics subsequent to RAVEP training. Muscle force and activation Prior studies have shown that patients with chronic LBP often exhibit atrophy or fatty infiltration in the lumbar multifidus and other deep core muscles [ 41 ], leading to functional impairment, delayed activation, and inadequate recruitment. This dysfunction may arise from persistent compensatory overactivation of superficial muscles due to inhibition of deep core muscles, leading to an inefficient stabilization pattern. OpenSim simulation results indicated notable post-intervention increases in bilateral LM forces among NSCLBP patients, with increases of 50.9% and 45.4%, respectively (p < 0.05). Consistent with the ultrasound analysis and muscle strength simulations revealed notable increases in LM muscle thickness, contraction rate, and strength were revealed among participants. The findings indicated that muscle hypertrophy alone does not fully explain the observed strength gains; rather, enhanced muscle recruitment and activation patterns during walking seem to be the main factors following the intervention. These findings also suggested that RAVEP training induced morphological alterations and reactivated levels in deep core muscles. Previous research supports this perspective, indicating that core stability training reeducates motor control in inhibited muscles, potentially improving muscle activation and positively influencing the management of chronic LBP [ 42 ] The simulation of LM activation before the intervention showed that the left limb stance phase had greater activation relative to the right. This could be because the right-side lumbar pain caused the trunk imbalance. Following the intervention, LM activation during the right limb stance phase exhibited increased symmetry with the left, suggesting that core stability training effectively improved muscle coordination, trunk control, and movement patterns. Kinematics This study found that after four weeks of the RAVEP training, patients with NSCLBP exhibited a decrease in lumbar spine joint angle during flexion-extension while walking, consistent with previous research findings.43–46 There are several explanations for our finds. Core strengthening exercises, as employed in RAVEP training, can effectively enhance lumbar spine stability [ 43 ] and improve spinal alignment to decrease the lumbar spine movement during daily activities. Greater core strength improves proprioception and postural control. This can reduce the strain and pain in the lower back [ 44 ] After four weeks of RAVEP training, the symmetry in coronal plane range of motion and axial rotation was improved in the lumbar spine. A reduction of 4.4% in the axial rotation angle was noted, suggesting improved rotational symmetry. The lateral stability of the lumbar spine during postural changes may significantly influence reported outcomes for patients. Prior studies indicated that core stabilization exercises positively influence the lumbar lordotic angle in patients with lumbar disc degeneration, suggesting improvements in spinal alignment and stability [ 45 ] These exercises markedly decrease sagittal translation at L4-L5 and L5-S1 levels. In the coronal plane, the lumbar lateral bending angle (swing amplitude) was smaller than that observed after RAVEP training. This suggests that the lateral stability of the lumbar spine was improved, consistent with a previous study involving adult idiopathic scoliosis, which demonstrated that a core exercise program could decrease the Cobb angle, reflecting an improvement in spinal alignment in the coronal plane [ 45 ] During the stance phase of the gait cycle, the hip joint showed no changes before and after RAVEP training. In the swing phase, the hip joint angle increased post-intervention relative to pre-intervention measurements, consistent with previous research findings47-49. This may relate to symptomatic individuals utilizing compensatory movements and modified load-sharing strategies, or reduced coordination abilities of the lumbar spine concerning the hip joint. Hip-spine syndrome exemplifies the compensatory mechanisms that arise from restricted hip motion, resulting in heightened lumbar spine loading and a potential increase in low back pain [ 47 ] Core strengthening may enhance hip range of motion and alleviate symptoms related to restricted hip mobility; however, its efficacy could be constrained by compensatory movements and altered coordination strategies. The knee joint angle alterations in lower limb joints during the swing phases showed improvement following RAVEP training, aligning with prior research findings. A previous study demonstrated that patients with NSCLBP exhibit modified gait patterns, marked by reduced hip and knee motion across various planes during distinct phases of the gait cycle [ 18 ] This reduction in range of motion may result in an increased knee joint angle at initial landing during the support phase, consistent with the findings of our study. NSCLBP patients exhibit a slower preferred walking speed [ 48 ], which may be attributed to the knee joint angle at the support phase as a compensatory strategy to sustain balance and alleviate pain. After RAVEP training, patients with NSCLBP exhibited a decrease in plantar flexion and dorsal extension during mid-stance, potentially reversing the altered ankle movement pattern in the sagittal plane associated with NSCLBP. Prior research indicated that this compensatory strategy of the ankle serves to prevent downward displacement of the center of gravity and to reduce mechanical load on the lumbar spine Lumbar intervertebral loading The results concerning lumbar compression and shear forces demonstrated a distinct bimodal pattern, intricately linked to the single-leg standing phase of the left and right feet within the gait cycle. After four weeks of RAVEP training, participants showed a minor rise in shear force at the L3-L4 segment. However, the shear forces at the L4-L5 and L5-S1 segments displayed a significant reduction, coupled with a pronounced improvement in LM strength. This finding aligns with biomechanical studies suggesting that the weakness of core muscles may lead to instability at the L4-L5 spinal segment. Numerous studies indicated that deep muscle fibers are essential for maintaining lumbar stability [ 50 ], especially during flexion, and likely regulate lumbar shear forces [ 51 ]. An experimental finding demonstrated that the resultant joint force, considering shear force allowance, is oriented towards a particular region within the vertebral body [ 52 ] We proposed that increased strength and activation of the deep core muscles generate a greater posterior stabilizing force along the lumbar spine, thus preserving physiological lumbar curvature, improving overall lumbar stability, and counteracting anterior-directed mechanical forces. Therefore, the aforementioned stabilizing effect likely explains the observed reductions in shear forces at the L4-L5 and L5-S1 segments. The shear force at the L3-L4 segment was unexpectedly elevated. Prior research has underscored the important role of the psoas muscle in lumbar motion and stability, specifically by its application of compressive forces on the anterior aspect of the L3-L4 segment [ 53 ] The heightened shear force at the L3-L4 segment may result from adaptations in the psoas muscle, as core stability training may improve its compressive function at the anterior lumbar spine, subsequently increasing the shear force at this segment. Lumbar intervertebral compression is a key biomechanical factor that influences the structural integrity and functionality of the spine. The application of compressive forces on intervertebral discs induces various mechanical and physiological alterations [ 54 ] Following the four-week RAVEP training, a decline in lumbar compression forces was observed across all spinal segments, though this decline did not reach statistical significance(p > 0.05). Consistent with prior research, improvements in the muscles and soft tissue function within the kinetic chain after the intervention may improve mechanical force transmission [ 55 ], which we suggest as a key factor in the reduction of lumbar compression forces observed. Lumbar torque decreased after the intervention, likely attributable to alterations in kinematics. Lumbar torque primarily arises from the swinging motion of the upper body and pelvis during walking, generating rotational forces on the intervertebral discs that may lead to degenerative changes. Similar results were seen in a previous study that employed surface electromyography to assess core muscle activity during exercises has shown that strengthening these muscles can increase lumbar torque [ 56 ], thereby enhancing overall functional performance and reducing the risk of musculoskeletal injuries. The strength and activation of the left internal oblique and right external oblique muscles after the four-week RAVEP training, which enhanced the lumbar spine’s resistance to rotational forces and increased rotational stability, contributed directly to a reduction in lumbar torque. Limitations and future work This study evaluates lumbar loading during walking following the four-week RAVEP training, utilizing pre- and post-intervention data from patients with NSCLBP and a modified FBLS model that integrates ultrasound-derived muscle measurements. However, the study presents several limitations. Firstly, OpenSim offers a range of preset mannequins for simulation purposes. However, these simplified models may not sufficiently capture the intricacies of real human motion. Secondly, this study highlighted the feasibility of remote AI-supervised virtual exercises for managing NSCLBP, influenced by factors such as smartphone usability and acceptance of tele-rehabilitation. Thirdly, the majority of participants were young adults aged 21 to 27 years, a demographic typically experiencing mild lumbar degeneration, which may be mitigated by enhanced muscle activity. Future research should include participants from a wider age, and motion spectrum. Conclusion This study indicated that RAVEP training markedly enhanced deep core muscle function, restored lumbar range of motion, decreased mechanical forces on the lumbar spine, and alleviated pain symptoms. The AI-guided real-time feedback component of RAVEP training improved the precision and coordination of local muscle contractions. The enhancements in lumbar muscle function enabled the restoration of optimal spinal posture and motion patterns, reduced excessive mechanical forces on the lumbar spine, alleviated pain-inducing factors, and ultimately improved outcomes for NSCLBP. Abbreviations LBP: Low back pain NSCLBP: Nonspecific chronic low back pain TrA: Transversus abdominis LM: Lumbar multifidus FBLS: Full-body lumbar spine AI: Artificial intelligence RAVEP: Remote AI-supervised virtual exercise program VAS: Visual Analog Scale NSAIDs: Nonsteroidal anti-inflammatory drugs APTA: American Physical Therapy Association GRF: Ground reaction forces EMG: Electromyography ODI: Oswestry Disability Index RMDQ: Roland-Morris Disability Questionnaire TUG: Timed Up and Go Declarations Ethics approval and consent to participate This study has the formal ethical approval from the Ethics Committee of Zhujiang Hospital of Southern Medical University (2023-KY-017). Competing interests No competing interests. Consent for publication Not applicable. Author Contributions Study concept and design: GH, MZ, QX; Acquisition of data: PL, ZZ, JH; Analysis and interpretation of data: PL, ZZ, JH; Drafting of the manuscript: PL, ZZ, MZ; Substantively revising: QZ, TF. All authors contributed to the review and revision of the manuscript. All authors have consented the final manuscript. Funding This work was supported by the grants from Guangzhou Municipal Science and Technology Bureau (2023A04J0444), Southern Medical University Youth Incubation Program (PY2018N057), Natural Science Foundation of China (82171174, 82371184), the Natural Science Foundation of China under Grant (52105305), Natural Science Foundation of Guangdong Province (2025A1515012782), and the Key Technologies R&D Program of Guangdong Province (2023B0303020003). Availability of data and materials The datasets generated during the current study are available from the corresponding author on reasonable request. Acknowledgements Not applicable. References Collaborators GDAI. Global burden of 369 diseases and injuries in 204 countries and territories, 1990-2019: A systematic analysis for the Global Burden of Disease Study 2019. Lancet 2020;396: 1204-22. Moser I, Sommer I, Gartlehner G. [who guideline for Non-Surgical management of chronic primary low back pain in adults in primary and community care settings]. Gesundheitswesen 2024;86: 715-22. Bittar M, Deodhar A. A critical view of WHO guidelines on management of low back pain. Nat Rev Rheumatol 2024;20: 201-2. Koes BW, van Tulder MW, Thomas S. Diagnosis and treatment of low back pain. BMJ 2006;332: 1430-4. Gianola S, Bargeri S, Del Castillo G, Corbetta D, Turolla A, Andreano A, et al. Effectiveness of treatments for acute and subacute mechanical non-specific low back pain: A systematic review with network meta-analysis. Br J Sports Med 2022;56: 41-50. Zafar T, Zaki S, Alam MF, Sharma S, Babkair RA, Nuhmani S, et al. Effect of Whole-Body vibration exercise on pain, disability, balance, proprioception, functional performance and quality of life in people with Non-Specific chronic low back pain: A systematic review and Meta-Analysis. J Clin Med 2024;13. Knezevic NN, Candido KD, Vlaeyen JWS, Van Zundert J, Cohen SP. Low back pain. Lancet 2021;398: 78-92. Balasch-Bernat M, Willems T, Danneels L, Meeus M, Goubert D. Differences in myoelectric activity of the lumbar muscles between recurrent and chronic low back pain: A cross-sectional study. BMC Musculoskelet Disord 2021;22: 756. Pinto SM, Boghra SB, Macedo LG, Zheng Y, Pang MYC, Cheung JPY, et al. Does motor control exercise restore normal morphology of lumbar multifidus muscle in people with low back pain? - a systematic review. J Pain Res 2021;14: 2543-62. Wong AYL, Parent EC, Funabashi M, Stanton TR, Kawchuk GN. Do various baseline characteristics of transversus abdominis and lumbar multifidus predict clinical outcomes in nonspecific low back pain? A systematic review. Pain 2013;154: 2589-602. Oshima Y, Watanabe N, Takeoka T, Kajikawa Y, Yotsumoto T, Majima T. Preoperative abnormal posture improves after total knee arthroplasty: Knee-Hip-Spine syndrome. J Arthroplasty 2025;40: 416-22. Govil G, Tomar L, Dhawan P. Knee-Spine syndrome: Management dilemma when knee osteoarthritis coexists with spine degeneration. Cureus 2022;14: e24939. Zambarano E, Glaviano N, Bouillon L, Norte G, Murray A. Effect of exhaustive exercise on Lumbopelvic-Hip complex stability, muscle activity, and movement patterns. J Electromyogr Kinesiol 2024;74: 102852. Zawadka M, Smolka J, Skublewska-Paszkowska M, Lukasik E, Zielinski G, Bys A, et al. Altered squat movement pattern in patients with chronic low back pain. Ann Agric Environ Med 2021;28: 158-62. de Jesus FLA, Fukuda TY, Souza C, Guimaraes J, Aquino L, Carvalho G, et al. Addition of specific hip strengthening exercises to conventional rehabilitation therapy for low back pain: A systematic review and meta-analysis. Clin Rehabil 2020;34: 1368-77. Alshehri MA, van den Hoorn W, Klyne DM, Hodges PW. Coordination of hip and spine in individuals with acute low back pain during unstable sitting. Spine J 2024;24: 768-82. von Arx M, Liechti M, Connolly L, Bangerter C, Meier ML, Schmid S. From stoop to squat: A comprehensive analysis of lumbar loading among different lifting styles. Front Bioeng Biotechnol 2021;9: 769117. Zhang Z, Zou J, Lu P, Hu J, Cai Y, Xiao C, et al. Analysis of lumbar spine loading during walking in patients with chronic low back pain and healthy controls: An OpenSim-Based study. Front Bioeng Biotechnol 2024;12: 1377767. Jones CMP, Day RO, Koes BW, Latimer J, Maher CG, McLachlan AJ, et al. Opioid analgesia for acute low back pain and neck pain (the OPAL trial): A randomised placebo-controlled trial. Lancet 2023;402: 304-12. Chiarotto A, Koes BW. Nonspecific low back pain. N Engl J Med 2022;386: 1732-40. Choo YJ, Chang MC. The effect of exercise on stabilizing and strengthening core muscles for patients with symptomatic herniated lumbar disc: A systematic review and meta-analysis. Asian J Surg 2024;47: 1703-4. Miyachi R, Miyazaki J. Relationship between lumbar motor control ability and spinal curvature in elderly individuals. Healthcare (Basel) 2020;8. Mataran-Penarrocha GA, Lara Palomo IC, Antequera Soler E, Gil-Martinez E, Fernandez-Sanchez M, Aguilar-Ferrandiz ME, et al. Comparison of efficacy of a supervised versus non-supervised physical therapy exercise program on the pain, functionality and quality of life of patients with non-specific chronic low-back pain: A randomized controlled trial. Clin Rehabil 2020;34: 948-59. Lopez-Marcos JJ, Diaz-Arribas MJ, Valera-Calero JA, Navarro-Santana MJ, Izquierdo-Garcia J, Ortiz-Gutierrez RM, et al. The added value of Face-to-Face supervision to a therapeutic Exercise-Based app in the management of patients with chronic low back pain: A randomized clinical trial. Sensors (Basel) 2024;24. Abedi A, Colella TJF, Pakosh M, Khan SS. Artificial intelligence-driven virtual rehabilitation for people living in the community: A scoping review. NPJ Digit Med 2024;7: 25. Day MA, Ciol MA, Mendoza ME, Borckardt J, Ehde DM, Newman AK, et al. The effects of telehealth-delivered mindfulness meditation, cognitive therapy, and behavioral activation for chronic low back pain: A randomized clinical trial. BMC Med 2024;22: 156. Costa F, Janela D, Molinos M, Moulder RG, Lains J, Bento V, et al. Digital rehabilitation for acute low back pain: A prospective longitudinal cohort study. J Pain Res 2022;15: 1873-87. Fatoye F, Gebrye T, Fatoye C, Mbada CE, Olaoye MI, Odole AC, et al. The clinical and Cost-Effectiveness of telerehabilitation for people with nonspecific chronic low back pain: Randomized controlled trial. JMIR Mhealth Uhealth 2020;8: e15375. Ekambaram D, Ponnusamy V. Real-Time monitoring and assessment of rehabilitation exercises for low back pain through interactive dashboard pose analysis using streamlit—a pilot study. Electronics (Basel) 2024;13: 3782. Fritz JM, Minick KI, Brennan GP, McGee T, Lane E, Skolasky RL, et al. Outcomes of telehealth physical therapy provided using Real-Time, videoconferencing for patients with chronic low back pain: A longitudinal observational study. Arch Phys Med Rehabil 2022;103: 1924-34. Lewkowicz D, Slosarek T, Wernicke S, Winne A, Wohlbrandt AM, Bottinger E. Digital therapeutic care and decision support interventions for people with low back pain: Systematic review. JMIR Rehabil Assist Technol 2021;8: e26612. Chou R, Qaseem A, Snow V, Casey D, Cross JTJ, Shekelle P, et al. Diagnosis and treatment of low back pain: A joint clinical practice guideline from the American College of Physicians and the American Pain Society. Ann Intern Med 2007;147: 478-91. George SZ, Fritz JM, Silfies SP, Schneider MJ, Beneciuk JM, Lentz TA, et al. Interventions for the management of acute and chronic low back pain: Revision 2021. J Orthop Sports Phys Ther 2021;51: G1-60. Djordjevic O, Djordjevic A, Konstantinovic L. Interrater and intrarater reliability of transverse abdominal and lumbar multifidus muscle thickness in subjects with and without low back pain. J Orthop Sports Phys Ther 2014;44: 979-88. Raabe ME, Chaudhari AMW. Biomechanical consequences of running with deep core muscle weakness. J Biomech 2018;67: 98-105. Raabe ME, Chaudhari AMW. An investigation of jogging biomechanics using the full-body lumbar spine model: Model development and validation. J Biomech 2016;49: 1238-43. McCrady AN, Masterson CD, Barnes LE, Scharf RJ, Blemker SS. Development of an ultrasound-based metric of muscle functional capacity for use in patients with neuromuscular disease. Muscle Nerve 2024;70: 1205-14. Li M, Wang K, Niu W, Zhang S. A musculoskeletal modeling of hand-foot crawling with different heights. J Bionic Eng 2020;17: 591-9. Khaledi A, Gheitasi M. Isometric vs isotonic core stabilization exercises to improve pain and disability in patients with non-specific chronic low back pain: A randomized controlled trial. Anesth Pain Med 2024;14: e144046. Zheng Y, Ke S, Lin C, Li X, Liu C, Wu Y, et al. Effect of core stability training monitored by rehabilitative ultrasound image and surface electromyogram in local core muscles of healthy people. Pain Res Manag 2019;2019: 9130959. Balevi Batur E. Assessing the correlation between the lumbar disc herniation degree and multifidus muscle fatty degeneration in chronic low back pain. Cureus 2024;16: e71237. Mohammadi V, Letafatkar A, Jafarnezhadgero AA. Effects of core stability training on kinematic and kinetic variables in patients with chronic low back pain. Physical Treatments-Specific Physical Therapy Journal 2023;13: 55-66. Frizziero A, Pellizzon G, Vittadini F, Bigliardi D, Costantino C. Efficacy of core stability in Non-Specific chronic low back pain. J Funct Morphol Kinesiol 2021;6. Salik Sengul Y, Yilmaz A, Kirmizi M, Kahraman T, Kalemci O. Effects of stabilization exercises on disability, pain, and core stability in patients with non-specific low back pain: A randomized controlled trial. Work 2021;70: 99-107. Shalamzari MH, Henteh MA, Shamsoddini A, Ghanjal A. Comparison of the effects of core stability and whole-body electromyostimulation exercises on lumbar lordosis angle and dynamic balance of sedentary people with hyperlordosis: A randomized controlled trial. BMC Sports Sci Med Rehabil 2024;16: 91. Dimitrijevic V, Viduka D, Scepanovic T, Maksimovic N, Giustino V, Bianco A, et al. Effects of Schroth method and core stabilization exercises on idiopathic scoliosis: A systematic review and meta-analysis. Eur Spine J 2022;31: 3500-11. Huppert A, Ambrosio L, Nwosu K, Pico A, Russo F, Vadala G, et al. Previous lumbar spine fusion increases the risk of dislocation following total hip arthroplasty in patients with hip-spine syndrome: A systematic review and meta-analysis. BMC Musculoskelet Disord 2024;25: 732. Yalfani A, Asgarpoor A. Comparison of cognitive functional therapy and neurofeedback training on kinetic gait in patients with chronic non-specific low back pain: A randomised controlled trial. Disabil Rehabil 2025: 1-10. Xiao F, Maas H, van Dieen JH, Pranata A, Adams R, Han J. Chronic non-specific low back pain and ankle proprioceptive acuity in community-dwelling older adults. Neurosci Lett 2022;786: 136806. Padwal J, Berry DB, Hubbard JC, Zlomislic V, Allen RT, Garfin SR, et al. Regional differences between superficial and deep lumbar multifidus in patients with chronic lumbar spine pathology. BMC Musculoskelet Disord 2020;21: 764. Bastos De Oliveira V, Albuquerque Brandao MC, Coelho De Albuquerque Pereira W, Fernandes De Oliveira L. Lumbar multifidus layers stiffness at L5-S1 level in prone and sitting posture measured by shear wave elastography. J Back Musculoskelet Rehabil 2024;37: 1241-8. Kim K, Kim YH, Lee S. Shear force allowance in lumbar spine under follower load in neutral standing posture. Acta Bioeng Biomech 2010;12: 49-53. Seyedhoseinpoor T, Taghipour M, Dadgoo M, Ebrahimi Takamjani I, Sanjari MA, Kazemnejad A, et al. Relationship between the morphology and composition of the lumbar paraspinal and psoas muscles and lumbar intervertebral motion in chronic low-back pain: An exploratory study. Clin Anat 2022;35: 762-72. Eremina G, Smolin A, Xie J, Syrkashev V. Development of a computational model of the mechanical behavior of the L4-L5 lumbar spine: Application to disc degeneration. Materials (Basel) 2022;15. Clael S, Campos LF, Correia KL, de Lucena JMS, Gentil P, Durigan JL, et al. Exercise interventions can improve muscle strength, endurance, and electrical activity of lumbar extensors in individuals with non-specific low back pain: A systematic review with meta-analysis. Sci Rep 2021;11: 16842. Arvanitidis M, Jimenez-Grande D, Haouidji-Javaux N, Falla D, Martinez-Valdes E. Eccentric exercise-induced delayed onset trunk muscle soreness alters high-density surface EMG-torque relationships and lumbar kinematics. Sci Rep 2024;14: 18589. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7344603","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":506584305,"identity":"19cb5096-695a-4084-9d0b-c89ac09db4e7","order_by":0,"name":"Pengcheng Lu","email":"","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Pengcheng","middleName":"","lastName":"Lu","suffix":""},{"id":506584306,"identity":"4ac45849-6c67-42d6-9cbe-331fe589821c","order_by":1,"name":"Zhuodong Zhang","email":"","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Zhuodong","middleName":"","lastName":"Zhang","suffix":""},{"id":506584307,"identity":"1235fe9b-e824-4112-a032-590745da2486","order_by":2,"name":"Jingjing Hu","email":"","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jingjing","middleName":"","lastName":"Hu","suffix":""},{"id":506584308,"identity":"0e16873e-b051-4b80-bfe5-8f5d98ce3b86","order_by":3,"name":"Qing Zeng","email":"","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Qing","middleName":"","lastName":"Zeng","suffix":""},{"id":506584309,"identity":"7cbd4df3-8196-4ac9-85d2-1e024766989a","order_by":4,"name":"Tao Fan","email":"","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Tao","middleName":"","lastName":"Fan","suffix":""},{"id":506584311,"identity":"a1659d83-8323-4efc-8e89-2f1f4b9d4972","order_by":5,"name":"Qiuyou Xie","email":"","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Qiuyou","middleName":"","lastName":"Xie","suffix":""},{"id":506584313,"identity":"56f50717-6bdf-486f-b097-e09342b36c3d","order_by":6,"name":"Manxu Zheng","email":"","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":false,"prefix":"","firstName":"Manxu","middleName":"","lastName":"Zheng","suffix":""},{"id":506584315,"identity":"9655e88b-8f05-48fb-8bc3-f1aabf10cb89","order_by":7,"name":"Guozhi Huang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAx0lEQVRIiWNgGAWjYDCCAzxgSo6fvbHx4QdStBhL9hxuNpYgRUuiwYz0NgEeYnTw3T57TOLnjtoEA8mHbQwSDHZyug0EtEiey0uT7D1zPM9cOrHtQQFDsrHZAQJaDM7wmEnwth0rtpyd2G4gwXAgcRsxWiT/th1L3HDzYJsED7FapHnbahI33GAkUovkGR5ja9m2A8BATgQGsgERfuE7w2N4821bHTAqjz98+KHCTo6gFig4DHMnccpBoI54paNgFIyCUTDyAACZd0SbfR5sJQAAAABJRU5ErkJggg==","orcid":"","institution":"Zhujiang Hospital, Southern Medical University","correspondingAuthor":true,"prefix":"","firstName":"Guozhi","middleName":"","lastName":"Huang","suffix":""}],"badges":[],"createdAt":"2025-08-11 09:23:17","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7344603/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7344603/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":90315833,"identity":"3e017d18-d09d-4840-a0f1-ce815a794ede","added_by":"auto","created_at":"2025-09-01 10:17:02","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":553448,"visible":true,"origin":"","legend":"\u003cp\u003eSchematic of the program: (A) and (B) are the user training sessions interface; (C) is the user’s mobile phone display interface; (D) is the schematic interface of key points on the body surface.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7344603/v1/045ea0122d18ae78f8b43444.png"},{"id":90317076,"identity":"7475dad7-42e1-43b7-8ea8-2d54502e9334","added_by":"auto","created_at":"2025-09-01 10:25:02","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":300060,"visible":true,"origin":"","legend":"\u003cp\u003eThe main muscle force of the lumbar: (A-H) muscle force of the right external oblique abdominal muscle, the left external oblique abdominal muscle, the right internal oblique abdominal muscle, the left internal oblique abdominal muscle, the left iliocostalis, the right iliocostalis, the left multifidus muscle, and the right multifidus muscle, respectively. *Represents significant differences between HC and LBP groups (p \u0026lt; 0.05).\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7344603/v1/e579586c1c049aa3527e2e59.png"},{"id":90315839,"identity":"3f1814bb-92ab-4b4a-86bc-853ce52e5638","added_by":"auto","created_at":"2025-09-01 10:17:02","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":210895,"visible":true,"origin":"","legend":"\u003cp\u003eThe main muscle activation of the lumbar: (A-H) muscle activation of the right external oblique abdominal muscle, the left external oblique abdominal muscle, the right internal oblique abdominal muscle, the left internal oblique abdominal muscle, the left iliocostalis, the right iliocostalis, the left multifidus muscle, and the right multifidus muscle, respectively. (I) Represents a comparison between the left iliocostalis muscle activation and EMG data; (J) Represents a comparison between the right iliocostalis muscle activation and EMG data.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7344603/v1/51d05dc063e1e0e37aacc079.png"},{"id":90315836,"identity":"7c386bd2-3a25-4b93-ad64-c31bfa86ef73","added_by":"auto","created_at":"2025-09-01 10:17:02","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":224987,"visible":true,"origin":"","legend":"\u003cp\u003eLumbar spine and lower limb range of motion: \u0026nbsp;(A-F) The angle of the lumbar flexion-extension, the lumbar lateral bending, the lumbar axial rotation, the hip motion, the knee motion, and the ankle motion in a gait cycle, respectively. Lumbar extension is positive, flexion is negative, lateral flexion to the right is positive, lateral flexion to the left is negative, left rotation is positive, and right rotation is negative. The hip joint flexion is positive, extension is negative; knee joint flexion is negative, extension is positive; ankle joint dorsi flexion is positive, plantar flexion is negative. The shaded area represents ±1 standard deviation.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7344603/v1/b2f0c18789f77e57e0a71328.png"},{"id":90315837,"identity":"361b861b-dcc0-4235-8190-a1766a6e8352","added_by":"auto","created_at":"2025-09-01 10:17:02","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":335165,"visible":true,"origin":"","legend":"\u003cp\u003eIntervertebral Loading of each lumbar segment: (A–C) Corresponding to the shear force, compressive force, and twisting force of L3-4, respectively. (D–F) Corresponding to the shear force, compressive force, and twisting force of L4-5, respectively. (G–I) Corresponding to the shear force, compressive force, and twisting force of L5-S1, respectively. (J) The shaded area represents ±1 standard deviation. The bar chart shows the shear force, compressive force, and twisting force of each section of L3-S1.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-7344603/v1/cc579d6c7e2933d7d13704f7.png"},{"id":93520679,"identity":"114d85b0-af8a-4e91-b61f-cff0cb871597","added_by":"auto","created_at":"2025-10-14 17:46:35","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2337706,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7344603/v1/bba94032-19b1-44e6-bba5-d0b7553a7ffb.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Clinical and Biomechanical Effects of AI-Supervised Exercise Telerehabilitation on Patients with Nonspecific Chronic Low Back Pain: A Single-arm Clinical Trial","fulltext":[{"header":"Introduction","content":"\u003cp\u003eLow back pain (LBP) represents a major public health concern and is one of the leading causes of disability worldwide [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. It affects approximately one in every thirteen individuals, according to 2020 reports indicating personal experiences with the condition [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. We categorize LBP into specific and nonspecific types, with approximately 80 to 90% of cases being nonspecific. Non-specific chronic low back pain (NSCLBP) refers to discomfort persisting for over three months, not attributable to any identifiable disease or abnormal structural issue [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. This condition has been shown to have a significant impact on individuals suffering from persistent pain, characterized by a rigid lower back that restricts mobility, impairs balance, and diminishes proprioception. The inadequate consideration of biomechanical factors in lumbar and lower-extremity muscle pathology in NSCLBP may hinder the development of effective interventions [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eIn current clinical practice, NSCLBP is often a diagnosis of exclusion due to the lack of precise etiological identification [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. This ambiguity has led to controversies regarding the relative roles of biomechanical dysfunction, central sensitization, and psychological influences in pain persistence. Muscle pathology may be the initial alteration to spinal biomechanics and nociceptive input [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Prior research indicates that individuals with NSCLBP exhibit significantly reduced thickness changes in the transversus abdominis (TrA) and lumbar multifidus (LM) compared to asymptomatic individuals during voluntary tasks [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Additionally, intramuscular electromyography revealed a delayed anticipatory activation of TrA/LM fibers during trunk loading or limb movement [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. These studies focus solely on local lumbar biomechanics, neglecting the influence of movements in other body regions on the lumbar spine, leading to an incomplete analysis of motion. The lower limb plays a critical role in spinal function and stability. The stability of the spine is affected by the lower limb joints, particularly the hip, which influences posture and movement patterns in patients with NSCLBP [\u003cspan additionalcitationids=\"CR12 CR13\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. A study examined the potential benefits of incorporating hip strengthening exercises into conventional rehabilitation therapy for enhancing pain relief and reducing disability in individuals with low back pain [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Individuals experiencing acute low back pain exhibit altered coordination between the hip and spine during balance-related tasks, such as sitting on an unstable surface [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. The altered coordination may influence the balance performance of the overall lower extremity, suggesting the necessity for targeted interventions to improve coordination and stability of the lumbar region and the overall lower extremity. Recent advancements in the full-body lumbar spine (FBLS) model of OpenSim provide a detailed approach to musculoskeletal modeling [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. This helps us better understand the effects of various exercises on the lumbar region and the entire kinetic chain, which potentially improve the treatment strategies for patients with NSCLBP. Our previous study utilized the FBLS model to analyze the lumbar biomechanical changes during walking in patients with NSCLBP compared to healthy individuals, revealing a significant reduction in lumbar rotation angles, moments, muscle forces, and activation among the patient group .\u003c/p\u003e\u003cp\u003eIn the management of NSCLBP, analgesics are commonly used; however, their effectiveness is often restricted relative to placebo, and they may induce adverse effects [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Furthermore, other modalities such as superficial heat and cold, massage, acupuncture, and ultrasound may provide short-term relief; however, the evidence supporting their long-term effectiveness remains insufficient [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. Exercise therapies are commonly recommended for NSCLBP patients and have shown moderate improvements in pain relief and functional capacity [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. Recent studies have found that core strength training interventions effectively alleviated symptoms related to lumbar disc herniation, which may be due to optimized load distribution and reduced stress on the lumbar spine [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Improved strength and activation of deep core muscles may concurrently restore the physiological curvature of the lumbar spine, leading to an optimal load-bearing posture [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. However, the effectiveness of exercise depends upon individual patient characteristics and preferences [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e] indicating that a personalized approach may provide the best outcome. Achieving lasting therapeutic effects in CNSLBP requires consistent and ongoing supervised exercise interventions to ensure optimal results The accurate execution of exercise movements is crucial, as improper technique can reduce effectiveness and increase the risk of secondary injuries [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e] Supervised exercise programs are advocated for NSCLBP patients due to their ability to ensure proper execution, facilitate individualized modifications, and offer immediate feedback on movement accuracy [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e] Hospital-based supervised exercises encounter several barriers, including transportation difficulties, scheduling conflicts, financial limitations, and shortages of healthcare personnel [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e] The challenges lead to an overloaded rehabilitation system, limiting access to essential services, especially for older adults [\u003cspan additionalcitationids=\"CR27\" citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eVirtual rehabilitation offers exercise programs directly to patients\u0026rsquo; homes, potentially reducing barriers to attendance and program completion [\u003cspan additionalcitationids=\"CR26 CR27\" citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. Virtual exercises can yield clinical outcomes similar to in-clinic supervised exercise programs by effectively decreasing pain and improving lumbar function through core muscle strengthening. Video-based guidance faces challenges in delivering real-time motion corrections [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan additionalcitationids=\"CR30\" citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e] Recent advancements in artificial intelligence (AI) and real-time motion capture technology[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e] have facilitated the development of AI-supervised exercise training that employs a standard household tablet\u0026rsquo;s camera for dynamic tracking of patients\u0026rsquo; movements. Our study uses the human key-point recognition technology to monitor the trajectory of exercises. It provides immediate feedback for posture correction and replicates the benefits of supervised in-clinic training, such as enhanced posture stability and injury prevention. The integration of AI into exercise programs to automate different training phases has the potential to improve the quality of home-based care.\u003c/p\u003e\u003cp\u003eTherefore, this study explores whether early-stage personalized training that targets the abnormal biomechanics of the lumbar spine and lower limbs can effectively improve patients\u0026rsquo; symptoms and abnormal biomechanics. This research aims to evaluate the efficacy of a remote AI-supervised virtual exercise program (RAVEP) in improving clinical outcomes and biomechanical metrics in individuals with NSCLBP. In addition to conventional scaling using anthropometric data, this study intended to adopt ultrasound imaging-derived muscle thickness to calibrate the hill-type muscle parameters of each patient in the FBLS model of OpenSim. Therefore, this personized FBLS model can accurately analysis of alterations in lumbar and lower limb kinematics, lumbar moments and loads, as well as muscle activation patterns during walking after the intervention. The study has the potential to improve our understanding of lumbar spine biomechanics following a virtually supervised exercise program and provides important theoretical insights for the remote management of NSCLBP.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eParticipants\u003c/h2\u003e\u003cp\u003eFrom July to December 2023, fourteen patients with NSCLBP (4 males, 10 females; mean age 25.71\u0026thinsp;\u0026plusmn;\u0026thinsp;6.79 years; weight 56.60\u0026thinsp;\u0026plusmn;\u0026thinsp;10.04 kg; height 162.64\u0026thinsp;\u0026plusmn;\u0026thinsp;5.13 cm; BMI 21.32\u0026thinsp;\u0026plusmn;\u0026thinsp;3.13) were recruited from Zhujiang Hospital, Southern Medical University. The Ethics Committee of Zhujiang Hospital, Southern Medical University approved the study (Approval Number: 2023-KY-017), and it was registered with the Chinese Clinical Trial Registry (Registration Number: ChiCTR2300073185). All participants provided written informed consent before participating in the study, which included consent for the publication of any potentially identifiable images or data. Inclusion criteria adhered to the diagnostic guidelines for NSCLBP by the American College of Physicians and the American Pain Society [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e] Fourteen patients were enrolled based on the following criteria: (1) Participants provided written informed consent and voluntarily engaged in the study; (2) had a clinical diagnosis of NSCLBP lasting over three months with a Visual Analog Scale (VAS) score exceeding thirty; (3) reported no history of back trauma, lumbar fractures, lumbar spondylolisthesis, ankylosing spondylitis, spinal deformities, spinal tuberculosis, spinal tumors, sacroiliac joint dysfunction, lumbar degenerative changes, lumbar disc herniation, spinal stenosis, osteoporosis, or other conditions causing low back pain, and no history of lumbar or abdominal surgery in the past two years; (4) had not utilized nonsteroidal anti-inflammatory drugs (NSAIDs) or other medications in the past six months; and (5) were aged between 18 and 65 years old. The exclusion criteria were as follows: (1) inability to walk independently or the presence of an abnormal gait; (2) diagnosis of severe systemic diseases (e.g., acute infection, malignant tumors, cardiovascular disease, acute cerebrovascular events, mental disorders, significant visual or auditory impairments, or other conditions considered inappropriate for vigorous exercise); (3) pregnancy.\u003c/p\u003e\u003cp\u003eThe sample size was calculated utilizing the pwr.t.test function in R (version 4.3.2, released on October 31, 2023). Previous studies of remote intervention therapy indicated an assumed effect size of Hedge\u0026rsquo;s g\u0026thinsp;=\u0026thinsp;0.88, with a statistical power of 1\u0026thinsp;\u0026minus;\u0026thinsp;β\u0026thinsp;=\u0026thinsp;0.81, and a significance level of α\u0026thinsp;=\u0026thinsp;0.05. A paired-sample, two-tailed t-test revealed a requirement of at least 12 participants. The final sample size was adjusted to 14 participants to account for a 10% dropout rate.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eInterventions\u003c/h3\u003e\n\u003cp\u003eThis study used a remote AI-supervised virtual exercise program (RAVEP) as an intervention. The program was delivered via the WeChat mini-program \u0026ldquo;RAVEP\u0026rdquo; developed by Beijing Yinshan Future Health Technology Co., Ltd. Participants accessed the program from their home or any other preferred location using a smartphone or tablet. The mini-program employs AI-driven image recognition to detect 32 critical body points, such as the head, torso, and limbs. This makes it enables precise tracking of exercise movements with continuous monitoring of the motion trajectories of the head, torso, and limbs. It checks for accuracy by the analysis of the angle and length ratios among key points.\u003c/p\u003e\u003cp\u003eThis mini-program has two main sections: \u0026ldquo;Assessment\u0026rdquo; and \u0026ldquo;Training\u0026rdquo;. Participants can access the \u0026ldquo;Assessment\u0026rdquo; section via a QR code. This allows them to complete a self-assessment that collects their medical history, NRS scores, ODI responses, and Roland-Morris questionnaire results. Participants automatically receive an evaluation report along with a tailored exercise prescription. The exercise selection in the \u0026ldquo;Training\u0026rdquo; section follows established protocols and aligns with the 2021 American Physical Therapy Association (APTA) Low Back Pain Guidelines [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e] A standardized exercise prescription was implemented for all participants, encompassing a weekly regimen of breathing techniques, core muscle strengthening, flexibility training, and McKenzie exercises according to the abnormal biomechanics of each patient according to the results of OpenSim analysis. Prior to each session, participants engaged in a preparatory meeting online video meeting, where therapists provided detailed explanations of the forthcoming exercises and addressed essential topics relevant to NSCLBP education, such as posture, pain management, etiology, and lumbar anatomy. The intensity and difficulty of exercises progressively escalate from week 1 to week 4. Participants first identify a suitable location, access the RREP mini-program, select their assigned exercise course, and assume a position on a yoga mat, either standing or lying down, while placing their smartphone horizontally on a stand approximately 1.5 meters away. The program delivers audio instructions and displays instructional exercise videos, while an AI recognition frame in the top-right corner of the screen concurrently monitors motion in real time, offering guidance to ensure proper form (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Following each exercise, the program delivered an immediate performance rating based on key-point coincidence ratios, categorized as \u0026ldquo;Pass,\u0026rdquo; \u0026ldquo;Good,\u0026rdquo; \u0026ldquo;Excellent,\u0026rdquo; or \u0026ldquo;Outstanding\u0026rdquo; to promote patient adherence. At the end of each session, participants completed the rate of perceived exertion scale and received a summary report detailing their completion rates, thus enabling the evaluation of exercise efficacy. During the second to fourth weeks, participants conducted a follow-up assessment via the mini-program, leading to adjustments in their exercise prescriptions for those indicating discomfort.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\n\u003ch3\u003eEvaluation\u003c/h3\u003e\n\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003eUltrasound measurement\u003c/h2\u003e\u003cp\u003eThe Terason uSmart 3300 ultrasound instrument was utilized to measure the thickness of the TrA and LM muscles in both resting and contracting states [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e] During the TrA measurement, patients were positioned supine with hips flexed to approximately 135 degrees and knees flexed to about 90 degrees. Participants were instructed to breathe in a calm manner. A linear probe was positioned above the iliac crest, anterior to the axillary line, to assess muscle thickness at the end of quiet inspiration, thereby inhibiting TrA activation and enabling the acquisition of the relaxed-TrA image. Participants were directed to fully exhale and progressively contract their abdomen as much as possible to obtain the activated-TrA image prior to the image acquisition process. Muscle thickness refers to the measurement between the hyperechoic fascial layers of the muscle. During the LM measurement, patients were positioned prone with a thin pillow placed under the abdomen to ensure neutral alignment of the lumbar spine. The contraction image was obtained by muscle activation; the participants lifted the contralateral arm off the bed and maintained it at 120 degrees of shoulder abduction and 90 degrees of elbow flexion. The LM muscle thickness was measured with a curved probe, defined as the distance between the dorsal edge of the L4-L5 zygapophyseal joint and the thoracolumbar hyperechoic fascia, which separates the muscle from the subcutaneous fat, excluding the fascia itself. All participants underwent ultrasound examinations and muscle thickness assessments conducted by the same experienced physician both before and after the intervention. Measurements were performed three times on each side, and the mean value was calculated for data analysis.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eMotion capture and electromyography measurement\u003c/h3\u003e\n\u003cp\u003ePrior to the test, participants received instructions to wear tight-fitting clothing and were provided with guidance on performing the walking tests in the designated area. Participants first assumed an anatomical position, characterized by an upright posture, facing forward, eyes level, feet together with toes pointing forward, and arms resting at the sides with palms forward, to collect static data. Participants were then instructed to walk across the force plates at a self-comfortable pace. After each pass, participants turned and returned to the starting point at a comfortable pace, repeating this process five times. A successful trial was characterized by the simultaneous contact of both feet with the two force plates during a gait cycle.\u003c/p\u003e\u003cp\u003eKinematic data were recorded utilizing a 6-camera infrared 3D motion capture system (BTS SMART-DX EVO2) at a sampling frequency of 100 Hz. A total of forty-nine reflective markers, each with a diameter of 14.0 mm, were affixed to the bodies of the participants. Both static and dynamic tests demonstrated an error margin of less than 0.1 mm within a volume of 4 m \u0026times; 3 m \u0026times; 3 m (L \u0026times; W \u0026times; H), comparable to other commercial biomechanics systems. Ground reaction forces (GRF) were recorded using two force plates (BTS P6000) at a sampling frequency of 1,000 Hz. Surface electromyography (EMG) data were obtained from the bilateral erector spinae and iliocostalis muscles using the BTS FREEEMG 300 system. The sampling frequency was 1,000 Hz, and the electrodes were placed according to SENIAM guidelines. The 3D kinematic, GRF, and surface EMG data were calibrated and synchronized within the BTS system to ensure that the data were collected at the same time.\u003c/p\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eMusculoskeletal modeling and simulation\u003c/h2\u003e\u003cp\u003eGait data were imported into OpenSim software (version 4.1) for kinematic and dynamic analysis using the Full Body Lumbar Spine (FBLS) model developed by Raabe and Chaudhari [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e] The model was scaled to correspond with each subject\u0026rsquo;s height, weight, and body proportions. Reflective marker trajectories were imported into OpenSim, and inverse kinematics were utilized to compute motion data. The model\u0026rsquo;s kinematics were modified using the reduced residuals algorithm to improve dynamic consistency with the measured GRF and joint moments. The process produced an adjusted model and new sets of inverse kinematics and inverse dynamics results.\u003c/p\u003e\u003cp\u003eIn the FBLS model of OpenSim, the LM is modeled as a circular cross-sectional, nonlinear spring structure, with muscle thickness represented by the cross-sectional diameter. Muscle thickness was assessed using the cross-sectional diameter of the LM model, and alterations in resting cross-sectional area were evaluated before and after the intervention. The model demonstrates a relationship between muscle cross-sectional area and maximum isometric contraction force. The post-intervention model was personalized by adjusting the maximum isometric contraction force of the muscles based on the ratio of each subject\u0026rsquo;s muscle cross-sectional diameter, measured through ultrasound measurements conducted before and after the intervention [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]\u003c/p\u003e\u003cp\u003eStatic Optimization was subsequently performed to assign net joint moments, derived from inverse dynamics, to specific muscle fibers on a frame-by-frame basis. This process generated muscle forces and activation levels throughout the motion. Muscle activation occurred when the calculated activation exceeded 5% [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e] The GRF data, adjusted muscle force, and kinematic data were input into the joint reaction module to calculate the compressive loads and shear forces at the L3-L4 and L4-L5, and L5-S1 segments. This study adjusted the scaling factor for the maximum isometric contraction force in the OpenSim muscle model according to alterations in resting muscle area, determined by the ratio of post-intervention to pre-intervention muscle area measured via musculoskeletal ultrasound.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003eStatistical analysis\u003c/h2\u003e\u003cp\u003eStatistical analyses were conducted using IBM SPSS (version 25). The Chi-square test was employed to examine the gender variable. The Shapiro-Wilk test was used to check the normality for continuous variables, whereas the Levene test was used to evaluate homogeneity of variance. A one-way repeated measures analysis of variance (ANOVA) was performed to evaluate variations in muscle thickness, VAS, flexion-extension and lateral bending peak moments, and muscle peak forces before and after the intervention. The mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation was used to report all outcomes that were continuous variables. A significance level (α) of 0.05 was set for all statistical comparisons.\u003c/p\u003e\u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003ePain and functional assessment\u003c/h2\u003e\u003cp\u003eThe study findings demonstrated substantial decreases in Visual Analogue Scale (VAS) scores for worst pain experienced in the past week, average pain in the past week, and current pain levels after the intervention (P\u0026lt;0.01). Furthermore, significant decreases were observed in the Oswestry Disability Index (ODI) and Roland-Morris Disability Questionnaire (RMDQ) scores (P \u0026lt;0.01), as well as in Timed Up and Go (TUG) test times (P \u0026lt;0.01) following the intervention. Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e presents the detailed questionnaire results for NSCLBP participants before and after the intervention.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eQuestionnaire before and after intervention (X\u0026thinsp;\u0026plusmn;\u0026thinsp;S) Note: The symbol \u0026ldquo;α\u0026rdquo; denotes the use of non-parametric tests, while paired t-tests are used for other statistical analyses.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eVariable\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePre-intervention (n\u0026thinsp;=\u0026thinsp;14)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePost-intervention(n\u0026thinsp;=\u0026thinsp;14)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eworst VAS score in the past week(mm)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e50\u0026thinsp;\u0026plusmn;\u0026thinsp;13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e29\u0026thinsp;\u0026plusmn;\u0026thinsp;13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ethe average VAS score recorded over the past week\u003csup\u003eα\u003c/sup\u003e(mm)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e39\u0026thinsp;\u0026plusmn;\u0026thinsp;08\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e2.0\u0026thinsp;\u0026plusmn;\u0026thinsp;12\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ethe current VAS score\u003csup\u003eα\u003c/sup\u003e(mm)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e34\u0026thinsp;\u0026plusmn;\u0026thinsp;05\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e14\u0026thinsp;\u0026plusmn;\u0026thinsp;13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eODI\u003csup\u003eα\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e7.8\u0026thinsp;\u0026plusmn;\u0026thinsp;3.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e5.3\u0026thinsp;\u0026plusmn;\u0026thinsp;2.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRMDQ\u003csup\u003eα\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e5.1\u0026thinsp;\u0026plusmn;\u0026thinsp;4.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e2.0\u0026thinsp;\u0026plusmn;\u0026thinsp;1.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTUG (s)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e10.4\u0026thinsp;\u0026plusmn;\u0026thinsp;1.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e8.5\u0026thinsp;\u0026plusmn;\u0026thinsp;0.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003eMuscle thickness, muscle force, and activation\u003c/h2\u003e\u003cp\u003eUltrasound measurements (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) demonstrated that four weeks of RAVEP training significantly increased the resting thickness of the bilateral LM and TrA in participants with NSCLBP (p\u0026lt;0.01); however, the contracted thickness of the left LM did not show a significant change (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05). The contracted thickness of the bilateral TrA and the right LM exhibited a significant increase (p\u0026lt;0.01). The contraction ratio of the bilateral LM and the left TrA showed a statistically significant improvement (p\u0026lt;0.01); however, the contraction rate of the bilateral TrA did not show a significant change (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eMuscle thickness was measured both before and after the intervention(X\u0026thinsp;\u0026plusmn;\u0026thinsp;S)\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003epre-intervention\u003c/p\u003e\u003cp\u003e(n\u0026thinsp;=\u0026thinsp;14)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003epost-intervention\u003c/p\u003e\u003cp\u003e(n\u0026thinsp;=\u0026thinsp;14)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eP\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eThe resting thickness of the left TrA (mm)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e3.3\u0026thinsp;\u0026plusmn;\u0026thinsp;0.9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e3.7\u0026thinsp;\u0026plusmn;\u0026thinsp;0.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eThe resting thickness of the right TrA (mm) \u003csup\u003eα\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e3.2\u0026thinsp;\u0026plusmn;\u0026thinsp;0.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e3.7\u0026thinsp;\u0026plusmn;\u0026thinsp;0.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eThe resting thickness of the left LM (mm) \u003csup\u003eα\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e26.2\u0026thinsp;\u0026plusmn;\u0026thinsp;4.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e27.1\u0026thinsp;\u0026plusmn;\u0026thinsp;4.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eThe resting thickness of the right LM (mm)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e25.4\u0026thinsp;\u0026plusmn;\u0026thinsp;3.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e27.3\u0026thinsp;\u0026plusmn;\u0026thinsp;3.6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eThe contracted thickness of the left TrA (mm) \u003csup\u003eα\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e4.7\u0026thinsp;\u0026plusmn;\u0026thinsp;1.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e5.3\u0026thinsp;\u0026plusmn;\u0026thinsp;1.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eThe contracted thickness of the right TrA (mm)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e4.7\u0026thinsp;\u0026plusmn;\u0026thinsp;1.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e5.4\u0026thinsp;\u0026plusmn;\u0026thinsp;1.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eThe contracted thickness of the left LM (mm) \u003csup\u003eα\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e34.5\u0026thinsp;\u0026plusmn;\u0026thinsp;4.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e38.4\u0026thinsp;\u0026plusmn;\u0026thinsp;5.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.14\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eThe contracted thickness of the right LM (mm)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e34.1\u0026thinsp;\u0026plusmn;\u0026thinsp;4.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e39.9\u0026thinsp;\u0026plusmn;\u0026thinsp;4.9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eThe contraction ratio of the left TrA (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e43.8\u0026thinsp;\u0026plusmn;\u0026thinsp;17.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e46.0\u0026thinsp;\u0026plusmn;\u0026thinsp;14.9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.732\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eThe contraction ratio of the right TrA (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e48.7\u0026thinsp;\u0026plusmn;\u0026thinsp;19.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e48.8\u0026thinsp;\u0026plusmn;\u0026thinsp;20.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.956\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eThe contraction ratio of the left LM (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e32.7\u0026thinsp;\u0026plusmn;\u0026thinsp;12.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e42.7\u0026thinsp;\u0026plusmn;\u0026thinsp;11.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.018\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eThe contraction ratio of the right LM (%)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c2\"\u003e\u003cp\u003e34.8\u0026thinsp;\u0026plusmn;\u0026thinsp;12.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e46.7\u0026thinsp;\u0026plusmn;\u0026thinsp;11.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;0.01\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"4\"\u003eNote: The symbol \u0026ldquo;α\u0026rdquo; denotes the use of non-parametric tests, while paired t-tests are used for other statistical analyses.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eThe variation in muscle force during walking was similar in both the pre- and post-intervention phases. The OpenSim static optimization model revealed significant increases in the maximum muscle forces of the bilateral LM following the intervention (p\u0026lt;0.05) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eG, H). Other simulated muscle forces did not show statistically significant changes (P\u0026gt;0.05); however, curve-fitting analysis revealed that during 60%-80% of the gait cycle, the force of the left external oblique abdominal muscles experienced a slight decrease (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB), while the forces of the left internal oblique abdominal muscles (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD) and right external oblique abdominal muscles (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA) showed slight increases. The simulated muscle activation patterns (Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eI and \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eJ) closely aligned with the observed trends in muscle force. Analysis of muscle activation (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA\u0026ndash;H) demonstrated increased activation of the bilateral LM, right internal oblique, and right external oblique muscles during 60\u0026ndash;80% of the gait cycle. Conversely, the activation of the left external oblique and internal oblique muscles exhibited a slight decrease during the same period. However, these alterations in muscle activation did not reach statistical significance (P\u0026thinsp;\u0026gt;\u0026thinsp;0.05).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003eKinematics\u003c/h2\u003e\u003cp\u003ePatients with NSCLBP exhibited consistent alterations in lumbar flexion, lateral flexion, and rotation angles while walking, both before and after the intervention (Figs.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA-C). Curve-fitting analysis indicated a reduction in the sagittal plane motion angle of the lumbar spine in patients with NSCLBP after the intervention. The maximum lumbar sagittal plane motion angle decreased from \u0026minus;\u0026thinsp;0.23\u0026thinsp;\u0026plusmn;\u0026thinsp;0.22\u0026deg; pre-intervention to -0.15\u0026thinsp;\u0026plusmn;\u0026thinsp;0.14\u0026deg; post-intervention; however, this difference was not statistically significant (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05). Figure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA illustrates the lumbar motion in the sagittal plane during walking, with fitted curve analysis indicating a reduction in motion among NSCLBP patients post-intervention. Figures\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC depict lumbar motion in the coronal and axial planes, respectively, before and after the intervention.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD\u0026ndash;F presents the alterations in hip, knee, and ankle joint angles during walking prior to and following the intervention. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD details hip joint motion during walking under both pre- and post-intervention conditions. The peak sagittal plane knee flexion angle in patients with NSCLBP decreased from 69.42\u0026thinsp;\u0026plusmn;\u0026thinsp;4.13\u0026deg; before the intervention to 68.54\u0026thinsp;\u0026plusmn;\u0026thinsp;4.89\u0026deg; after the intervention (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05). Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE illustrates the detailed results of knee joint motion during walking. After the intervention, participants with NSCLBP demonstrated decreased dorsiflexion and plantarflexion angles at the ankle. The peak ankle joint range of motion declined from 16.95\u0026thinsp;\u0026plusmn;\u0026thinsp;5.62\u0026deg; prior to the intervention to 15.15\u0026thinsp;\u0026plusmn;\u0026thinsp;4.76\u0026deg;following the intervention, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eF (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003eLumbar intervertebral loading\u003c/h2\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e presents the simulation curves depicting changes in the intervertebral compression force, sagittal shear force, and twisting force at the L3-L4, L4-L5, and L5-S1 segments during a single gait cycle, both prior to and following the intervention. Curve-fitting analysis indicated that the peak shear forces at the L3-4, L4-5, and L5-S1 segments in NSCLBP participants were 0.77\u0026thinsp;\u0026plusmn;\u0026thinsp;0.29, 0.90\u0026thinsp;\u0026plusmn;\u0026thinsp;0.31, and 1.07\u0026thinsp;\u0026plusmn;\u0026thinsp;0.36 times body weight, respectively, before the intervention, and 0.78\u0026thinsp;\u0026plusmn;\u0026thinsp;0.22, 0.83\u0026thinsp;\u0026plusmn;\u0026thinsp;0.21, and 1.02\u0026thinsp;\u0026plusmn;\u0026thinsp;0.29 times body weight, respectively, after the intervention (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05). The shear force at the L3-4 segment exhibited a minor increase, whereas reductions were observed at the L4-5 and L5-S1 segments. Figures\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA, \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD, and \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eG illustrate the alterations in lumbar shear forces during walking before and after the intervention. Curve-fitting analysis revealed that the peak compressive forces at the L3-4, L4-5, and L5-S1 segments in NSCLBP participants were 3.33\u0026thinsp;\u0026plusmn;\u0026thinsp;0.59, 3.49\u0026thinsp;\u0026plusmn;\u0026thinsp;0.64, and 3.70\u0026thinsp;\u0026plusmn;\u0026thinsp;0.68 times body weight, respectively, prior to the intervention, and decreased to 3.18\u0026thinsp;\u0026plusmn;\u0026thinsp;0.33, 3.32\u0026thinsp;\u0026plusmn;\u0026thinsp;0.23, and 3.50\u0026thinsp;\u0026plusmn;\u0026thinsp;0.25 times body weight, respectively, following the intervention (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05). Reductions in compressive forces were observed across all three lumbar segments. Figures\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB, \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE, and \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eH present the detailed changes in lumbar compressive forces during walking before and after the intervention. Curve-fitting analysis indicated that the peak torsional moments at the L3-4, L4-5, and L5-S1 segments in NSCLBP participants were 1.12\u0026thinsp;\u0026plusmn;\u0026thinsp;0.22, 1.24\u0026thinsp;\u0026plusmn;\u0026thinsp;0.23, and 1.30\u0026thinsp;\u0026plusmn;\u0026thinsp;0.24 times body weight, respectively, before the intervention, and subsequently decreased to 1.04\u0026thinsp;\u0026plusmn;\u0026thinsp;0.14, 1.16\u0026thinsp;\u0026plusmn;\u0026thinsp;0.16, and 1.23\u0026thinsp;\u0026plusmn;\u0026thinsp;0.18 times body weight, respectively, after the intervention (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05). All three lumbar segments showed decreases in torsional moments. Figures\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC, \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eF, and \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eI depict the specific alterations in lumbar torsional moments during walking, both pre- and post-intervention. However, these differences did not achieve statistical significance (p\u0026thinsp;\u0026gt;\u0026thinsp;0.05).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study evaluates the efficacy of RAVEP training in patients with NSCLBP by analyzing clinical outcomes and biomechanical factors, such as kinematics, dynamics, lumbar load, and muscle force. A biomechanical analysis of pre- and post-intervention data indicated significant improvement in dynamics, muscle force, and overall clinical performance.\u003c/p\u003e\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\u003ch2\u003eClinical performance\u003c/h2\u003e\u003cp\u003ePost-RAVEP training demonstrated a significant improvement in pain and functional outcome for NSCLBP patients. Functional tests revealed significant decreases in patients\u0026rsquo; weekly worst, average, and current pain levels, as well as overall pain alleviation. The ODI and RMDQ scores decreased dramatically, indicating that NSCLBP had a smaller impact on daily activities. Furthermore, shorter TUG test durations indicated better lower limb mobility after the intervention. These findings suggest that RAVEP training significantly reduces pain and improves functional performance in patients with NSCLBP. These findings align with prior research showing that both isometric and isotonic core stabilization exercises significantly alleviate pain and disability in patients with NSCLBP compared to no intervention [\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e] Just as the recent guidelines also recommended the various exercise therapies for NSCLBP patients, which have shown moderate improvements in pain relief and functional capacity [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThis study then showed that NSCLBP patients exhibited significant bilateral increases in LM thickness and TrA thickness in both resting and contracted states, as well as a marked increase in contraction rate in bilateral LM after the RAVEP training. Similar results were also seen in a comparative study that reported a notable increase in the thickness of the TrA and LM, along with enhanced contractile capacity, after an eight-week core training program in healthy people [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e] The results indicated that extended core training leads to significant muscle hypertrophy in the core region, thereby increasing muscle cross-sectional area and enhancing force production. Core exercise significantly improves the biomechanical structure of the lumbar spine through this adaptation. This enhancement probably contributes to the noted advancements in lumbar biomechanics subsequent to RAVEP training.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\u003ch2\u003eMuscle force and activation\u003c/h2\u003e\u003cp\u003ePrior studies have shown that patients with chronic LBP often exhibit atrophy or fatty infiltration in the lumbar multifidus and other deep core muscles [\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e], leading to functional impairment, delayed activation, and inadequate recruitment. This dysfunction may arise from persistent compensatory overactivation of superficial muscles due to inhibition of deep core muscles, leading to an inefficient stabilization pattern. OpenSim simulation results indicated notable post-intervention increases in bilateral LM forces among NSCLBP patients, with increases of 50.9% and 45.4%, respectively (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Consistent with the ultrasound analysis and muscle strength simulations revealed notable increases in LM muscle thickness, contraction rate, and strength were revealed among participants. The findings indicated that muscle hypertrophy alone does not fully explain the observed strength gains; rather, enhanced muscle recruitment and activation patterns during walking seem to be the main factors following the intervention. These findings also suggested that RAVEP training induced morphological alterations and reactivated levels in deep core muscles. Previous research supports this perspective, indicating that core stability training reeducates motor control in inhibited muscles, potentially improving muscle activation and positively influencing the management of chronic LBP [\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e] The simulation of LM activation before the intervention showed that the left limb stance phase had greater activation relative to the right. This could be because the right-side lumbar pain caused the trunk imbalance. Following the intervention, LM activation during the right limb stance phase exhibited increased symmetry with the left, suggesting that core stability training effectively improved muscle coordination, trunk control, and movement patterns.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec18\" class=\"Section2\"\u003e\u003ch2\u003eKinematics\u003c/h2\u003e\u003cp\u003eThis study found that after four weeks of the RAVEP training, patients with NSCLBP exhibited a decrease in lumbar spine joint angle during flexion-extension while walking, consistent with previous research findings.43\u0026ndash;46 There are several explanations for our finds. Core strengthening exercises, as employed in RAVEP training, can effectively enhance lumbar spine stability [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e] and improve spinal alignment to decrease the lumbar spine movement during daily activities. Greater core strength improves proprioception and postural control. This can reduce the strain and pain in the lower back [\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e] After four weeks of RAVEP training, the symmetry in coronal plane range of motion and axial rotation was improved in the lumbar spine. A reduction of 4.4% in the axial rotation angle was noted, suggesting improved rotational symmetry. The lateral stability of the lumbar spine during postural changes may significantly influence reported outcomes for patients. Prior studies indicated that core stabilization exercises positively influence the lumbar lordotic angle in patients with lumbar disc degeneration, suggesting improvements in spinal alignment and stability [\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e] These exercises markedly decrease sagittal translation at L4-L5 and L5-S1 levels. In the coronal plane, the lumbar lateral bending angle (swing amplitude) was smaller than that observed after RAVEP training. This suggests that the lateral stability of the lumbar spine was improved, consistent with a previous study involving adult idiopathic scoliosis, which demonstrated that a core exercise program could decrease the Cobb angle, reflecting an improvement in spinal alignment in the coronal plane [\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e]\u003c/p\u003e\u003cp\u003eDuring the stance phase of the gait cycle, the hip joint showed no changes before and after RAVEP training. In the swing phase, the hip joint angle increased post-intervention relative to pre-intervention measurements, consistent with previous research findings47-49. This may relate to symptomatic individuals utilizing compensatory movements and modified load-sharing strategies, or reduced coordination abilities of the lumbar spine concerning the hip joint. Hip-spine syndrome exemplifies the compensatory mechanisms that arise from restricted hip motion, resulting in heightened lumbar spine loading and a potential increase in low back pain [\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e] Core strengthening may enhance hip range of motion and alleviate symptoms related to restricted hip mobility; however, its efficacy could be constrained by compensatory movements and altered coordination strategies. The knee joint angle alterations in lower limb joints during the swing phases showed improvement following RAVEP training, aligning with prior research findings. A previous study demonstrated that patients with NSCLBP exhibit modified gait patterns, marked by reduced hip and knee motion across various planes during distinct phases of the gait cycle [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e] This reduction in range of motion may result in an increased knee joint angle at initial landing during the support phase, consistent with the findings of our study. NSCLBP patients exhibit a slower preferred walking speed [\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e], which may be attributed to the knee joint angle at the support phase as a compensatory strategy to sustain balance and alleviate pain. After RAVEP training, patients with NSCLBP exhibited a decrease in plantar flexion and dorsal extension during mid-stance, potentially reversing the altered ankle movement pattern in the sagittal plane associated with NSCLBP. Prior research indicated that this compensatory strategy of the ankle serves to prevent downward displacement of the center of gravity and to reduce mechanical load on the lumbar spine\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e\u003ch2\u003eLumbar intervertebral loading\u003c/h2\u003e\u003cp\u003eThe results concerning lumbar compression and shear forces demonstrated a distinct bimodal pattern, intricately linked to the single-leg standing phase of the left and right feet within the gait cycle. After four weeks of RAVEP training, participants showed a minor rise in shear force at the L3-L4 segment. However, the shear forces at the L4-L5 and L5-S1 segments displayed a significant reduction, coupled with a pronounced improvement in LM strength. This finding aligns with biomechanical studies suggesting that the weakness of core muscles may lead to instability at the L4-L5 spinal segment. Numerous studies indicated that deep muscle fibers are essential for maintaining lumbar stability [\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e], especially during flexion, and likely regulate lumbar shear forces [\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e]. An experimental finding demonstrated that the resultant joint force, considering shear force allowance, is oriented towards a particular region within the vertebral body [\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e] We proposed that increased strength and activation of the deep core muscles generate a greater posterior stabilizing force along the lumbar spine, thus preserving physiological lumbar curvature, improving overall lumbar stability, and counteracting anterior-directed mechanical forces. Therefore, the aforementioned stabilizing effect likely explains the observed reductions in shear forces at the L4-L5 and L5-S1 segments. The shear force at the L3-L4 segment was unexpectedly elevated. Prior research has underscored the important role of the psoas muscle in lumbar motion and stability, specifically by its application of compressive forces on the anterior aspect of the L3-L4 segment [\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e] The heightened shear force at the L3-L4 segment may result from adaptations in the psoas muscle, as core stability training may improve its compressive function at the anterior lumbar spine, subsequently increasing the shear force at this segment.\u003c/p\u003e\u003cp\u003eLumbar intervertebral compression is a key biomechanical factor that influences the structural integrity and functionality of the spine. The application of compressive forces on intervertebral discs induces various mechanical and physiological alterations [\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e] Following the four-week RAVEP training, a decline in lumbar compression forces was observed across all spinal segments, though this decline did not reach statistical significance(p\u0026thinsp;\u0026gt;\u0026thinsp;0.05). Consistent with prior research, improvements in the muscles and soft tissue function within the kinetic chain after the intervention may improve mechanical force transmission [\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e], which we suggest as a key factor in the reduction of lumbar compression forces observed.\u003c/p\u003e\u003cp\u003eLumbar torque decreased after the intervention, likely attributable to alterations in kinematics. Lumbar torque primarily arises from the swinging motion of the upper body and pelvis during walking, generating rotational forces on the intervertebral discs that may lead to degenerative changes. Similar results were seen in a previous study that employed surface electromyography to assess core muscle activity during exercises has shown that strengthening these muscles can increase lumbar torque [\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e], thereby enhancing overall functional performance and reducing the risk of musculoskeletal injuries. The strength and activation of the left internal oblique and right external oblique muscles after the four-week RAVEP training, which enhanced the lumbar spine\u0026rsquo;s resistance to rotational forces and increased rotational stability, contributed directly to a reduction in lumbar torque.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec20\" class=\"Section2\"\u003e\u003ch2\u003eLimitations and future work\u003c/h2\u003e\u003cp\u003eThis study evaluates lumbar loading during walking following the four-week RAVEP training, utilizing pre- and post-intervention data from patients with NSCLBP and a modified FBLS model that integrates ultrasound-derived muscle measurements. However, the study presents several limitations. Firstly, OpenSim offers a range of preset mannequins for simulation purposes. However, these simplified models may not sufficiently capture the intricacies of real human motion. Secondly, this study highlighted the feasibility of remote AI-supervised virtual exercises for managing NSCLBP, influenced by factors such as smartphone usability and acceptance of tele-rehabilitation. Thirdly, the majority of participants were young adults aged 21 to 27 years, a demographic typically experiencing mild lumbar degeneration, which may be mitigated by enhanced muscle activity. Future research should include participants from a wider age, and motion spectrum.\u003c/p\u003e\u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study indicated that RAVEP training markedly enhanced deep core muscle function, restored lumbar range of motion, decreased mechanical forces on the lumbar spine, and alleviated pain symptoms. The AI-guided real-time feedback component of RAVEP training improved the precision and coordination of local muscle contractions. The enhancements in lumbar muscle function enabled the restoration of optimal spinal posture and motion patterns, reduced excessive mechanical forces on the lumbar spine, alleviated pain-inducing factors, and ultimately improved outcomes for NSCLBP.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eLBP: Low back pain\u003c/p\u003e\n\u003cp\u003eNSCLBP: Nonspecific chronic low back pain\u003c/p\u003e\n\u003cp\u003eTrA: Transversus abdominis\u003c/p\u003e\n\u003cp\u003eLM: Lumbar multifidus\u003c/p\u003e\n\u003cp\u003eFBLS: Full-body lumbar spine\u003c/p\u003e\n\u003cp\u003eAI: Artificial intelligence\u003c/p\u003e\n\u003cp\u003eRAVEP: Remote AI-supervised virtual exercise program\u003c/p\u003e\n\u003cp\u003eVAS: Visual Analog Scale\u003c/p\u003e\n\u003cp\u003eNSAIDs: Nonsteroidal anti-inflammatory drugs\u003c/p\u003e\n\u003cp\u003eAPTA: American Physical Therapy Association\u003c/p\u003e\n\u003cp\u003eGRF: Ground reaction forces\u003c/p\u003e\n\u003cp\u003eEMG: Electromyography\u003c/p\u003e\n\u003cp\u003eODI: Oswestry Disability Index\u003c/p\u003e\n\u003cp\u003eRMDQ: Roland-Morris Disability Questionnaire\u003c/p\u003e\n\u003cp\u003eTUG: Timed Up and Go\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study has the formal ethical approval from the Ethics Committee of Zhujiang Hospital of Southern Medical University (2023-KY-017).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eStudy concept and design: GH, MZ, QX; Acquisition of data: PL, ZZ, JH; Analysis and interpretation of data: PL, ZZ, JH; Drafting of the manuscript: PL, ZZ, MZ; Substantively revising: QZ, TF. All authors contributed to the review and revision of the manuscript. All authors have consented the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the grants from Guangzhou Municipal Science and Technology Bureau (2023A04J0444), Southern Medical University Youth Incubation Program (PY2018N057), Natural Science Foundation of China (82171174, 82371184), the Natural Science Foundation of China under Grant (52105305), Natural Science Foundation of Guangdong Province (2025A1515012782), and the Key Technologies R\u0026amp;D Program of Guangdong Province (2023B0303020003).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets generated during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eCollaborators GDAI. Global burden of 369 diseases and injuries in 204 countries and territories, 1990-2019: A systematic analysis for the Global Burden of Disease Study 2019. Lancet 2020;396: 1204-22.\u003c/li\u003e\n\u003cli\u003eMoser I, Sommer I, Gartlehner G. [who guideline for Non-Surgical management of chronic primary low back pain in adults in primary and community care settings]. Gesundheitswesen 2024;86: 715-22.\u003c/li\u003e\n\u003cli\u003eBittar M, Deodhar A. A critical view of WHO guidelines on management of low back pain. Nat Rev Rheumatol 2024;20: 201-2.\u003c/li\u003e\n\u003cli\u003eKoes BW, van Tulder MW, Thomas S. Diagnosis and treatment of low back pain. BMJ 2006;332: 1430-4.\u003c/li\u003e\n\u003cli\u003eGianola S, Bargeri S, Del Castillo G, Corbetta D, Turolla A, Andreano A, et al. Effectiveness of treatments for acute and subacute mechanical non-specific low back pain: A systematic review with network meta-analysis. Br J Sports Med 2022;56: 41-50.\u003c/li\u003e\n\u003cli\u003eZafar T, Zaki S, Alam MF, Sharma S, Babkair RA, Nuhmani S, et al. Effect of Whole-Body vibration exercise on pain, disability, balance, proprioception, functional performance and quality of life in people with Non-Specific chronic low back pain: A systematic review and Meta-Analysis. J Clin Med 2024;13.\u003c/li\u003e\n\u003cli\u003eKnezevic NN, Candido KD, Vlaeyen JWS, Van Zundert J, Cohen SP. Low back pain. Lancet 2021;398: 78-92.\u003c/li\u003e\n\u003cli\u003eBalasch-Bernat M, Willems T, Danneels L, Meeus M, Goubert D. Differences in myoelectric activity of the lumbar muscles between recurrent and chronic low back pain: A cross-sectional study. BMC Musculoskelet Disord 2021;22: 756.\u003c/li\u003e\n\u003cli\u003ePinto SM, Boghra SB, Macedo LG, Zheng Y, Pang MYC, Cheung JPY, et al. Does motor control exercise restore normal morphology of lumbar multifidus muscle in people with low back pain? - a systematic review. J Pain Res 2021;14: 2543-62.\u003c/li\u003e\n\u003cli\u003eWong AYL, Parent EC, Funabashi M, Stanton TR, Kawchuk GN. Do various baseline characteristics of transversus abdominis and lumbar multifidus predict clinical outcomes in nonspecific low back pain? A systematic review. Pain 2013;154: 2589-602.\u003c/li\u003e\n\u003cli\u003eOshima Y, Watanabe N, Takeoka T, Kajikawa Y, Yotsumoto T, Majima T. Preoperative abnormal posture improves after total knee arthroplasty: Knee-Hip-Spine syndrome. J Arthroplasty 2025;40: 416-22.\u003c/li\u003e\n\u003cli\u003eGovil G, Tomar L, Dhawan P. Knee-Spine syndrome: Management dilemma when knee osteoarthritis coexists with spine degeneration. Cureus 2022;14: e24939.\u003c/li\u003e\n\u003cli\u003eZambarano E, Glaviano N, Bouillon L, Norte G, Murray A. Effect of exhaustive exercise on Lumbopelvic-Hip complex stability, muscle activity, and movement patterns. J Electromyogr Kinesiol 2024;74: 102852.\u003c/li\u003e\n\u003cli\u003eZawadka M, Smolka J, Skublewska-Paszkowska M, Lukasik E, Zielinski G, Bys A, et al. Altered squat movement pattern in patients with chronic low back pain. Ann Agric Environ Med 2021;28: 158-62.\u003c/li\u003e\n\u003cli\u003ede Jesus FLA, Fukuda TY, Souza C, Guimaraes J, Aquino L, Carvalho G, et al. Addition of specific hip strengthening exercises to conventional rehabilitation therapy for low back pain: A systematic review and meta-analysis. Clin Rehabil 2020;34: 1368-77.\u003c/li\u003e\n\u003cli\u003eAlshehri MA, van den Hoorn W, Klyne DM, Hodges PW. Coordination of hip and spine in individuals with acute low back pain during unstable sitting. Spine J 2024;24: 768-82.\u003c/li\u003e\n\u003cli\u003evon Arx M, Liechti M, Connolly L, Bangerter C, Meier ML, Schmid S. From stoop to squat: A comprehensive analysis of lumbar loading among different lifting styles. Front Bioeng Biotechnol 2021;9: 769117.\u003c/li\u003e\n\u003cli\u003eZhang Z, Zou J, Lu P, Hu J, Cai Y, Xiao C, et al. Analysis of lumbar spine loading during walking in patients with chronic low back pain and healthy controls: An OpenSim-Based study. Front Bioeng Biotechnol 2024;12: 1377767.\u003c/li\u003e\n\u003cli\u003eJones CMP, Day RO, Koes BW, Latimer J, Maher CG, McLachlan AJ, et al. Opioid analgesia for acute low back pain and neck pain (the OPAL trial): A randomised placebo-controlled trial. Lancet 2023;402: 304-12.\u003c/li\u003e\n\u003cli\u003eChiarotto A, Koes BW. Nonspecific low back pain. N Engl J Med 2022;386: 1732-40.\u003c/li\u003e\n\u003cli\u003eChoo YJ, Chang MC. The effect of exercise on stabilizing and strengthening core muscles for patients with symptomatic herniated lumbar disc: A systematic review and meta-analysis. Asian J Surg 2024;47: 1703-4.\u003c/li\u003e\n\u003cli\u003eMiyachi R, Miyazaki J. Relationship between lumbar motor control ability and spinal curvature in elderly individuals. Healthcare (Basel) 2020;8.\u003c/li\u003e\n\u003cli\u003eMataran-Penarrocha GA, Lara Palomo IC, Antequera Soler E, Gil-Martinez E, Fernandez-Sanchez M, Aguilar-Ferrandiz ME, et al. Comparison of efficacy of a supervised versus non-supervised physical therapy exercise program on the pain, functionality and quality of life of patients with non-specific chronic low-back pain: A randomized controlled trial. Clin Rehabil 2020;34: 948-59.\u003c/li\u003e\n\u003cli\u003eLopez-Marcos JJ, Diaz-Arribas MJ, Valera-Calero JA, Navarro-Santana MJ, Izquierdo-Garcia J, Ortiz-Gutierrez RM, et al. The added value of Face-to-Face supervision to a therapeutic Exercise-Based app in the management of patients with chronic low back pain: A randomized clinical trial. Sensors (Basel) 2024;24.\u003c/li\u003e\n\u003cli\u003eAbedi A, Colella TJF, Pakosh M, Khan SS. Artificial intelligence-driven virtual rehabilitation for people living in the community: A scoping review. NPJ Digit Med 2024;7: 25.\u003c/li\u003e\n\u003cli\u003eDay MA, Ciol MA, Mendoza ME, Borckardt J, Ehde DM, Newman AK, et al. The effects of telehealth-delivered mindfulness meditation, cognitive therapy, and behavioral activation for chronic low back pain: A randomized clinical trial. BMC Med 2024;22: 156.\u003c/li\u003e\n\u003cli\u003eCosta F, Janela D, Molinos M, Moulder RG, Lains J, Bento V, et al. Digital rehabilitation for acute low back pain: A prospective longitudinal cohort study. J Pain Res 2022;15: 1873-87.\u003c/li\u003e\n\u003cli\u003eFatoye F, Gebrye T, Fatoye C, Mbada CE, Olaoye MI, Odole AC, et al. The clinical and Cost-Effectiveness of telerehabilitation for people with nonspecific chronic low back pain: Randomized controlled trial. JMIR Mhealth Uhealth 2020;8: e15375.\u003c/li\u003e\n\u003cli\u003eEkambaram D, Ponnusamy V. Real-Time monitoring and assessment of rehabilitation exercises for low back pain through interactive dashboard pose analysis using streamlit\u0026mdash;a pilot study. Electronics (Basel) 2024;13: 3782.\u003c/li\u003e\n\u003cli\u003eFritz JM, Minick KI, Brennan GP, McGee T, Lane E, Skolasky RL, et al. Outcomes of telehealth physical therapy provided using Real-Time, videoconferencing for patients with chronic low back pain: A longitudinal observational study. Arch Phys Med Rehabil 2022;103: 1924-34.\u003c/li\u003e\n\u003cli\u003eLewkowicz D, Slosarek T, Wernicke S, Winne A, Wohlbrandt AM, Bottinger E. Digital therapeutic care and decision support interventions for people with low back pain: Systematic review. JMIR Rehabil Assist Technol 2021;8: e26612.\u003c/li\u003e\n\u003cli\u003eChou R, Qaseem A, Snow V, Casey D, Cross JTJ, Shekelle P, et al. Diagnosis and treatment of low back pain: A joint clinical practice guideline from the American College of Physicians and the American Pain Society. Ann Intern Med 2007;147: 478-91.\u003c/li\u003e\n\u003cli\u003eGeorge SZ, Fritz JM, Silfies SP, Schneider MJ, Beneciuk JM, Lentz TA, et al. Interventions for the management of acute and chronic low back pain: Revision 2021. J Orthop Sports Phys Ther 2021;51: G1-60.\u003c/li\u003e\n\u003cli\u003eDjordjevic O, Djordjevic A, Konstantinovic L. Interrater and intrarater reliability of transverse abdominal and lumbar multifidus muscle thickness in subjects with and without low back pain. J Orthop Sports Phys Ther 2014;44: 979-88.\u003c/li\u003e\n\u003cli\u003eRaabe ME, Chaudhari AMW. Biomechanical consequences of running with deep core muscle weakness. J Biomech 2018;67: 98-105.\u003c/li\u003e\n\u003cli\u003eRaabe ME, Chaudhari AMW. An investigation of jogging biomechanics using the full-body lumbar spine model: Model development and validation. J Biomech 2016;49: 1238-43.\u003c/li\u003e\n\u003cli\u003eMcCrady AN, Masterson CD, Barnes LE, Scharf RJ, Blemker SS. Development of an ultrasound-based metric of muscle functional capacity for use in patients with neuromuscular disease. Muscle Nerve 2024;70: 1205-14.\u003c/li\u003e\n\u003cli\u003eLi M, Wang K, Niu W, Zhang S. A musculoskeletal modeling of hand-foot crawling with different heights. J Bionic Eng 2020;17: 591-9.\u003c/li\u003e\n\u003cli\u003eKhaledi A, Gheitasi M. Isometric vs isotonic core stabilization exercises to improve pain and disability in patients with non-specific chronic low back pain: A randomized controlled trial. Anesth Pain Med 2024;14: e144046.\u003c/li\u003e\n\u003cli\u003eZheng Y, Ke S, Lin C, Li X, Liu C, Wu Y, et al. Effect of core stability training monitored by rehabilitative ultrasound image and surface electromyogram in local core muscles of healthy people. Pain Res Manag 2019;2019: 9130959.\u003c/li\u003e\n\u003cli\u003eBalevi Batur E. Assessing the correlation between the lumbar disc herniation degree and multifidus muscle fatty degeneration in chronic low back pain. Cureus 2024;16: e71237.\u003c/li\u003e\n\u003cli\u003eMohammadi V, Letafatkar A, Jafarnezhadgero AA. Effects of core stability training on kinematic and kinetic variables in patients with chronic low back pain. Physical Treatments-Specific Physical Therapy Journal 2023;13: 55-66.\u003c/li\u003e\n\u003cli\u003eFrizziero A, Pellizzon G, Vittadini F, Bigliardi D, Costantino C. Efficacy of core stability in Non-Specific chronic low back pain. J Funct Morphol Kinesiol 2021;6.\u003c/li\u003e\n\u003cli\u003eSalik Sengul Y, Yilmaz A, Kirmizi M, Kahraman T, Kalemci O. Effects of stabilization exercises on disability, pain, and core stability in patients with non-specific low back pain: A randomized controlled trial. Work 2021;70: 99-107.\u003c/li\u003e\n\u003cli\u003eShalamzari MH, Henteh MA, Shamsoddini A, Ghanjal A. Comparison of the effects of core stability and whole-body electromyostimulation exercises on lumbar lordosis angle and dynamic balance of sedentary people with hyperlordosis: A randomized controlled trial. BMC Sports Sci Med Rehabil 2024;16: 91.\u003c/li\u003e\n\u003cli\u003eDimitrijevic V, Viduka D, Scepanovic T, Maksimovic N, Giustino V, Bianco A, et al. Effects of Schroth method and core stabilization exercises on idiopathic scoliosis: A systematic review and meta-analysis. Eur Spine J 2022;31: 3500-11.\u003c/li\u003e\n\u003cli\u003eHuppert A, Ambrosio L, Nwosu K, Pico A, Russo F, Vadala G, et al. Previous lumbar spine fusion increases the risk of dislocation following total hip arthroplasty in patients with hip-spine syndrome: A systematic review and meta-analysis. BMC Musculoskelet Disord 2024;25: 732.\u003c/li\u003e\n\u003cli\u003eYalfani A, Asgarpoor A. Comparison of cognitive functional therapy and neurofeedback training on kinetic gait in patients with chronic non-specific low back pain: A randomised controlled trial. Disabil Rehabil 2025: 1-10.\u003c/li\u003e\n\u003cli\u003eXiao F, Maas H, van Dieen JH, Pranata A, Adams R, Han J. Chronic non-specific low back pain and ankle proprioceptive acuity in community-dwelling older adults. Neurosci Lett 2022;786: 136806.\u003c/li\u003e\n\u003cli\u003ePadwal J, Berry DB, Hubbard JC, Zlomislic V, Allen RT, Garfin SR, et al. Regional differences between superficial and deep lumbar multifidus in patients with chronic lumbar spine pathology. BMC Musculoskelet Disord 2020;21: 764.\u003c/li\u003e\n\u003cli\u003eBastos De Oliveira V, Albuquerque Brandao MC, Coelho De Albuquerque Pereira W, Fernandes De Oliveira L. Lumbar multifidus layers stiffness at L5-S1 level in prone and sitting posture measured by shear wave elastography. J Back Musculoskelet Rehabil 2024;37: 1241-8.\u003c/li\u003e\n\u003cli\u003eKim K, Kim YH, Lee S. Shear force allowance in lumbar spine under follower load in neutral standing posture. Acta Bioeng Biomech 2010;12: 49-53.\u003c/li\u003e\n\u003cli\u003eSeyedhoseinpoor T, Taghipour M, Dadgoo M, Ebrahimi Takamjani I, Sanjari MA, Kazemnejad A, et al. Relationship between the morphology and composition of the lumbar paraspinal and psoas muscles and lumbar intervertebral motion in chronic low-back pain: An exploratory study. Clin Anat 2022;35: 762-72.\u003c/li\u003e\n\u003cli\u003eEremina G, Smolin A, Xie J, Syrkashev V. Development of a computational model of the mechanical behavior of the L4-L5 lumbar spine: Application to disc degeneration. Materials (Basel) 2022;15.\u003c/li\u003e\n\u003cli\u003eClael S, Campos LF, Correia KL, de Lucena JMS, Gentil P, Durigan JL, et al. Exercise interventions can improve muscle strength, endurance, and electrical activity of lumbar extensors in individuals with non-specific low back pain: A systematic review with meta-analysis. Sci Rep 2021;11: 16842.\u003c/li\u003e\n\u003cli\u003eArvanitidis M, Jimenez-Grande D, Haouidji-Javaux N, Falla D, Martinez-Valdes E. Eccentric exercise-induced delayed onset trunk muscle soreness alters high-density surface EMG-torque relationships and lumbar kinematics. Sci Rep 2024;14: 18589.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Nonspecific chronic low back pain, AI-supervised virtual exercise, Opensim, Ultrasound imaging","lastPublishedDoi":"10.21203/rs.3.rs-7344603/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7344603/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e\u003cp\u003eNonspecific chronic low back pain (NSCLBP) is a major cause of disability worldwide, and the long-term efficacy of traditional treatments is limited. Although highly recommended, core stabilization exercises are often limited by adherence and access to supervised programs. Recent advances in artificial intelligence (AI) have created new opportunities for remote, real-time supervision of exercise. This study aims to evaluate the clinical and biomechanical effects of a remote AI-supervised virtual exercise program (RAVEP) on patients with NSCLBP.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eFourteen patients with NSCLBP underwent a four-week RAVEP intervention, which was delivered via a smartphone application that integrated human key-point recognition and real-time feedback. The program incorporated breathing, core strengthening, and McKenzie exercises. The effects before and after the intervention were assessed using ultrasound imaging of the lumbar multifidus (LM) and transversus abdominis (TrA), gait analysis using 3D motion capture and surface electromyography, and musculoskeletal modeling using a full-body lumbar spine model in OpenSim. The model was scaled using patient-specific anthropometric data, and the muscle properties were calibrated using ultrasound-derived muscle thickness to assess personalized lumbar kinematics, intervertebral loads, and muscle forces and activations. In addition, the evaluation of pain and clinical function was conducted using the Visual Analog Scale (VAS), Oswestry Disability Index (ODI), Roland-Morris Disability Questionnaire (RMDQ), and Timed Up and Go (TUG) tests.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eFollowing the RAVEP, patients with NSCLBP showed significant improvements in both pain (VAS) and clinical function (ODI, RMDQ, TUG) (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01). Ultrasound assessment revealed increased LM and TrA thickness, as well as contraction ratios after the intervention. The results of the OpenSim simulation indicated significant increases in LM force (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05), and decreased compressive and torsional loads at L4-L5 and L5-S1 during gait. Kinematic results showed more symmetrical lumbar movement. Increased activation of deep core muscles was found; however, the differences in activation did not reach statistical significance.\u003c/p\u003e\u003ch2\u003eDiscussion\u003c/h2\u003e\u003cp\u003eOur study showed that the RAVEP could significantly improve the pain, functioning, and core muscle performance in NSCLBP patients. AI-supervised real-time feedback in the RAVEP helped patients to perform their home exercises properly, which improved lower back stability and lessened stress on the spine. Personized musculoskeletal modeling and simulation provide a new understanding of how virtual rehabilitation works. This study demonstrates that RAVEP may serve as an effective and accessible strategy for the management of NSCLBP.\u003c/p\u003e","manuscriptTitle":"Clinical and Biomechanical Effects of AI-Supervised Exercise Telerehabilitation on Patients with Nonspecific Chronic Low Back Pain: A Single-arm Clinical Trial","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-01 10:16:57","doi":"10.21203/rs.3.rs-7344603/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"a9b53519-abd5-427f-aa72-732a6249a911","owner":[],"postedDate":"September 1st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-10-14T17:38:22+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-01 10:16:57","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7344603","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7344603","identity":"rs-7344603","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}