The use of the impedance measurements to distinguish between fresh and frozen–thawed chicken breast muscle

• Impedance was able to distinguish the fresh chicken breasts from frozen–thawed ones.
• We used learning vector quantization neural network (LVQNN) in this study.
• LVQNN could predict the frozen–thawed cycles of chicken breasts.
• We compared the results based on magnitude, phase angle and the mixture of them.
• Mixture index could predict the frozen–thawed cycles of chicken breasts better.

An impedance system was built to differentiate fresh chicken breasts from those that had been frozen and thawed. Inserting needle electrode pairs of the detecting probe aligned with the longitudinal direction of muscle myofibers (PL) gave more satisfactory results. Learning vector quantization neural network (LVQNN) and partial least square-discriminant analysis (PLS-DA) were employed to acquire the prediction accuracy. The results demonstrated that the model using LVQNN achieved a satisfactory prediction accuracy, with a discrimination accuracy for fresh breasts of 100%. Additionally, the recognition results for a single frozen–thawed cycle were greater than 90%, and for two cycles were greater than 88%. The values obtained from PLS-DA were somewhat lower than for LVQNN, being 100% for fresh samples, in excess of 90% for single frozen–thawed cycle and more than 84% for those that had been multiple frozen–thawed. In conclusion, these results showed that the impedance system is a simple and effective application for the discrimination of fresh chicken breasts from frozen–thawed ones.