Abstract: The feasibility of rapidly and objectively evaluating the flavor of pork balls using an electronic nose was investigated. The odor of four pork meatballs with different proportions of lean (100%, 90%, 80% and 70%) was analyzed by the electronic nose, and their volatile compounds were determined by head space-solid phase microextraction (HS-SPME) coupled to gas chromatography-mass spectrometry (GC-MS). Besides, the aroma intensity was scored by a trained sensory panel. Both linear discriminant analysis (LDA) and neural network algorithm (NNA) were used to identify and classify the electronic nose data. Partial least squares regression (PLSR) was applied to correlate the aroma compounds identified by electronic nose sensors and GC-MS data. The results showed that there were significant differences in the aroma scores of four meatball samples (P < 0.01). The meatballs with higher proportion of fat had higher sensory scores. Analysis by LDA and NNA demonstrated that the aroma profiles of all four samples could be discriminated by the electronic nose. A total of 67 volatile aroma compounds were detected by GC-MS, mainly including aldehydes, alcohols and ketones. Differences in volatile flavor compounds were mainly responsible for the differences in sensory scores. The PLSR model established exhibited good correlation between the electronic nose data and the relative contents of major volatile compounds. The prediction model established by stepwise regression (R2 > 0.9, P < 0.01) indicated that the electronic nose can be used to predict the aroma of meatballs with reliable results.

Key words: pork ball flavor, sensory evaluation, electronic nose, neural network, gas chromatography-mass spectrometry, partial least squares regression