Affective Imagery Modifies Sweetness Sensitivity

Subjects

Fifty-two student volunteers (37 female and 15 male) from Changshu Institute of Technology (CIT), China were chosen for the experiments. They were all self-reported right-handed and had normal eyesight or at least were corrected to normal by glasses. None of them was color blind and their ages were between 21 to 30 years old (average 24 ± 3 years). They did not have any taste- or smell-related disease before. All the participants were well briefed about the details of their performance. They all agreed and signed on the written informed consent declaration to volunteer as subjects in these experiments. The study was approved by the Institute Ethics Committee (IEC) of CIT, according to the Ethics Guidelines.

Sugar Solution

Sucrose was dissolved in distilled water to prepare the sugar solution with concentration of 1.5, 3.1, 3.9 and 4.7 g/L, respectively. All the solutions were prepared in volumetric flasks one night before and kept on the table at the room temperature between 20 to 25 °C. During the experiments the solutions were provided to the subjects in a series of half-filled odorless white paper cups (25 ml).

Imagination and Taste Training

The experiments were carried out at the sensory science laboratory, Changshu Institute of Technology, China. The participants were selected by the criterion that if they could detect sweetness and identify the difference of the sugar solutions (1.5, 3.1, 3.9, 4.7 g/L). Since the new subjects might not be able to identify all the mild difference between 1.5, 3.1, 3.9, 4.7 g/L, they were initially requested to test the difference between 0, 3.1 and 4.7 g/L. If yes, they were further to test the difference between 1.5, 3.1, 3.9, 4.7 g/L. At least they should be able to identify the difference of three solutions out of four cups and make the order of increasing sweetness correctly. The selected subjects were trained for two days to perform the imagination and taste experiments. To avoid the influence of hunger status of the subjects, the experimental data of the subjects were collected at a fixed time of the day (around one hour after food intake) while repeating the experiments on different days. During the first two days the subjects were trained to test sugar water of different concentrations to get used to the sweetness level in the laboratory.

To study the imagery influence on taste and to control the subjects’ imagery process, eight popular and well-accepted symbolic figures (see Figure 1) were chosen. The figures included: smiling face, sad face, neutral face, thumb-up, thumb-down, point finger, tick and cross symbols. All the figures were printed in black and white on one A4 size paper. Each subject was visually exposed to the figures on the paper for two minutes and the paper would be removed after that. Later the subject was requested by auditory machine speaker to recall or to imagine the figures during taste experiments. As control, the subjects were requested to taste the sugar solutions of different concentrations randomly without being exposed to visual stimuli or imagery cues. The sweetness sensitivity measured in the control experiments was analyzed and kept as a base line to observe the possible modification of the sweet taste sensitivity of subjects with additional imagery cues.

OPEN IN VIEWER

Figure 1.

The Eight Emoticons Applied in the Experiments as Imagery Clues. The upper illustrates smiling face, neutral face and sad face; the middle represents thumb-up, pointing finger and thumb-down; the below are the tick and cross symbols.

Procedure

During the tasting experiments, the participants sat in front of the table. Four cups of sugar solutions with different concentrations were placed on the table next to the participants. Before tasting, the participants would hear a request to recall or imagine one of the emoticons for 5 seconds. Following they sipped the sugar solution from the paper cup (around 12 ml) into the mouth and moved the tip of the tongue slightly, keeping the solution in the mouth for 5 seconds and spitted it out. During the following 50 seconds pause, the subjects rinsed mouth twice with distilled water, answered the questionnaire whether they detected sweet taste with the corresponding solution. We used ‘1’ and ‘0’ to record the results of the taste experiment. When the participants tasted the solution and detected the sweetness, ‘1’ is recorded; otherwise ‘0’ is recorded. The sugar solutions of different concentrations were provided to the participants in random order and the sound request to recall one of the eight figures was played randomly in a complete block design. Every participant needed to perform and complete the task with imagination of all the figures. Each participant repeated the whole set of sugar concentration for all the visual imageries 10 times. All the experiments were carried out at room temperature 20–25°C.

To quantify the preference associations of the visual imagery and the sweetness sensitivity, all the subjects were asked to rate to the preference scores of each imagery figure before the taste experiments. The range of the scores was from 1 to 7: ‘1’ means strong dislike; ‘2’ means dislike; ‘3’ means a bit dislike; ‘4’ means ordinary, does not matter; ‘5’ means a bit like; ‘6’ means like; ‘7’ means like very much. According to the average scores of all the subjects, we could sort the order of preference ratings of the imagery figures.

Data Analysis

All the data were recorded and saved in the computer (Window system 7) and were analyzed offline with MATLAB 7.9 (The MathWorks, Natick, MA). We calculated the sweetness detection ratio of each visual imagery for each person. The sweetness detection ratio = the number of times when sweet taste detected with one symbol imagination/the number of times of total experiments repeated with the corresponding symbol imagination. For each imagery, the sweetness detection ratio of each concentration was first calculated for each person. The average detection ratio and the standard deviation across all the subjects were calculated accordingly. For each concentration, multiple comparisons using one-way ANOVA was applied to test the significance of differences in the sweetness detection ratios with different imagination cues.

Affective Imagery Influences Sweetness Detection

With short recall or imagination of the affective emoticons, the sweetness detection of the subjects fluctuates. When subjects have the positive imagery before taste, such as smiling face, thumb-up and tick symbols, the sweetness detection increases significantly, compared with the negative symbols, such as sad face, thumb-down and cross. When the subjects recall the neutral symbols such as normal face or point finger, the sweetness detection ratio is almost between the ratio of positive and negative imageries. It is worthy to note, with different sugar concentrations, the fluctuations of the sweetness detection operate at different amplitudes. At low sugar concentrations, i.e. 1.5 and 3.1 g/L (the black dotted and the black lines in Figure 2), the sweetness detection ratio is around 30%–50% and 50%-80% respectively. When the sugar concentration increases to 3.9 g/L or 4.7 g/L (the grey dotted and the grey lines in Figure 2), the sweetness detection ratio is around 80%-90% and 85%-90% respectively. It is clear that the fluctuation of the sweetness detection ratio is much higher at the low concentration than at the relatively higher concentration. The reason might be the response saturation. On the other hand, at round 50% sugar detection ratio level, which is defined by the sweetness threshold, the sweetness detection is at the most uncertain area for the subjects. Namely, at the most ambiguous cognitive condition the taste perception is most effectible with imageries. Similar phenomenon has been observed in our previous studies (Liang et al., 2013, 2016). It has shown that the external visual influence on taste is limited around the sub-threshold sweet level. Compared to the control condition, i.e. when the subjects are without visual input or imagery, the positive imagery increases the sweetness detection. Particularly at the 1.5 and 3.1 g/L sugar water, positive imagery enhances the sweet sensitivity significantly, followed by the sweetness detection induced by neutral imagery symbols. (with one way ANOVA, p = 0.004 for 1.5 g/L and 0.0001 for 3.1 g/L)

OPEN IN VIEWER

Figure 2.

Visual Imagery Influences the Sweetness Detection. The grey line with circle, grey dotted line with triangle, black line with cross and black dotted line with star represent the sweetness detection of the sugar solution at 4.7, 3.9, 3.1 and 1.5 g/L respectively. X-label represents the visual imagery of the subjects before tasting the sugar solution. Control means subjects taste sugar water without visual input nor imagination. Y-label represents the sweetness detection. 1 means 100% of sweetness detection. Light grey horizontal dotted lines demonstrate the sugar detection level of four different concentrations respectively.

Pooled and Normalized Sweetness Detections With Affective Imageries

Although the fluctuations of sweetness detection of different sugar concentrations operate at different ranges, the patterns of the fluctuation are similar with all the sugar concentrations. To pool and compare the responses with different visual imagery directly, we normalized the data by dividing the responses of each sugar concentration with the responses of control condition respectively (detail in method data analysis). The normalized responses of the four different sugar concentrations are averaged and plotted in Figure 3. It supports the hypothesis that positive imagery enhances the sweetness sensitivity. With paired t-test, the sweetness detection ratios with the positive imagery (smile face, thumb-up and tick) are significantly higher than negative imagery (p < 0.05). Although the averaged sweetness detections of neutral imageries (normal face and point finger) lie between the positive and negative imageries, no statistically significant effect has been found. Hence, we may suggest that when subjects imagine positive pictures, they could be more sensitive to sweet taste than when they imagine negative pictures.

OPEN IN VIEWER

Figure 3.

Normalized Sweetness Sensitivity of All the Sugar Concentrations With Different Types of Imagery. X-label represents the same as in Figure 2. Y-label represents the normalized sweetness detection. Vertical lines indicate the standard deviation of the sweetness detection of all four different sugar solutions for each type of imagery.

Influence of Imagery on Sweetness Perception at Different Sugar Concentrations

From above analysis it is clear that the sweetness perception increases with positive imagery. How exactly is the sweetness perception associated with the subjects’ preference for pleasantness? To observe the associations in detail, we plotted the normalized sweetness detection as the function of hedonic scores for each sugar concentration, respectively (Figure 4). With different sugar concentrations, the data points have suggested that the sweetness sensitivity increases with hedonic scores. To quantify the associations more precisely, we have simulated linearly the relationship for each sugar concentration. As expected, the slope of each linear association of different sugar concentrations varies. When we shift the simulated lines vertically, so that all lines can start from the same point (Figure 4 inset) it demonstrates clearly that the slopes of 1.5 and 3.1 g/L are larger than those of 3.9 and 4.7 g/L. It suggests that the sweetness sensitivity enhances with hedonic ratings, particularly at low concentration, i.e. the sweetness detection above 30% and below 70%. More specifically, 1.5 and 3.1 g/L are below and around threshold level (50% sweetness detection), such situation provides the probably most uncertain zone for taste detection, where the taste perception would be mostly influenced by hedonics ratings.

OPEN IN VIEWER

Figure 4.

Association of Sweetness Sensitivity and Hedonic Scores. The grey circle, grey triangle, black cross and black star represent the sweetness sensitivity of sugar solution at 4.7, 3.9, 3.1 and 1.5 g/L respectively. The grey and grey dotted, black and black dotted are the simulated linear association of sweetness sensitivity and hedonic ratings for each concentration. X-axis represents the averaged hedonic scores of eight imagery figures. Note, the hedonic score of blank condition is set with value 4 by default. Y-axis represents the normalized sweetness detection ratio. The right-below inset illustrate the four simulated lines which were shifted to the same start point.

Imagination and Taste Perception

Previous research has demonstrated that mental imagery is processed similarly and activated similar brain cognitive areas as external environment sensory information process (Berger & Ehrsson, 2013). By only imagination auditory or visual object, the endogenous brain information may integrate with the exogenous sensory input from the real environment. Consistent with top-down theory, many studies have suggested that expectation, attention, awareness, and even memory load could modulate taste perception (Grabenhorst & Rolls, 2008; McClure et al., 2004; Woods et al., 2011; Zampini et al., 2007). For instance, visual inputs such as label or colour of drinks may induce subjects’ expectation and bias, and thus modulate the sweet taste perception (Woods et al., 2011; Zampini et al., 2007). All these findings support the top-down mechanism, that the primary taste cortex activity could be influenced by affective cognition centre of higher level. Our previous observations have shown that circular shapes and semantic familiar words enhance sweetness sensitivities (Liang et al., 2013, 2016). It has been suggested that the circular shapes lead to higher hedonic ratings than angular shapes, and the familiarity is positively associated with hedonic ratings (Delplanque et al., 2008). Both situations induce cognitive ease of the brain, and thus influence the taste perception. Here in this study our data further confirmed that visual imagery from the brain influences the sweetness perception. The positive affective imagery with higher hedonic ratings enhances the sweetness sensitivity (Figures 2 and 3). The pronounced sweetness modification by different imagery appears at around sweetness threshold level (Figure 2). The imagery induced cognitive status modulates sweet taste perception. It is worthy to note, that it is at the sub-threshold or around threshold level, which is an uncertain zone of brain perception, that additional input from external sensory system or the internal brain may modify the perception bias. Here in our experiments, instead of visual input, imagery is inducing similar modification of taste perception. Hence, from such observation, we may suggest that the neural networks between imagery and taste perception might share some similar pathways of visual-taste cross-modal interaction.

Imagery, Hedonics and Cognition

The subjects who were requested to recall one of popular emoticons were creating in their mind the picture and enrich in their inner world with the imagery. Implicit mental imagery occurs when perceptual information is accessed from memory (Kosslyn et al., 2001). Stokes et al. have demonstrated that visual imagery is mediated via top-down activation of functionally distinct, yet spatially overlapping population codes for high-level visual representation (Stokes et al., 2009). An fMRI study has found that the brain areas activated by visual imagery and perception are similar to a large extent, and the overlap is most pronounced in frontal and parietal regions than temporal and occipital regions (Ganis et al., 2004). The frontal and parietal lobes are the locations for the higher-level visual processing and for the emotion processing. For instance orbitofrontal cortex activity may reflect the top-down modulation of activation through expectations of drink liking (Kringelbach, 2005). Hippocampal modulation implies that expectation effects are recalled from memory and then go on to influence primary taste cortex (McClure et al., 2004).These brain areas are involved in the affective cognition process and may control the cognitive status of the brain, and thus modify the sensory perception.

Future Possibilities

Along the line of our hypothesis, the brain’s cognitive status is determined by sensory inputs of the external environment and the internal brain itself. Here the brain’s cognitive status is a complex term including emotion, awareness, cognition etc. To further test our hypothesis, we may manipulate the brain’s cognitive status by different sensory modalities and different cognitive patterns. For instance, how does different cognitive status induced by memory load influence the taste perception? How does the different cognitive status induced by other sensory modalities such as music or different tactile stimuli influence the taste perception? It would be important and interesting to systematically manipulate the cognitive status and test the taste or other sensory perception. Hence, a lot of systematic work needs to be done to understand further the cognitive status and perception in behavioral and neural levels.