Handheld NIR device: A non-targeted approach to assess authenticity of fish fillets and patties

This study evaluates the reliability of a handheld NIR device in distinguishing fillets and patties of Atlantic cod (n = 80) from those of haddock (n = 90), in comparison with a FT-NIR benchtop spectrometer. The authentication issue was faced by Linear Discriminant Analysis (LDA) and Soft Independent Modelling of Class Analogy (SIMCA), pre-treating spectral data with different algorithms, and validating models both internally and externally. The best LDA models gave 100% correct classification in prediction. Sensitivity > 65% and specificity > 74% in prediction were calculated for the best SIMCA models. No significant differences (P > .05) were found between the two instruments by McNemar test. Thus, the work demonstrated that a handheld NIR device can be a simple, cost-effective, and reliable alternative to benchtop spectrometers in fish fillet and patty authentication. These important findings can help in improving commercial fraud fight, extending the possibility to authenticate fish species also in processed products.