Process-Modulated Flavor Formation and Establishment of Predictive Modeling for Aroma in Spicy Anchovies.

Research on spicy anchovies lacks dedicated sensory frameworks, reliable aroma identification, and systematic processing-flavor insights. In this study, 21 spicy anchovy samples with different processing parameters were selected as research objects. The effects of process modifications on the sensory attributes and aroma composition of spicy anchovies were investigated through sensory evaluation and aroma analysis. A product-specific flavor wheel (5 modalities, 136 terms) with 17 key descriptors was built via Quantitative Descriptive Analysis. GC-O combined with AEDA/AECA identified 13 key aroma compounds in the commercial sample. HS-SPME-GC-MS detected 73 volatiles across all samples, among which olefins (34 species) were dominant and their formation was linked to lipid oxidation and high-temperature processing. Odor activity values and sensory data revealed that a frying temperature of 180 °C promoted nonanal and (E)-β-ocimene to enhance "fried seafood aroma"; Xiaomila chili pepper boosted "initial spiciness" via capsaicin; and high Sichuan pepper masked "fishy off-flavor" via linalyl acetate. A prediction model for aroma sensory attributes was established and the prediction correlations for "braised beef in soy sauce aroma" and "fried seafood aroma" were relatively high (r = 0.90 and 0.96, respectively). This study provides theoretical guidance for the flavor improvement of spicy anchovies.