Determination of Recognition Threshold and Just Noticeable Difference in the Sensory Perception of Pungency of Zanthoxylum bangeanum

Abstract

Zanthoxylum bungeanum, also called the Szechuan pepper, is a pungent-flavored spice. Together with the chili pepper, it constitutes the typical flavor of famous Chinese Szechuan cuisine. With the worldwide popularity of Sichuan cuisine and pungent food, the research on sensory perception of pungency of Z. bungeanum becomes more important and necessary. This article first determined the sensory recognition threshold and just noticeable difference of pungency of Z. bungeanum. The pungent extract, containing hydroxyl α-sanshool (63.6%), hydroxyl β-sanshool (22.8%), and hydroxyl γ-sanshool (7.4%) from Z. bungeanum was used as the tested sample prepared in ethanol–water solution. Ten selected assessors participated in three sessions of sensory tests in standard sensory evaluation booths established according to ISO 8589 by following two-alternative forced choice procedure. The data showed that the recognition threshold was 0.036 mg pungent extract per 1 mL ethanol–water solution and just noticeable difference exponentially increased with the increase of the concentration of pungent extract solution. The response of pungency among assessors followed Weber’s Law with a stable Weber fraction (0.118) from 0.0360 to 0.500 mg pungent extract per 1 mL. Nevertheless, the fraction changed when the concentration of pungent extract solution was above 0.500 mg per 1 mL.

Keywords:

• Determination
• Threshold
• Sensory
• Perception
• Pungency

INTRODUCTION

Zanthoxylum bungeanum, also called the Szechuan pepper, is a pungent-flavored spice. Together with the chili pepper, it constitutes the typical flavor of famous Chinese Sichuan cuisine. With the worldwide popularity of Sichuan cuisine and extensive application of Z. bungeanum in the food industry, research on the sensory perception of pungency of Z. bungeanum becomes more important and necessary. To date, although researchers worldwide have conducted some studies on the pungency of Z. bungeanum, their work related to sensory perception is scant. Instead, the reported researches mainly focused on two aspects: (1) extraction and separation of the pungency compound from Z. bungeanum,^[1–3] including instrumental quantitative analysis methods;^[4,5] (2) the biological mechanism underlying its pungency.^[6–9]

It is well known that one basic study of sensory perception is determining thresholds. As for the pungency perception of the Szechuan pepper or sanshool compounds, only some work on recognition threshold but no just noticeable differences (JNDs) has been reported presently. Etsuko and others^[10] reported that the average threshold values of α-sanshool, β-sanshool, γ-sanshool, δ-sanshool, and hydroxyl α-sanshool, hydroxyl β-sanshool were 1.3, 1.4, 0.9, 0.9, 3.8, and 7.8 g per 1 mL, respectively. Fu^[11] also reported that the estimated recognition threshold of pungent oil and Szechuan pepper were 2.25, 35.5 mg per 100 mL, respectively. Except for these reports we mentioned, no other related work could be found. Fortunately, in spite of inadequate study of pungency perception of the Szechuan pepper or sanshool compounds, the research work on pungency of the chili pepper or capsaicin was relatively abundant. It was demonstrated that the pungency of sanshool compounds and capsaicin invoke the same kind of trigeminal nerve sensation^[6] and that the pungency sensation from capsaicin and sanshool compounds is, to some extent, similarly related to the activation of TRPV1 and TRPA1 in sensitive neurons.^[8] However, pungent agents from Szechuan peppers were also reported to excite sensory neurons by inhibiting two-pore potassium channels. These finding means that we can use the previous sensory research of capsaicin^[12–15] for reference before we will begin to study the thresholds of pungency.

Additionally, from a technical perspective, determination of both recognition threshold and JNDs of pungency of the Szechuan pepper is also very important. It is not only useful for grading of dried Szechuan peppers and their products based on pungency intensity, but also very helpful for developing new products to satisfy the expected pungent flavor by target consumers. For food flavors, JNDs means a lot to set the acceptable and unacceptable levels of flavor components. For example, a negative change, such as reduction in an expensive ingredient in a formula, will not be readily discernible to consumers, or a positive change, such as an increment in the amount of a consumers’ favorable ingredient, will be very apparent to consumers.^[16] Moreover, according to the law of Weber-Fechner,^[17] we concluded that the value of thresholds can also be used for quality grading by providing a perceived range each grade of products.

Previous studies have provided several methods to determine sensory thresholds, such as Scoville method,^[12] magnitude estimation,^[18] and alternative forced-choice (n-AFC)^[19] methods. These are workable methods, but there are some potential limitations. The Scoville method is based on determination of recognition threshold in serially diluted samples was the classical method for quantification of heat intensity. However, it has been taken no consideration for the interference of matrix (such as ethanol, tween-80) and the statistical basis for data processing procedure was also missing. The results of magnitude estimation methods have ratio problems such as strong favored-number or round-number effects^[20] and showed high variability and poor reproducibility.^[21] The n-AFC method has been described as an authoritative standard^[19] and recommended in two of the major sensory evaluation textbooks.^[17] It has proven to be a useful method for estimating thresholds of flavor and taste materials.^[22] Unlike the method of magnitude estimation and Scoville implemented by experienced assessors, n-AFC has been usually carried out by two kinds of groups, untrained consumers or trained assessors. On one hand, Orellana has confirmed that the threshold determined by untrained consumers was defined as the stimulus concentration at which the proportion of correct responses was 76%.^[15] From his study, we deduce that assessors used for the measurement of pungency threshold in different districts (such as in Sichuan or Shanghai) will present different threshold values, on account that Sichuan people, who consume Szechuan peppers frequently, will present a higher values than people living in Shanghai, who do so less frequently.^[23,24] On the other hand, Gonzalez-Viñas and others^[25] organized trained assessors to determine the group threshold of basic, umami and metallic tastes by the 3-AFC method and found that the method provided a rapid and simple means of estimating recognition thresholds. As for JND, there is rare research on JND for pungency, L Orellana-Escobedo^[15] determined five JND values using five reference sample solutions with different capsaicinoids concentration by 2-AFC methodology and he also summarized that JND values changed proportionally as the reference sample solution changed. Therefore, in this study, the pungency threshold values, including recognition threshold and JND threshold, was attained by 2-AFC method^[19] and calculated by best estimated threshold (BET; ISO: 13301 2002) with 10 experienced (trained) assessors instead of untrained consumers participated in each sensory test. Finally, the relationship between JNDs and the physical concentration was also studied to explore the response regularity in the sensory perception of pungency of Z. bungeanum.

MATERIALS AND METHODS

Stock Preparation

The pungent extract, containing hydroxyl alpha-sanshool (63.6%), hydroxyl beta-sanshool (22.8%), and hydroxyl gamma-sanshool (7.4%), was extracted and purified from Z. bungeanum by high-performance liquid chromatography (HPLC) and used to make the stock solution (10 mg pungent extract per 1 mL absolute ethyl alcohol). Afterward, the stock solution was kept at 4°C.

Sample Preparation for Recognition Threshold

Stock solution was diluted with spring water to concentrations of 0.625, 1.25, 2.50, 5.00, 10.0, 20.0, and 40.0×10^–2 mg per 1 mL. The control solution was 0.1% EtOH aqueous solution.

Sample Preparation for JND Threshold

Twenty reference samples covering the lowest to the highest pungency intensities were prepared for JND threshold determination, and then each reference (or control) sample corresponding to five dilutions was ready, respectively. Ultimately, 120 pungent solutions were obtained (Table 1). All the sample solutions and rinsing water in our study were kept at room temperature before testing and identified with randomly assigned numbers.

TABLE 1 Concentration of the pungent extract used for JND threshold determination

Pungencyintensity	Concentration of referencesolutions (c, mg/mL)	Concentration of sample solutions (c, mg/mL)
		1st	2nd	3rd	4th	5th
Low	0.0360	0.0378	0.0397	0.0417	0.0438	0.0459
	0.0438	0.0460	0.0483	0.0507	0.0532	0.0559
	0.0532	0.0559	0.0587	0.0616	0.0647	0.0679
	0.0647	0.0680	0.0713	0.0749	0.0786	0.0826
	0.0786	0.0830	0.0867	0.0910	0.0955	0.100
Moderate	0.0955	0.100	0.105	0.111	0.116	0.122
	0.116	0.122	0.128	0.134	0.141	0.148
	0.141	0.148	0.156	0.163	0.171	0.180
	0.163	0.171	0.180	0.189	0.198	0.208
	0.198	0.208	0.218	0.229	0.241	0.253
	0.241	0.253	0.266	0.279	0.293	0.308
	0.279	0.293	0.308	0.323	0.339	0.356
	0.308	0.323	0.340	0.357	0.374	0.393
Strong	0.323	0.339	0.356	0.374	0.393	0.412
	0.373	0.392	0.411	0.432	0.453	0.476
	0.411	0.432	0.453	0.476	0.500	0.525
	0.453	0.476	0.500	0.525	0.551	0.578
	0.500	0.525	0.550	0.578	0.607	0.638
	0.578	0.607	0.638	0.669	0.703	0.738
	0.600	0.630	0.662	0.695	0.729	0.766

Assessor Training

Ten assessors (five males and five females) aged from 20–26 participated in this study. Prior to their evaluation, they were confirmed to show no pungency-loving or exclusion feelings, were verified to have normal taste and pungency sensitivity, and meet the requirements of ISO 8586.^[26] Assessor training consisted of two 1-hour sessions which involved introducing pungency flavor at low concentration of pungent solutions and the 2-AFC method used to determine thresholds. Assessor were first told to hold each sample in their mouth for 180 s before expectorating it, and then, the assessors discussed and developed appropriate descriptors of pungency and reached consensus on the following descriptors of tingling and astringent taste. Afterward, assessors were instructed to be familiar with the 2-AFC method. They were asked to give a response of more pungent sample between three randomized pairs of target (10.0×10^–2 mg pungent extract per 1 mL ethanol–water solution, extremely low intensity but with the tingling sensation) and blank (10.0 × 10^–2 mg EtOH per 1 mL water solution) trials. If all assessors made correct responses, the 10 experienced and trained assessors were asked to participate in three sessions of sensory tests in standard sensory evaluation booths established according to ISO 8589^[27] by the two-alternative forced choice (2-AFC) procedure.

Recognition Threshold and JND Determination

Threshold was determined by the 2-AFC methodology.^[19] Considering the order effects, assessors were given one fixed reference solution (containing 0.1% EtOH aqueous solution for recognition threshold while less sanshool compounds contained for JNDs) and the corresponding test solution (0.625 ~ 40.0×10^–2 mg sanshool compounds per 1 mL and Table 1, respectively). The presentation order of all samples was randomized and counterbalanced. Each assessor took 20 mL sample, kept the sample in their mouth for 180 s, rinsed the palate using spring water, repeated the same procedure for the second sample, and then chose which solution was more pungent. Afterward, the five min interstimulus interval was chosen based on the sensory literature.^[18] At the end of test, the maximum concentration of negative responses and the minimum concentration of three consecutive positive responses were recorded by the sensory analyst. Note that the value of thresholds were specific for people aged from 20–26.

Statistical Analysis

The individual BET was determined as the geometric mean of the highest concentration missed and the next higher concentration. The group BET was calculated as the geometric mean of the individual BET. Standard deviation log10 provided a measure of the panel’s variation.

RESULTS AND DISCUSSION

Recognition threshold

Personal thresholds of each assessor and the panel recognition threshold determined by the 2-AFC method and calculated by the BET method are presented in Table 2. None of assessors could perceive the presence of pungency exactly at the lowest concentration (0.625×10^–2 mg/mL) and only one of them could detect at the concentration next to the second. Meanwhile, almost all assessors gave positive responses when the concentration of pungent extract was greater than 20.0×10^–2 mg/mL. According to the requirement of ISO 13301,^[19] personal recognition threshold was calculated by the BET method, and then the panel recognition threshold was taken as the geometric mean of personal BET. Therefore, in this study, personal threshold was determined on the basis of the maximum concentration of negative responses and the minimum concentration among three consecutive positive responses using the BET method. Thus, the panel’s (aged from 20–26) mean recognition threshold was 0.036 mg/mL pungent extract in ethanol–water solution.

TABLE 2 Recognition threshold value of the pungent extract measured by two alternative-forced-choice method

Assessors	Concentration (10^–2 mg/mL)							Best estimatethreshold (BET,10^–2 mg/m)
	0.625	1.25	2.50	5.00	10.0	20.0	40.0
No. 1	–	+	+	–	+	+	+	0.884
	–	+	+	+	–	+	+	0.884
	–	–	+	+	–	+	+	1.77
No. 2	–	–	+	+	+	+	+	1.77
	–	+	–	+	+	+	–	3.54
	–	–	+	+	+	+	+	1.77
No. 3	+	–	+	+	+	–	+	1.77
	–	–	–	+	+	+	+	3.54
	–	–	+	+	–	+	+	1.77
No. 4	–	–	+	+	+	–	+	1.77
	–	–	–	+	+	+	–	3.54
	–	–	+	+	–	+	–	1.77
No. 5	–	–	–	–	+	+	+	7.07
	–	+	–	–	–	+	+	14.1
	+	–	–	–	+	+	+	7.07
No. 6	–	–	+	–	+	+	+	7.07
	–	–	+	+	+	+	–	1.77
	–	+	–	+	+	–	+	3.54
No. 7	–	–	–	+	–	+	+	14.1
	–	–	+	–	+	+	–	7.07
	–	–	–	–	+	+	+	7.07
No. 8	+	–	–	+	+	+	+	3.54
	+	–	+	+	–	+	+	1.77
	–	–	+	+	+	+	–	1.77
No. 9	–	–	–	–	+	+	+	7.07
	–	–	–	–	+	+	+	7.07
	–	–	–	–	+	+	–	7.07
No. 10	–	–	+	–	+	+	+	7.07
	–	–	–	–	+	+	+	7.07
	–	–	+	–	+	+	+	7.07
Sensory panel mean threshold (10^–2 mg/mL) = Geometric mean of personalBET of each assessors								3.62
Standard deviation								0.34

“–” indicates that the assessor selected the wrong sample of the set of the two; “+” indicates that the assessor selected the correct sample.

Etsuko and others^[10] have reported that the recognition threshold of hydroxyl α-sanshool and hydroxyl β-sanshool in ethanol–water solution was 0.038 and 0.078 mg/mL, respectively. Although it was absent, the reference value of hydroxyl γ-sanshool, Kenji^[28] and Kashiwada^[2] reported that hydroxyl α-sanshool and hydroxyl γ-sanshool, having one cis and three or four trans-double bonds in the molecule, were strongly pungent, while hydroxyl β-sanshool, which has all-trans double bonds, was tasteless. As we stated above, in this study, only 7.4% of the pungent extract consisted of hydroxyl γ-sanshool, 63.6% hydroxyl α-sanshool, and 22.8% hydroxyl β-sanshool. Therefore, our result (0.036 mg per 1 mL) is in keeping with the results from Etsuko’s study.^[10]

JND Threshold

JND values indicate the amount of additional stimulus required to perceive a difference between two samples which both contain the stimulus. Twenty JND values at different referential concentrations calculated by BET method are presented in Table 3. In order to explore the response regularity in the sensory perception of pungency of Z. bungeanum, further studies on the relationship between JNDs and the physical concentration were made. Weber’s law is one of the most famous psychophysics formulas to describe the relationship of stimulus-intensity and perception-intensity and state that the JND between two stimuli is a fixed proportion of the value of the stimuli being judged. Weber fraction (k) is calculated using the form:

TABLE 3 JND threshold value of the pungent extract at different concentrations

Pungencyintensity	Concentration for referencesolutions (c, mg/mL)	JND (Δc, mg/mL)	Δc/c
Low	0.0360	0.00330	0.0917
	0.0438	0.00373	0.0852
	0.0532	0.00486	0.0914
	0.0647	0.00585	0.0904
	0.0786	0.00698	0.0888
	0.0955	0.00861	0.0902
Moderate	0.116	0.0106	0.0914
	0.141	0.0140	0.0993
	0.163	0.0153	0.0939
	0.198	0.0187	0.0944
	0.241	0.0230	0.0954
	0.279	0.0271	0.0971
	0.308	0.0305	0.0990
Strong	0.323	0.0326	0.101
	0.373	0.0382	0.102
	0.411	0.0525	0.128
	0.453	0.0549	0.121
	0.500	0.0562	0.112
	0.578	0.102	0.177
	0.600	0.128	0.213

where k is a constant, Δc is the increase in the physical stimulus needed to be judged different (JND), and c is the starting level of the stimulus (concentration of reference solution). Table 3 presented that calculated Weber fraction values were similar (0.0852–0.112) among tests conducted with reference solution ranging from 0.036 to 0.500 mg per 1 mL. However, the reference solution with the concentration ranging from 0.500 to 0.600 mg per 1 mL showed a different Weber fraction (0.177, 0.213). Then, according to the fitting curve (Fig. 1), we learned that the slope of the fitting curve was 0.118. It also meant that the pungency response to Z. bungeanum extract is in congruence with Weber’s law with a stable Weber fraction (0.118; Fig. 1). However, when the concentration of pungent extract solution was above 0.500 mg per 1 mL, the Weber fraction (k) increased above 0.118 (0.178, 0.213, respectively; Table 3), indicating that a larger increase in concentration was required to produce the same increase in perceived pungency.

FIGURE 1 The relationship between Delta C and the concentration of pungency extract solution (α = 0.05).

This phenomenon may be attributed to the nature of the stimulus produced by pungent extract, which produces fatigue when the aftertaste of a high-intensity sample interferes with the tasting of a subsequent sample. A desensitization effect may also explain the sensory perception of pungency of the Szechuan pepper, as a similar variation trend has been recorded for the chili pepper.^[15] Furthermore, Goldstein^[29] has already demonstrated that Weber’s law holds a stable fraction for most senses as long as the stimulus intensity is not too far from the threshold. This information suggests that an ethanol–water solution with a concentration of 0.500 mg pungent extract per 1 mL is probably far enough from the recognition threshold concentration (0.036 mg pungent compound per 1 mL) to cause deviations from Weber’s law.

CONCLUSIONS

The recognition threshold for the sensory perception of Z. bungeanum was 0.036 mg per 1 mL. JND thresholds exponentially increased with the increase of the concentration of pungent extract solution. Weber fraction (k) kept stable as 0.118 from 0.0360 mg to 0.500 mg pungent extract per 1 mL. When the concentration of pungent extract solution was above 0.500 mg per 1 mL, the Weber fraction changed observably. Recognition threshold and differential thresholds of pungent compounds can be used to guide the design and production of the new pungent products. Moreover, it is also useful for quality grading of the dried Chinese pepper or other pungent products according to its pungency intensity. However, the threshold determined in the present article is only applicable to the ethanol–water medium at room temperature, because flavor compounds show different threshold values in different kinds of media, serving temperatures, fat levels, and intensity of tastes, textures, and varieties of flavor. Therefore, interaction effects of different factors to the sensory perception of pungency of Z. bungeanum will be carried out and reported in our follow-up studies.

ACKNOWLEDGMENT

The authors want to express their heartfelt gratitude for the participation of assessors in our sensory study.

FUNDING

The work was supported by “Research on the Chemical Basis of Pungency Intensity in Z. Bungeanum (National Science Foundation Project, item number: 31171695),” and “Research on Development and Application of Flavor Reference Sample Used in Sensory Evaluation (The Fundamental Research Funds for the Central Project, item number: 562014Y-3348).”

Additional information

Funding

The work was supported by “Research on the Chemical Basis of Pungency Intensity in Z. Bungeanum (National Science Foundation Project, item number: 31171695),” and “Research on Development and Application of Flavor Reference Sample Used in Sensory Evaluation (The Fundamental Research Funds for the Central Project, item number: 562014Y-3348).”