The overall acceptability of food is largely based on sensory factors such as appearance, flavor, and texture, which are perceived by the senses directly (Bourne 2002). To advance the area of dynamic sensory perception, we need to consider the current dynamic sensory techniques available, the interplay between oral physiology and sensory perception, and the relationships between physical characteristics and sensory perception. These are discussed further below.
The selection of appropriate dynamic sensory methods
Examining flavor and textural changes during a masticatory sequence requires sensory techniques that can assess perception considering the additional dimension of time. The main techniques that have been applied for this purpose can be categorized into 2 quite different approaches: temporal dominance of sensation (TDS) and TI measurements (such as discrete point TI, continuous TI). TI has been around for some 40 y (the evolution of time-intensity methodology for sensory evaluation was reviewed comprehensively by Cliff and Heymann (1993), whereas TDS is a relatively new method developed by “Centre European des Sciences du Guot” in the year 1999 (Pineau and others 2009). A good description of these techniques can be found in Kemp and others (2009).
Lenfant and others (2009) used TDS to evaluate the dynamics of texture perceptions during oral processing of different breakfast cereals. TDS allows the subject to record their dominant sensation (in this case hardness, crackliness, crispness, brittleness, lightness, stickiness, grittiness, or dryness) for each product at different points throughout a masticatory sequence. Twenty-five untrained subjects were used to create a vocabulary for the breakfast cereals of interest, which were later reduced to the 8 most frequently used terms. A glossary for these terms was created and the subjects practiced the use of these terms using different commercially available (nonbreakfast cereal) products. After training on the use of the TDS method and the protocol used, 3 g samples were assessed and the dominant attribute, which changed with time, was selected on a computer screen throughout the chewing process. Due to the masticatory process being highly variable among subjects, each TDS curve was standardized by reducing the x-axis was reduced from x= 0 (1st scoring) to x= 1 (swallowing). The aim here was to standardize the data according to individual mastication time; however, differences between the time taken from ingestion to 1st scoring were unaccounted for. Overall, hardness and crackliness attributes were dominant in the early stages of mastication (and registered ≤5 s after ingestion), which then gave way to the perception of crispness. Brittleness was dominant in the middle of a masticatory sequence and stickiness was highly dominant at the end. Stickiness referred to sticking to the palate and teeth during oral processing, but it is possible that subjects also assessed this as cohesiveness, which is thought to be critical for the initiation of swallowing (Prinz and Lucas 1995; Lillford 2001; Woda and others 2006).
To date, TI method is defined as one of the most frequently used descriptive sensory analyses to determine intensity of a specific attribute(s) over a period of time (Cliff and Heymann 1993; Peyvieux and Dijksterhuis 2001; Sprunt and others 2002; McGowan and Lee 2006; Ross 2009). However, TI has several limitations as discussed by Ledauphin and others (2006). In particular, it is very time consuming when more than one attribute is of interest. Dual-attribute TI was developed with the intention to reduce the time used for sensory evaluation by requiring panelists to score simultaneously 2 attributes by moving a cursor on a x–y axis (each axis refers to one attribute) (Pineau and others 2009). Nevertheless, this method cannot be extended to more than 2 attributes at one time. Continuous TI has also been reported as producing inconsistent data with Saint-Eve and others (2006) selecting 3 discrete points for intensity ratings of aroma in flavored yoghurt complexes rather than using continuous TI. Therefore, when multiple attributes need to be assessed during the same run, the newer TDS method warrants further investigation.
Despite this noteworthy introduction of a new dynamic process for assessing sensory evaluation, there is currently little data comparing these 2 approaches formally, particularly in solid foods. This possibly is due to the newness of the technique. However, there is a need for us to understand more fully how subjects interact with these techniques. The decision processes that subjects are using for TDS where dominant attributes are being selected can be viewed as a series of ranking assessments over a period of time as opposed to TI where one or two attributes are being assessed for their intensity. Pineau and others (2009) compared TI and TDS for assessing 5 dairy products. Although both methodologies gave similar patterns of sensations, it was concluded that TDS demonstrated more clearly the sequence of the sensations over time. Labbe and others (2009) compared TDS with sensory profiling for gels containing different levels of odorants, citric acid, cooling agent, and xanthan gum. They found differences between dynamic perceptions and those gained immediately postconsumption, in particular for bitterness and coldness sensations, and concluded that TDS may be more relevant for understanding complex perceptions such as refreshing. TDS and TI methods are compared in Table 1.
Table 1–. Comparison between temporal dominance of sensations (TDSs) and time intensity (TI) methods.
An important consideration in comparing masticatory behavior with sensory perception is the influence of training on oral processing. It is well known that there are large variations among individuals for all parameters of mastication, even when potentially influencing variables (for example, age, gender, dental status) are controlled for (Lassauzay and others 2000). Training on sensory attributes could be another potential source of variation. Mioche and Martin (1998) found significant differences between trained and untrained panelists for EMG muscular work and insignificant differences for the number of chews and chewing time. González and others (2002) found trained subjects to use longer chewing times, smaller chewing frequencies, and larger muscular work than untrained subjects. Intraindividual variability was also greater for trained subjects, and this was thought to be related to these subjects having too much knowledge of the products and them receiving samples slightly different from what they were expecting. González and others (2002) found gender not to be an influencing variable. This is in contrast to other studies, which have found males to display higher EMG activities, higher vertical amplitudes, and slightly higher frequencies although no difference between the total number of cycles used compared with females (Peyron and others 2004a).
The decision of which dynamic sensory technique is most suited to investigate the effects of food properties and/or oral processing on sensory evaluation requires some consideration. Currently, our understanding of which techniques are most suitable is at a general level based on numbers of attributes that need to be assessed and the complexity of the food system. Until we can develop model food products where only 1 sensory attribute changes over time, TDS is a technique worthy of further investigation for this area of research. However, there is now a need for more studies to investigate this technique further, for example, looking at factors such as effects of training panelists, attribute selection, and how panelist's sensory data is standardized across the whole eating process. From this research a more standardized approach to using the techniques may be formulated.
Interplay between oral physiology and sensory perception
It is believed that many assessments are made during the first bite and that the 1st chewing cycle is often exploratory and is generally different from subsequent chewing cycles, for example, the frequency of a chewing cycle is usually slower for the 1st cycle (Foster and others 2006; Peyron and others 2002). Duizer and others (1996) compared instrumental TPA with sensory tenderness measured over an entire masticatory sequence using TI sensory evaluation and masseter activity measured by surface EMG for 5 different beef samples. It was found that maximum tenderness perception occurred anywhere between the 1st and 4th chew, illustrating that more than just the first bite is required to measure tenderness, as previously proposed by Boyar and Kilcast (1986). It was also found that age differences among the 5 meat samples relating to tenderness were best detected during late mastication.
Unfortunately, only a few studies have looked at sensory adaptations throughout a masticatory sequence, and very different foods were used. While Duizer and others (1996) focused on meat tenderness, de Wijk and others (2008a) investigated sensory attributes in vanilla custard desserts. Here, the authors compared VMG activities from the throat and temporalis with the assessments of thick, creamy, melting, fatty, rough, and liking attributes using 11 starch-based vanilla custard desserts (with different fat contents, viscosities, and polystyrene particle sizes) from 10 subjects. A 5 mL sample of each custard was tested orally for 5 s while 1 of the 6 attributes was assessed using a 10-cm line scale. Activities measured from the temporalis muscles were mostly affected by jaw movements whereas those measured from the throat were mostly affected by tongue movements. As seen by de Wijk and others (2003 and 2006b), increasing thickness ratings related to reduced up and down movements of the tongue. High melting ratings were associated with higher throat VMG activities and hence increased up and down tongue movements. Both VMG activities were related to creaminess ratings, indicating more complex movements are required to assess creaminess, which is believed to be a complex sensation relating to the food's viscosity, lubrication, and flavor (de Wijk and others 2006a). Although creaminess is typically correlated with liking, different oral behaviors were used to assess creaminess and liking. Both were associated with low VMG activities at the temporalis, whereas liking was associated with low throat VMG activities, and creaminess was associated with high throat VMG activities. The point during the 5 s of oral processing at which these attributes were related to oral movements differed: 1 to 2 s after ingestion for liking compared with 4 to 5 s after ingestion for creaminess. de Wijk and others (2008a) stated that liking (hedonic) and creaminess (analytical) sensations activate different pathways and motor programmes in the brain that determine oral movements.
Although, in theory it is possible to minimize the effects of aroma when studying the interplay between sensory perception and oral processing by allowing subjects to wear nose clips that inhibits retronasal aroma release, the taste modality (sweet, sour, bitter, salt, umami, and possibly fatty taste), trigeminal responses (for example, astringency), and effects of saliva cannot be minimized as easily. Further, a model system that minimizes these effects may be a step too far away from a real eating situation. Neyraud and others (2005) used viscoelastic gels with varying concentrations of quinine, 0 to 1446 μmol/kg, and measured the activity of the masticatory muscles and intensity of several sensory attributes (bitterness, sourness, sweetness, firmness, and acceptability) over time. With increasing quinine concentration, bitterness perception increased while acceptability and sweetness decreased, chew time decreased (from 30 to 22 s), and clearance time increased (from 7 to 14 s). Neyraud and others (2005) also found correlations between chewing and sensory perceptions. Subjects who chewed for longer reported higher bitterness and lower acceptability ratings. Subjects who used more muscle effort found the gels to be sweeter, and subjects who had higher levels of quinine in their saliva rated the gels as more bitter. This study indicates clearly that acceptability correlated with taste will decrease chew time regardless of the textural characteristics of the products. Saliva, which is known to be important during the breakdown of food in the mouth, also affects sensory perception in custard and mayonnaise. Engelen and others (2007) found that subjects with high total protein concentrations in their saliva reported low flavor ratings, low slippery lip-tooth feelings, and fatty after-feel. High α-amylase activity correlated with a reduced vanilla flavor sensation in custard and decreased creamy after-feel. This is thought to be due to the enzyme breaking down the starch in the custard to give a less viscous product, resulting is less surface area and reduced flavor release. Subjects with low α-amylase activity also had stronger slippery lip-tooth sensations. Pionnier and others (2004) used a modified TI methodology to study the relationships between perception, flavor release, and oral parameters during mastication of one type of cheese. Interindividual differences in aroma and taste compounds were related to interindividual differences in flavor release, masticatory measurements, and salivation.