2.3. Sensory methods

A trained sensory panel of experts in profiling techniques (n = 11; 10 female, 1 male), with a minimum of 6 months training, were given further training on WPC profiling and sequential profiling (minimum 5 h). Sensory evaluation was carried out at room temperature (25 ± 2 °C) in isolated booths.

Quantitative descriptive analysis (QDA) (Stone, Sidel, Oliver, Woolsey, & Singleton, 1974) was performed using a consensus vocabulary developed by the panel during training (34 attributes; 3 appearance, 6 odour, 6 taste, 6 flavour, 6 mouthfeel, 6 aftereffects, Table A1). The panel assigned mouthfeel characteristics in order to separate important attributes describing distinct sensations. These consensus mouthfeel attributes were: body; furring, the roughening of the tongue; chalky, to describe the sensation of particulate matter; mouthcoating; astringency, specific to the puckering of the cheeks; and drying, to describe the sensation of the reduction of saliva in the mouth.

WPC samples were evaluated in duplicate according to a balanced design using unstructured line scales with appropriate anchors. Samples were presented monadically in opaque white cups (20 mL), unsalted crackers and warm filtered tap water were provided as palate cleansers between samples during an enforced break (2 min). Evaluation was carried out under artificial daylight.

Sequential profiling was carried out to establish the perception of seven sensory attributes over repeated consumption of eight aliquots (5 mL) of samples, with 1-min breaks between aliquots. Samples were scored after consumption of each aliquot (T0), and following 30 (T30) and 60 s (T60) time delays, as described by Methven et al. (2010) (Compusense at-hand, Ontario, Canada). Thus there are eight aliquots tasted for each of four samples (WPC00, WPC05, WPC10 and WPC20), scored at three time points (T0, T30 and T60).

The seven attributes scored were bitter, sour, metallic, cooked milk flavour, mouthcoating, chalky and drying. The maximum number of attributes that we recommend to score within one sequential profiling session is 7, determined through training with the panel. These were chosen carefully from the full QDA profile. Bitter, sour and metallic are taste attributes associated with whey protein beverages (Martini and Walsh, 2012, Whetstine et al., 2005). Cooked milk flavour was selected as this attribute showed significant differences between samples in the QDA data as both an odour and flavour attribute. Mouthcoating, chalky and drying were selected by the panel as dominant mouthfeel attributes, and the QDA data showed increases upon heating for all three attributes.

Samples were coded with three-digit numbers and all eight aliquots of one sample were presented together with the same code; the panellists were not blinded to the sequential nature of the evaluation. Warm filtered tap water and unsalted crackers were provided as palate cleansers in the 2 min enforced break between samples; however panellists were instructed not to use these between the eight aliquots of the same sample. Panellists were instructed to consume the total volume of each aliquot and to coat the mouth with the sample before swallowing. Two samples were scored in each session. Evaluation was carried out under red lighting and aliquots were served in opaque black cups to mask appearance differences between samples. Nine of the trained panellists were present for sequential profiling.