Participants
Seven female adolescents aged between 12 and 14 years were recruited through a school screening program during which the angle of trunk rotation (ATR) in the thoracic and lumbar regions was assessed using the Adam’s forward bending test. For adolescents who showed possible signs of scoliosis (an ATR 3), during school screening, their spinal condition was further evaluated by ultrasound imaging using ScolioscanTM29,30 to measure the spinal deformity, which was evaluated again after completing the biofeedback posture training. Female adolescents were excluded from the study if they had a history of surgical or orthotic treatment for AIS. The curve type was based on the Lenke classification system 31. The demographic data of the seven recruited female adolescents are provided in Table 1. The study followed the tenets of the Declaration of Helsinki as developed by the World Medical Association, and the research protocol was approved by the Human Subjects Ethics Sub-committee of the Hong Kong Polytechnic University. All adolescents participated voluntarily, and informed assent from the adolescents and written informed consent for participation from their parents were solicited before beginning the study, as per institutional guidelines.
sEMG measurement before and after training
The parameters of the sEMG values were formulated based on the standards for the Surface EMG for Noninvasive Assessment of Muscles (SENIAM) 32. The myoelectrical activities of the paraspinal muscles of the adolescents were measured in terms of sEMG activities using a preamplified sensor, MyoScan (model T9503M), and a data acquisition system, Flexcomp (model T7555M; both acquired from Thought Technology, Montreal, Canada), within 1 week before starting the posture training and after completing the training. Briefly, the sEMG electrodes were placed onto the paraspinal muscles, namely the trapezius, latissimi dorsi, thoracic erector spinae, and lumbar erector spinae, in pairs to determine the muscle activities along the spine in habitual sitting postures. The sEMG signals were sampled at a rate of 2048 Hz with a 10–500-Hz band pass filter and a 60-Hz notch filter to eliminate artefacts and noise. The measurements were taken three times for 3 minutes each time. The RMS sEMG values for each trial were calculated using BioGraph Infiniti software, and the scores for three trials were averaged. The RMS sEMG ratio for each tested muscle pair of each participant was then calculated using the following equation 16:
𝑅𝑀𝑆 𝑠𝐸𝑀𝐺 𝑅𝑎𝑡𝑖𝑜 = 𝑅𝑀𝑆 𝑠𝐸𝑀𝐺 (𝑐𝑜𝑛𝑣𝑒𝑥)/𝑅𝑀𝑆 𝑠𝐸𝑀𝐺 (𝑐𝑜𝑛𝑐𝑎𝑣𝑒)
This ratio is an index of symmetry in the myoelectric activities of the tested paraspinal muscle pairs in habitual sitting postures 10. A ratio of 1 indicates that the tested muscle pair has symmetric sEMG activity between the concave and convex sides.
Biofeedback posture training
An sEMG biofeedback posture training protocol was delivered using Thought Technology BioGraph Infiniti software (Montreal, Canada). Each adolescent received 30 sessions of training delivered in one to two sessions per week for approximately 6 months. During each training session, the adolescents underwent a 10-minute EMG setup procedure with an impedance check, which was followed by the baseline measurement. During the baseline measurement, the participants were required to sit for 3 minutes in a relaxed state to measure their muscle activities in terms of sEMG signals. During posture training, they were instructed to sit in an ideal recommended posture and maintain this posture for 5 minutes. Animated indicators on the biofeedback screen appeared whenever the posture fell below the threshold of specific individualized requirements. These indicators were aimed at reducing the RMS sEMG ratio between the left and right sides of the four pairs of muscles—the trapezius, latissimi dorsi, thoracic erector spinae, and lumbar erector spinae—toward 1 and the individual sEMG values of each of those muscles to less than 5 µV. As 12 indexes (4 ratios and 8 individual values) were required to be measured to activate the animated indicators, the exact threshold was individually tailored for an overall success rate of approximately 80%, which also motivated the adolescents to engage in the training. This posture training routine was administered five times in each session, with a 2-minute resting period between each time. Each training session lasted approximately 60 minutes.
Statistics
All statistical analyses were performed using the SPSS Statistic Version 25 program for Windows (Armonk, NY: IBM Corp). The RMS sEMG ratios were subjected to one-sample t-tests to identify any asymmetry (deviation from the test value of 1) in sEMG activities over the trapezius, latissimi dorsi, thoracic erector spinae, and lumbar erector spinae before and after the biofeedback training. A repeated measures analysis of variance (ANOVA) was performed to compare the symmetry in the sEMG activities of the four paraspinal muscle pairs before and after the biofeedback training, followed by a post-hoc pairwise comparison if the multivariate results were significant. The level of significance was set at p < 0.05. The difference in the degree of the spinal curve angle from baseline to post-training was measured using ScolioscanTM (Telefield Medical Imaging Limited, Hong Kong).