Do trained assessors generalize their knowledge to new stimuli?

Previous work showed that trained assessors are better at discriminating and describing familiar chemico-sensorial stimuli than novices. In this study, we evaluated whether this superiority holds true for new stimuli. We first trained a group of subjects to characterize beer flavors over a two year period. After training was accomplished, we compared the performance of these trained assessors with the performance of novice subjects for discrimination and matching tasks. The tasks were performed using both well-learned and new beers. Trained assessors outperformed novices in the discrimination task for learned beers but not for new beers. But on the matching task, trained assessors outperformed novices for both learned and new beers. These results suggest that perceptual learning is harder to generalize than verbal learning.     

Origins of Flavour in Whiskies and a Revised Flavour Wheel: a Review

The nature and origins of flavour in whiskies are reviewed with the aim of developing a revised and simplified flavour wheel for training of sensory assessors. Scotch whiskies are perceived as having distinctive characters, generally recognised in pattern recognition (perception, macroscopic brain processing), rather than being subjected to a deconstruction process of evaluating attributes (sensation, microscopic brain processing). Although consumers use simple recognition judgements on whisky flavour in categorical assimilation, industry has a requirement for monitoring spirit quality that necessitates a more reductionist approach. Whisky flavour wheels identify attributes, specific components of flavour character, which can be demonstrated to sensory assessors using reference standards. The advent of cyclodextrin bound reference standards has enabled communication of information on flavour character in training of assessors, as exploited in the brewing industry. A revised flavour wheel, with characters illustrated by reference standards, is proposed to assist sensory training on attributes of whisky flavour character.     

Sensory profile of a Fernet appetizer

The odour and flavour profiles of Fernet appetizer were determined by a panel of 51 male and 65 female assessors. The performance of the two sexes was very similar but there were still slight differences in ratings of intensities of various partial flavours. No differences were observed in the accuracy. In some cases odour notes were rated differently from the analogous flavour notes. Correlations between intensities assigned to various flavour and odour notes were found sensitive indicators of differences between male and female assessors. Significances of various correlation coefficients often widely differed between the set of men and that of women. The results proved moderate differences of odour and flavour perception, respectively, by men on one hand and women on the other hand.     

Prediction of sensory characteristics of lamb meat samples by near infrared reflectance spectroscopy

This study was implemented to evaluate the potential of visible and near infrared reflectance (NIR) spectroscopy to predict sensory characteristics related to the eating quality of lamb meat samples. A total of 232 muscle samples from Texel and Scottish Blackface lambs was analyzed by chemical procedures and scored by assessors in a taste panel (TP). Then, these parameters were predicted from Vis/NIR spectra. The prediction equations showed that the absorbance data could explain a significant but relatively low proportion of the variability (R(2)<0.40) in the taste panel traits (texture, juiciness, flavour, abnormal flavour and overall liking) of the lamb meat samples. However, a top-tail approach, looking at the spectra of the 25 best and worst samples as judged by TP assessors, provided more meaningful results. This approach suggests that the assessors and the spectrophotometer were able to discriminate between the most extreme samples. This may have practical implications for sorting meat into a high quality class, which could be branded, into a low quality class sold for a lower price for less demanding food use. Regarding the chemical parameters, both intramuscular fat and water could be more accurately predicted by Vis/NIR spectra (R(2)=0.841 and 0.674, respectively) than sensory characteristics. In addition, the results obtained in the present study suggest that the more important regions of the spectra to estimate the sensory characteristics are related to the absorbance of these two chemical components in meat samples.     

基于秩相关与聚类分析的评价员排序能力评估

将最常用的甜、酸基本味分别设置成6个浓度体系作为评价员排序能力筛选的参比样,考虑到感官疲劳的客观现象,12位评价员对每个味道系列浓度进行3次重复排序,在分别采用Spearman检验和Kendall检验评估评价员排序正确性与重复性的基础上,采用聚类分析融合正确性与重复性数据特征,综合评估分析每位评价员的排序能力。发现5号评价员表现最差需要剔除;3号评价员需要进一步训练;6位评价员(编号为2、4、9、10、11、12号)排序能力表现优异,可直接开展检测试验;4位评价员(编号为1、6、7、8号)排序能力良好。同时发现Kendall检验系数W值与Spearman检验系数rs平均值具有整体正相关趋势,而W值能更精确区分评价员之间的能力差异。该研究可为感官分析实验室中评价员的筛选与维护提供理论依据和实践经验。     

Relationships of Overall Estery Aroma Character in Lagers with Volatile Headspace Congener Concentrations

In lager beers the intensity of “estery” aroma character is regarded as an important component of sensory quality, but its origins are somewhat uncertain. Overall “estery” aroma intensity was predicted from capillary gas chromatographic (GC) data following solid phase micro extraction (SPME) of headspaces. Estery character was scored in 23 commercial lagers using rankrating, allowing assessors (13) constant access to a range of appropriate standards. From univariate data analysis, all assessors behaved similarly and lagers fell into three significantly different groups: low (1), high (1) and intermediate (21). The quantification of 36 flavour volatiles by SPME of headspaces was reproducible and principal component analysis explained 91% total variance. Multiple linear regression could utilise only a restricted (26) set of flavour volatiles, whereas partial least square regression, that considered all flavour components, showed significant differences and improved prediction. However, an artificial neural network that could compensate for non‐linearities and interactions in ester perception gave the most robust prediction at R² = 0.88.     

Sensory quality in retailed organic,free range and corn-fed chicken breast

Discriminations between organic, free-range, corn-fed, and conventional retailed chicken breasts (20) were explored using free choice and conventional profiling. Generalised Procrustes analyses of total free choice data explained 45.3% variance in two significant factors with one group of assessors perceiving differences in flavour between organic and conventional products whereas others did not. Clear grouping based on category was apparent from pair-wise Procrustes matching of matrices of appearance and texture data. Principal component analysis of conventional profiling data explained 92% variance in three significant components with appearance and texture discrimination, with clustering based on product category, dominating the product space but there was no significant differentiation in aroma and flavour. Discriminant partial least square regression provided a relationship between appearance and product category. It was concluded that the primary differentiation between chicken breast from differing production regime was on the basis of appearance and texture, but that a sub-group of assessors could also differentiate on the basis of aroma and flavour.     

Sensory evaluation of young goat meat

Loins from 12 young goats (three extensively reared-entire males, EE; three extensively reared-castrated males, EC; three intensively reared-entire males, IE; four intensively reared-castrated males, IC) were assessed by a trained panel of eight assessors for odour, flavour and texture using 100 mm unstructured scales. Procrustes analysis was used to produce the spatial configuration of the samples. The first principal axis explained 44.9% of the variation and was mainly based on textural attributes. The second principal axis accounted for 21.3% of the variation and was mainly related to odour and flavour. Samples obtained from groups EE and EC were more tender and juicy, whereas, IE animals produced a more cohesive meat. Groups IE and IC produced a more stringy meat, with a higher meaty odour and flavour. The rearing system affected texture more than odour and flavour, whereas sex had little effect on textural attributes compared with odour and flavour.     

Influence of colour intensity on the perception of colour and sweetness in various fruit-flavoured yoghurts

A study was made of the influence of colour concentration on the perception of flavour, sweetness and colour intensity, and also of opinions about colour suitability, in yoghurts with strawberry, lemon, fruit of the forest and orange flavours. The results indicated that, even with the same content of each fruit flavour and sugar, the greater the concentration of colourant, the greater was the intensity of taste perceived by the assessors in yoghurts with strawberry, orange and fruit of the forest flavours. With regard to perception of sweetness, only in the yoghurts with fruit of the forest flavour was it found that the greater the concentration of colourant, the greater was the sensation of sweetness.In the range of colourant concentrations studied for the four flavours, it was found that the colour considered most suitable corresponded to the yoghurts with intermediate colourant concentrations and that, for the four flavours studied, the greater the concentration of colourant, the greater was the sensation of intensity of colour.     

Measurement of Thresholds for Reference Compounds for Sensory Profiling of Scotch Whisky†

Sensory profiling of Scotch whiskies, important for research and product analysis, requires training of assessors including the identification of important whisky aroma characters. Seventeen compounds were selected as potential reference standards for common whisky sensory characteristics. Nineteen assessors determined detection and recognition thresholds for each reference standard dissolved in 3 year old grain whisky at 23% (v/v) ethanol using both ascending dilution series and Latin square presentations. With the former, order effects and errors of habituation and anticipation were apparent. However, the Latin square presentation was satisfactory for threshold measurements in whisky, yielding good correlations and reproducibility, and enabling outliers to be reassessed independently. Slopes obtained from plotting probit against concentration for detection and recognition data, psychometric functions of aroma perception, varied between reference compounds and 90% recognition concentrations were obtained for each aroma standard. Sixteen of the compounds assessed were suitable for use as reference flavour standards for training of whisky assessors.     

Relationships between the gross,non-volatile and volatile compositions and the sensory attributes of eight hard-type cheeses

Relationships between individual flavour attributes of eight hard-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 individual texture attributes and gross compositional constituents. A trained panel of 15 assessors described the sensory characteristics of the cheeses using eight odour, twenty flavour, four appearance and nine texture attributes. Volatile compounds were isolated using a model-mouth device. FAA, FFA and gross compositional constituents were determined using standard methods. Relationships were determined by using partial least squares regression coupled with a new jack-knife method for identification and elimination of non-contributing variables. Eight flavour attributes were found to be correlated with subsets of volatile compounds, FAA, FFA and gross compositional constituents. For instance, the “nutty” flavour of Emmental was found to be positively correlated with the concentrations of propionic acid, ethyl acetate and 2-pentanone. “Nutty” flavour was negatively correlated with the concentrations of salt in moisture and pH 4.6-soluble nitrogen (pH 4.6-SN). Four texture attributes were correlated with subsets of gross compositional constituents. For example, “firmness” was positively correlated with concentrations of protein, calcium and phosphorous and negatively correlated with pH value and level of pH 4.6-SN.     

Swiss-type and Swiss–Cheddar hybrid-type cheeses: effects of manufacture on sensory character and relationships between the sensory attributes and volatile compounds and gross compositional constituents

The effects of manufacturing steps, including rate of acidification in the vat, hot room step, curd wash and ripening temperature, on the sensory character of Swiss-type and Swiss–Cheddar hybrid-type cheeses were determined. In addition, relationships were determined between sensory attributes of cheeses and their gross compositional constituents and volatile compounds. Ten assessors described the sensory characteristics using 12 odour, 19 flavour, four appearance and nine texture attributes. Gross compositional constituents were determined using standard methods. Volatile compounds were isolated using a model mouth apparatus that included a mastication device. Statistical methods were used to reduce the original sensory vocabulary to four odour, eight flavour, three appearance and five texture attributes. Results of a principal component analysis on the sensory data showed clear differences in sensory character between cheese types. For example, the Swiss-type with a low rate of acid production in the vat and a hot room step had a more 'nutty' and 'sweet' flavour than the modified Swiss-type with higher acid production and no hot room step. However, the modified Swiss–Cheddar hybrid with a curd wash step, a higher acid production and no hot room step also had a 'nutty' and 'sweet' flavour. Ripening temperature also has an effect on sensory character. Relationships between sensory character and volatile compounds and/or gross compositional constituents were determined by using partial least squares regression. One odour and one flavour attribute were shown to be correlated with subsets of volatile compounds and gross compositional constituents. Four texture attributes were correlated with subsets of gross compositional constituents. The availability of information on the manufacturing process increased the interpretability of the models.