Methods: In this study, 267 matched pairs of AIS and controls were recruited. The participants underwent EMG measurements at their first presentation and did not receive any treatment for 6 months at which point they underwent EMG and radiographs. Early curve progression was defined as >5° in Cobb angle at 6 months. The root mean square of the EMG (rms-EMG) signal was recorded with the participants in sitting and back extension. The rms-EMG ratio at the upper end vertebrae, apical vertebrae (AV), and lower end vertebrae (LEV) of the major curve was calculated. Results: The rms-EMG ratio in the scoliosis cohort was high compared with that in the controls (sitting: 1.2±0.3 vs. 1.0±0.1, p<0.01; back extension: 1.1±0.2 vs. 1.0±0.1, p<0.01). An AV rms-EMG ratio in back extension, with a cutoff threshold of ≥1.5 in the major thoracic curve and ≥1.3 in the major lumbar curve, was a risk factor for early curve progression after 6 months without treatment (odds ratio, 4.1; 95% confidence interval, 2.8–5.9; p<0.01). Increases in side deviation (SD) (distance between the AV and the central sacral line) were related to a higher rms-EMG ratio in LEV of the major thoracic curve (baseline: rs=0.2, p=0.03; 6 months: rs=0.3, p<0.01). Conclusions An EMG discrepancy was detected in the scoliosis cohort, which was related to increases in SD in the major thoracic curve. The AV rms-EMG ratio in back extension was correlated with curve progression after 6 months of no treatment.