Enhanced Time-Frequency Analysis of sEMG signals using Superlet Transform

The superlet transform (SLT) has recently received increasing attention in the literature. SLT is a signal processing tool for time-frequency (TF) analysis designed to mitigate the inherent trade-off between temporal and frequency resolution. While SLT has been successfully applied to several biological signals, including electroencephalography, electrocardiogram and photoplethysmography, its performance in surface electromyography (sEMG) has not yet been systematically assessed. In this study, we evaluated the multiplicative adaptive SLT using simulated sEMG burst (1 and 0.5 s), employing Short-Time Fourier Transform (STFT) and Continuous Wavelet Transform (CWT) as benchmarks. Transient detection was quantified through mean absolute error in time (MAEt) and frequency (MAEf), and TF resolution was further determined via full width at half maximum (FWHM). In addition, one experimental sEMG burst (0.28 s) recorded during human gait was analyzed to provide qualitative evaluation of TF representations and to examine intrinsic variability of spectral and temporal estimates. Results from simulated data showed that SLT provided a balanced compromise between temporal and spectral resolution, matching STFT in frequency estimation and improving temporal characterization for moderate parameter configurations. SLT consistently outperformed CWT in frequency, while maintaining comparable temporal resolution. Furthermore, SLT exhibited stable performance across signal durations, frequencies and parameter variation. FWHM analysis revealed that SLT degrades temporal information less than CWT when increasing the number of cycles, enabling a better detection of fast transient oscillatory events. Qualitative analysis of the experimental sEMG confirmed clearer TF representations and reduced variability. Overall, SLT emerges as a robust and versatile tool for sEMG analysis, although its effectiveness depends on parameter selection and is still limited by computational demands. Future work should focus on optimizing parameters and improving computational efficiency to establish SLT as a practical and robust approach for sEMG research and clinical applications.