Assessment of Table Olives’ Organoleptic Defect Intensities Based on the Potentiometric Fingerprint Recorded by an Electronic Tongue

Abstract

Table olives are prone to the appearance of sensory defects that decrease their quality and in some cases result in olives unsuitable for consumption. The evaluation of the type and intensity of the sensory negative attributes of table olives is recommended by the International Olive Council, although not being legally required for commercialization. However, the accomplishment of this task requires the training and implementation of sensory panels according to strict directives, turning out in a time-consuming and expensive procedure that involves a degree of subjectivity. In this work, an electronic tongue is proposed as a taste sensor device for evaluating the intensity of sensory defects of table olives. The potentiometric signal profiles gathered allowed establishing multiple linear regression models, based on the most informative subsets of signals (from 24 to 29 recorded during the analysis of olive aqueous pastes and brine solutions) selected using a simulated annealing meta-heuristic algorithm. The models enabled the prediction of the median intensities (R 2 ≥ 0.942 and RMSE ≤ 0.356, for leave-one-out or repeated K-fold cross-validation procedures) of butyric, musty, putrid, winey-vinegary, and zapateria negative sensations being, in general, the predicted intensities within the range of intensities perceived by the sensory panel. Indeed, based on the predicted mean intensities of the sensory defects, the electroche micalchemometric approach developed could correctly classify 86.4% of the table olive samples according to their trade category based on a sensory panel evaluation and following the International Olive Council regulations (i.e., extra, 1st choice, 2nd choice, and olives that may not be sold as table olives). So, the satisfactory overall predictions achieved demonstrate that the electronic tongue could be a complementary tool for assessing table olive defects, reducing the effort of trained panelists and minimizing the risk of subjective evaluations.

A taste sensor device for unmasking admixing of rancid or winey-vinegary olive oil to extra virgin olive oil

“…Harzalli et al Computers and Electronics in Agriculture 144 (2018) 222-231 2018), confirming the competence of this type of taste sensor device to give qualitative or quantitative responses towards positive or negative sensory attributes usually found in olive oil Slim et al, 2017) or in table olives (Marx et al, 2017a(Marx et al, , 2017c. Thus, in the present work, E-tongue-LDA-SA models were established based on selected sub-sets of sensors for discriminating non-adulterated EVOO or intentionally-adulterated EVOO with different levels of LOO-R or LOO-WV (2.5%, 5%, 10%, 20% or 40%, in v/v).…”

“…Harzalli et al Computers and Electronics in Agriculture 144 (2018) 222-231 the potentiometric E-tongue capability to detect adulterated EVOO due to the intentionally addition of different percentages of LOO-R or LOO-WV, was evaluated using linear discriminant analysis (LDA) coupled with the meta-heuristic simulated annealing (SA) variable selection algorithm (Bertsimas and Tsitsiklis, 1992;Kirkpatrick et al, 1983;Cadima et al, 2004). The electrochemical-chemometric strategy followed was similar to previous works of the research team Rodrigues et al, 2016;Veloso et al, 2016Veloso et al, , 2018Marx et al, 2017aMarx et al, , 2017bMarx et al, , 2017cSlim et al, 2017;Souayah et al, 2017). The signal profiles generated during the electrochemical analysis of the hydro-ethanolic extracts of non-adulterated and intentionally-adulterated olive oils were subjected to a linear discriminant analysis (LDA) in combination with a meta-heuristic simulated annealing (SA) algorithm, in order to establish predictive E-tongue-LDA-SA models capable of classifying olive oil according to their adulteration level.…”

Perception of olive oils sensory defects using a potentiometric taste device

“…The capability of a potentiometric E-tongue with cross-sensitivity and non specific lipid polymeric membranes to correctly classify olive oils according to the sensory intensity perception levels (i.e., intense, medium and light) of positive attributes (fruity, bitter and pungency) has been shown by our research team [13]. A similar electrochemical device was also used to classify table olives according to the sensory quality category based on the intensity of the defect predominantly perceived (DPP), to differentiate organoleptic negative attributes that may be perceived in table olives, using standard solutions and real samples as well as to quantify the intensity of the DPP in table olive and respective brine solutions as well as to evaluate table olives' gustatory attributes (e.g., acid, bitter and salty sensations) [22][23][24]. Furthermore, this type of E-tongue device showed quantitative responses towards polar compounds (aldehydes, esters and alcohols) usually found in olive oils and that are related to their sensory positive attributes (e.g., green and fruity) [25].…”

Discrimination of Olive Oil by Cultivar, Geographical Origin and Quality Using Potentiometric Electronic Tongue Fingerprints

“…The potentiometric E-tongue performance for olive oil classification according to the autochthonous Tunisian olive cultivar (i.e., cvs Chemlali or Sahli) and for each singlecultivar olive oil by its geographical origin (cv Chemlali: Kairouan, Sidi Bouzid and Sfax regions; and, cv Sahli: Kairouan, Mahdia or Sousse regions) as well as its quality grade (EVOO, VOO or LOO) was evaluated using LDA coupled with the meta-heuristic simulated annealing (SA) variable selection algorithm [34][35][36], following a previous described electrochemical-chemometric strategy [27][28][29][30][31][32]37]. The sensitivities (correct classification rates) of the selected LDA-SA models were evaluated using the leave-one-out crossvalidation (LOO-CV) and the repeated K-fold (repeated K-fold-CV) techniques.…”

“…Olive oil (≈ 10 g) was extracted using 100 mL of a water-ethanol solution (80:20, v/v), which enabled the extraction of polar compounds related to different sensory attributes of olive oil, and to which the lipid polymeric sensors show qualitative and quantitative responses, due to electrostatic or hydrophobic interactions [22,[24][25][26][27][28][29][30][31]. Ethanol was of analytical grade (Panreac, Barcelona) and water was deionized type II.…”