Enhancing Accuracy of 2D ¹H–¹³C HSQC NMR Spectra Under Low Sampling Rates via Non-Harmonic Analysis and Band Segmentation

Multidimensional nuclear magnetic resonance (NMR) spectroscopy is a powerful tool for analyzing molecular structures; however, its practical application is hindered by prolonged acquisition times and reduced resolution due to spectral overlap. Although non-uniform sampling (NUS) offers a partial solution by reducing measurement duration, conventional reconstruction methods struggle at ultra-low sampling rates, often producing artifacts and missing spectral features. This study introduces a novel hybrid framework that integrates non-harmonic analysis (NHA) with a band-segmented processing strategy—termed non-uniform non-harmonic analysis (NUNHA)—for high-fidelity two-dimensional NMR spectrum reconstruction under NUS conditions. NHA enables grid-free, continuous frequency adaptation, grids, allowing for the precise extraction of subtle signals while minimizing artifacts. By incorporating the band-segmented methodology, the framework effectively reduces cross-band interference and achieves high-quality spectral reconstruction—even at ultra-low sampling rates (for example, 6%). Experimental results demonstrate that NUNHA yields spectral resolutions comparable to conventional high-rate sampling methods, enabling accurate recovery and enhanced analysis of complex molecular structures using NMR spectroscopy. These findings highlight the potential of NUNHA for accelerating high-throughput metabolomics and enabling real-time structural analysis in future multidimensional NMR applications.