A COMPARISON OF THREE INSTRUMENTAL TESTS FOR PREDICTING SENSORY TEXTURE PROFILES OF CHEESE

Multivariate analysis techniques were used to seek correlations between texture sensory attributes assessed by a trained professional panel and instrumental measurements (compression, puncture and penetration) carried out on various types of cheeses. Twenty-nine cheeses were assessed by the panel and instruments. Correlation was sought using Partial Least Squares regression. Hardness (R=0.87), springiness (R=0.98) and cohesiveness of mass (R=0.89) were best predicted by instrumental data from a cone penetration test. The prediction of cohesiveness was acceptable using any of the three instrumental tests performed (0.76相似文献   

PREDICTION OF RICE TEXTURE BY SPECTRAL STRESS STRAIN ANALYSIS: A NOVEL TECHNIQUE FOR TREATING INSTRUMENTAL EXTRUSION DATA USED FOR PREDICTING SENSORY TEXTURE PROFILES

Sensory texture characteristics of cooked rice for three cultivars (74 samples) were predicted using an extrusion cell and a novel data analysis method (i.e. Spectral Stress Strain Analysis). Eight sensory texture characteristics were evaluated and force values from the instrumental tests were used in combination with Partial Least Squares regression to evaluate predictive models for each of the sensory attributes studied. Relative Ability of Prediction (RAP) values were evaluated for each model; they ranged from 0.06 to 0.85. Satisfactory models are proposed for the two major texture characteristics of cooked rice, namely hardness (RAP=0.85) and stickiness as evaluated by adhesion to lips (RAP=0.76). Other sensory attributes such as roughness of mass (RAP=0.73) and toothpack (RAP=0.81) were also satisfactorily predicted. Sensory attributes such as toothpull (RAP=0.12) and loose particles (RAP=0.06) could not be predicted using the Spectral Stress Strain Analysis.     

PREDICTION OF SENSORY TEXTURE OF FETA CHEESE MADE FROM ULTRAFILTERED MILK BY UNIAXIAL COMPRESSION AND SHEAR TESTING

The ability to predict sensory texture properties of Feta cheese made from ultrafiltered milk (UF-Feta) from uniaxial compression, small shear deformation measurements (frequency sweep, strain sweep, relaxation) and indices of proteolysis was studied. In principal component analysis (PCA) some of the instrumental variables were highly correlated, e.g. the moduli from uniaxial compression and shear measurements; and strain at fracture from uniaxial compression and indices of proteolysis. PCA of the six sensory attributes determined by a trained panel showed that mainly one type of information was present in the sensory results. Partial Least Squares regression (PLS) of all results revealed that stress at fracture from uniaxial compression was the individual instrumental parameter having the highest correlation with the sensory texture attributes. Of these, the three firmness attributes were best predicted by the instrumental parameters. As the shear measurements were not very useful for prediction of sensory texture properties by themselves, and as the increase in prediction precision by inclusion of these measurements was marginal, it is suggested that either stress at fracture alone, or together with three other parameters from uniaxial compression should be used to describe texture properties of UF-Feta cheese.     

INFLUENCE OF UNIAXIAL COMPRESSION RATE ON RHEOLOGICAL PARAMETERS AND SENSORY TEXTURE PREDICTION OF COOKED POTATOES

The effect of uniaxial compression rate (20–1000 mm/min) on the parameters: Stress (σ_ftrue), strain (ε_fHencky) and work to fracture (W_f), modulus of deformability (E_d), maximum slope before fracture (E_max) and work during 75% compression (W_total) was investigated for ten potato varieties. Multivariate data analysis was used to study the correlation between and within the sensory and nonsensory measurements by Principal Component Analysis (PCA) which showed σ_ftrue, E_max, W_f, and W_total to explain the same type of information in the data, and ε_fHencky versus E_d another type of information in the data. The deformation rate had a large effect on ε_fHencky. Nine sensory texture attributes covering the mechanical, geometrical and moistness attributes were evaluated. Relationships between uniaxial compression data at various deformation rates and the sensory texture attributes were studied by Partial Least Squares Regression (PLSR). A minor effect of deformation rate on the correlation with the sensory texture properties was obtained. Mechanical properties were predicted to a higher extent than the geometrical attributes and moistness. The prediction of the mechanical, geometrical and moistness attributes increased largely by using uniaxial compression supplemented by chemical measures such as dry matter and pectin methylesterase, but here no relevant effect of deformation rate was obtained.     

INSTRUMENTAL AND SENSORY CHARACTERIZATION OF COOKED POTATO TEXTURE

Uniaxial compression, Texture Profile Analysis (TPA) and chemical measurements were related to sensory texture evaluation of potato quality during storage. Principal component analysis grouped the varieties into three types of variation: mealiness versus firmness and springiness (PC1), moistness versus adhesiveness (PC2) and hardness versus adhesiveness and moistness (PC3). In uniaxial compression the variable 'stress', 'work up to fracture' and 'total work during compression' described the same type of information in the data. These uniaxial data and most of the TPA data were highly correlated. Uniaxial compression data (stress, strain, modulus of deformability), starch structural data (area, roundness, aspect ratio), specific gravity and pectin methyl esterase activity discriminated between the varieties and harvest times. Partial Least Squares Regression showed stress, strain, modulus of deformability and specific gravity to be the most important variables in distinguishing between two groups of sensory texture attributes explaining 65% of the total variance in the sensory data. Coefficients of correlation between predicted and measured sensory attributes were in the range 0.36–0.79. The TPA data were not found to be relevant substitutions for the sensory attributes.     

UNIAXIAL COMPRESSION OF UF-FETA CHEESE RELATED TO SENSORY TEXTURE ANALYSIS

Rheological characteristics of seven Feta cheeses with different textures and produced from ultrafiltered milk (UF-Feta cheeses) were evaluated by uniaxial compression and sensory texture analysis. The effect of uniaxial deformation rate (50–2500 mm/min) on four rheological parameters: Stress at fracture s?_f), Hencky strain at fracture (?_f), deformability modulus (E) and work to fracture (W_f) was examined. Three Principal Components (PC) described 76, 16 and 4% respectively, of the variation in the uniaxial compression data set (4 parameters at 12 deformation rates). Statistically α_f, E and W_f described the same type of information in the data set. Six sensory texture attributes of the UF-Feta cheeses were evaluated by a sensory texture panel: nonoral firmness, nonoral brittleness, nonoral spreadability, oral crumbliness, oral firmness and oral stickiness. One PC described 93% of the variation in the sensory texture data and grouped the sensory variables into two negatively correlated groups: nonoral firmness nonoral brittleness, oral firmness and oral crumbliness versus nonoral spreadability and oral stickiness. Correlations and Partial Least Squares regression (PLS) between instrumental and sensory texture variables showed that nonoral and oral firmness were the nonoral and oral sensory variables best predicted from instrumental measurements. α_f, E and W_f were all able to predict nonoral and oral firmness. Of the instrumental parameters, α_f generally gave the best correlation to nonoral firmness at all deformation rates. Above a deformation rate of 50 mm/min correlations between α_f and nonoral firmness were almost independent of deformation rate, and at any deformation rate correlations between α_f and oral firmness     

牛肉酶解物对牛肉特征香味形成的影响研究

以五种不同水解度牛肉酶解物分别制备得到5种热反应牛肉香精。利用定量感官分析和GCMS-O分析考察了5种热反应牛肉香精香气成分的变化。结果显示,添加DH29.13%酶解液的牛肉香精样品其牛肉味、肉香味和仿真度相比于其它样品是最强的。GC-O分析也发现DH29.13%的牛肉酶解液赋予形成了种类较多的香气活性化合物,而没有添加牛肉酶解物的肉味香精缺失了这些特征化合物。利用PLSR进行了香气活性化合物、感官评价和牛肉酶解液的肽分子量分布之间的相关性分析,进一步解释了DH为29.13%的牛肉酶解液是赋予热加工牛肉风味的最合适风味前体。     

Eating quality of raw carrots––correlations between flavour compounds,sensory profiling analysis and consumer liking test

Sensory profiling, analysis of aroma, sugar and dry matter, as well as consumer liking test were used to characterise the quality of six carrot cultivars, grown at two locations in Denmark. The carrot samples were examined at harvest, and after three months of cold storage. Carrot cultivar had an effect on most sensory and flavour compound variables, due particularly to the influence of one cultivar; and also location influenced carrot quality. Storage of carrots was characterised by an increase of a range of aroma components, but the changes in flavour compounds were not correspondingly observed by the sensory analysis. By Partial Least Squares Regression (PLSR) two thirds of the flavour compound variables was found to correlate significantly with one or more of the nine sensory attributes; and all of the sensory attributes were significantly correlated with one or more of the consumer liking test variables bitterness, sweetness and liking.     

Sensory evaluation of fermented milks: vocabulary development and the relations between sensory properties and composition and between acceptability and sensory properties

Methodology for sensory profiling of fermented milks has been studied. A vocabulary of sensory attributes has been developed which encompasses most of the difference between samples. Principal Component Analysis, incorporating factor rotation, allowed simplification of the variability between samples to five Principal Components capable of clear interpretation. Sensory mapping was found to be a useful tool for categorizing fermented milks. The acceptability of the fermented milks was successfully modelled, by Partial Least Squares Regression, in terms of a limited number of key attributes. The model explained 88.4% of the variance. The relations between sensory attributes and the composition of the fermented milks were considered using Multiple Linear Regression. Although a number of statistically significant relations were derived they were of poor to modest value for purposes of prediction .     

Vane Rheometry for Textural Characterization of Cheddar Cheeses: Correlation with Other Instrumental and Sensory Measurements

The relationship between instrumental (vane method, texture profile analysis (TPA), uniaxial compression) and sensory texture measurements of Cheddar cheeses was investigated. A Haake VT 550 viscotester equipped with a four-bladed vane rotor was used for the vane test. Instrumental TPA was performed with a TA.XT2 Texture Analyser, and compression variables were calculated from TPA data. Vane parameters were significantly correlated with respective variables of compression and TPA (r=0.56-0.91), and sensory tests (r=0.54-0.88). Multivariate analysis indicated that seven sensory attributes of ten commercial Cheddar cheeses were satisfactorily predicted (calibration regression coefficient,R_cal >0.62) by variables of the vane, uniaxial compression and TPA tests. In particular, cheese firmness and cohesiveness evaluated by sensory panel were well described by vane stress and apparent strain. The results validate the vane method as an alternative to the existing cheese testing methods for rapid evaluation of cheese texture.     

Comparison between linear and nonlinear PLS methods to explain overall liking from sensory characteristics

The influence of sensory characteristics on overall liking can be statistically studied with Partial Least Squares (PLS) regression methods. To correctly model nonlinear dependence relationships, some nonlinear PLS extensions are useful. The purpose of the present paper is to compare performances and results of three PLS methods, using a real data set: regular PLS with sensory attributes as explanatory variables; PLS with attributes and their respective squares; and a new nonlinear PLS extension, called ASPLS. In case of a nonlinear dependence relationship between sensory characteristics and hedonic responses, this last method is shown to be worth considering.     

Relationships between sensory attributes and the volatile compounds,non-volatile and gross compositional constituents of six blue-type cheeses

Relationships between odour and flavour attributes of six blue-type cheeses and their volatile compounds, free amino acids (FAA), free fatty acids (FFA) and gross compositional constituents were determined. Relationships were also determined between texture attributes and gross compositional constituents. Fifteen assessors described the odour, flavour, appearance and texture profile of cheeses. Volatile compounds were isolated using a model-mouth apparatus. FAA, FFA and gross compositional constituents were determined using standard methods. Using Partial Least Squares Regression two odour and five flavour attributes were found to correlate with subsets of volatile compounds, FAA, FFA and gross compositional constituents. For example, “mouldy” flavour was positively correlated with the concentrations of pH 4.6-soluble nitrogen and 2-pentanone, 2-heptanone, 2-octanone and 2-nonanone. Three texture attributes were found to correlate with subsets of gross compositional constituents. For example, “crumbly” texture was positively correlated with concentration of fat and protein and negatively correlated with levels of moisture in the non-fat substance and moisture.     

Rapid determination of pork sensory quality using Raman spectroscopy

Existing objective methods to predict sensory attributes of pork in general do not yield satisfactory correlation to panel evaluations, and their applications in meat industry are limited. In this study, a Raman spectroscopic method was developed to evaluate and predict tenderness, juiciness and chewiness of fresh, uncooked pork loins from 169 pigs. Partial Least Square Regression models were developed based on Raman spectroscopic characteristics of the pork loins to predict the values of the sensory attributes. Furthermore, binary barcodes were created based on spectroscopic characteristics of the pork loins, and subjected to multivariate statistical discriminant analysis (i.e., Support Vector Machine) to differentiate and classify pork loins into quality grades ("good" and "bad" in terms of tenderness and chewiness). Good agreement (>83% correct predictions) with sensory panel results was obtained. The method developed in this report has the potential to become a rapid objective assay for tenderness and chewiness of pork products that may find practical applications in pork industry.