Rapid quantification of phenolic content and antioxidant activity in cookies produced with amazonian palm fruit flour using Micro-NIR spectrometer and PLS regression

Abstract

There are several reports of the potential benefits of phenolic compound (PC) in food products, due to their antioxidant activities (AC). However, in recent years, new research results have demonstrated that PC has potential health risks due to the reduction in absorption of protein nutrients and cytotoxic effects. The PC and AC quantifications are laborious and time-consuming methods, therefore it is necessary to develop simple, fast and precise method to determine these parameters, not only in the raw materials, but also in food products. Therefore, this study focused on the potential of Micro-NIR spectrometer data modeled with partial least square regression to predict PC and AC in processed food (cookies) prepared with peach palm (PP), that is rich in PC. The cookies were prepared using 12.5 to 100 % of PP flour in substitution to wheat flour (WF). The NIR model for AC, determined by the ferric reducing antioxidant power (FRAP) method, shows R2cv = 0.93 (regression coefficient of cross-validation step); RMSECV = 0.05; R2p = 0.87 (regression coefficient of prediction step); RMSEP = 0.04; RPD = 2.73, and by 2,2-azinobis (3-ethylbenzothiazoline-6-sulfonic acid) radical capture (ABTS) exhibit R2cv = 0.83; RMSECV = 3.72; R2p = 0.70; RMSEP = 4.12; RPD = 1.76, and for PC, determined by Folin-Ciocalteu, shows R2cv = 0.86; RMSECV = 0.44; R2p = 0.80; RMSEP = 0.43; RPD = 2.04. These excellent results, mainly for FRAP and PC, demonstrated that portable NIR spectrometers could be a fast, simple and reliable method to predict PC and AC in cookies prepared with different proportion of PP flour and WF. Similar models can also be developed to predict PC and AC in other food products.