Early detection and genetic insights into salt tolerance in rice by linking physiology and spectral reflectance during seed germination.

Soil salinization threatens global agriculture, and seed germination is especially salt-sensitive. Although seed imbibition is a critical early phase, salt responses during this stage remain poorly understood. Reliance on final morphological traits limits nondestructive early phenotyping, thereby hindering efficient screening and mechanistic insight. To overcome this limitation, we integrated continuous physiological measurements with hyperspectral data to model dynamic responses of key indicators and further leveraged hyperspectral-derived traits for early prediction and genetic dissection of salt tolerance. The results showed that: (1) Salt stress induced a distinct temporal cascade of physiological changes during seed imbibition. Proline accumulation was the earliest response, initiating at 6 h, followed by marked increases in abscisic acid (ABA) and malondialdehyde (MDA) by 12 h, while gibberellin (GA) levels and α-amylase activity were strongly suppressed by 24 h. (2) Integrating hyperspectral reflectance with PLSR and principal component analysis (PCA) enabled high-throughput, real-time physiological monitoring and nondestructive early salt-tolerance classification. PLSR simultaneously estimated five physiological traits (R