2.2. Samples
Representation of final chocolate shapes and their names are shown in Table 1. The shapes named Catstongue 1 and Catstongue 2 had very similar architectures; the former being longer and thinner than the latter one.
To limit the study to the impact of the shape on in-mouth perception, samples were made from the same recipe of dark
chocolate at 70% cocoa content and by using the same batch of chocolate mass.
2.3. Sensory evaluation Sensory tests were conducted with 12 trained female subjects with an average age of 45 years old. They were screened for their sensory abilities and had previous experience in descriptive evaluation
of foods. Training was provided so they were able to describe and discriminate samples in a reliable way.
The two methodologies employed to characterise the products and underline the sensory differences were:
Monadic profiling. This method gives a description of the product’s in-mouth properties and allows quantifying the
intensity of each attribute. This static method was used as a first step to identify the attributes that discriminate samples
and to select shapes that differed the most. TimeeIntensity (TI). This dynamic methodology allows following the evolution of the intensity of a single sensory attribute over time. In the present study, TI was applied to
cocoa flavour to characterise, over time, the shapes that differed the most according to the monadic profiling.
Subjects were instructed to let the chocolate melt in the mouth and to avoid biting into it. Movement of tongue was not restricted and subjects had the possibility to move the chocolate piece in the mouth. The sensory glossary encompassing attributes, their defi-nitions and their protocol of evaluation was agreed among subjects
(Table 2). For each of the samples, attributes were scored on unstructured scales anchored from “weak” to “strong” (0e10). In order to avoid carry-over effects and to ensure independent evaluations
of samples, subjects were requested to wait for 10 min between samples and to clean their mouth by eating apple slices
and rinsing with bottled water.In the timeeintensity experiment, subjects were instructed to
start the evaluation when putting the sample in the mouth by clicking at the left anchor of the unstructured scale (“weak”). Then,they scored continuously the cocoa flavour intensity over time by moving horizontally the anchor on the intensity scale. The total duration of the evaluation was set at 10 min, meaning that subjects had to score the cocoa intensity even after the product had completely melted and had been swallowed. Otherwise, test conditions were unchanged from sensory profiling test. For both evaluations, samples were delivered individually on white plastic trays labelled with three digit codes. Samples were stored at 20 C and evaluated at room temperature.
Sensory tests were carried out in individual sensory booths
designed according to the ISO 8589 standards (1988). A red light
was set up so that small differences in gloss due to the curvature of
the shapes were minimised.
In each test, samples were assessed twice and evaluated on
screen. The order of presentation of samples was balanced
according to a presentation plan based on Latin Squares.
Data acquisition of monadic profiling and timeeintensity tests
was done with Fizz software (Biosystèmes, Couternon, France). For
the timeeintensity test, the acquisition frequency was set at 2.6 s1
.