Surimi (or fish paste) products are the most representative processed fishery foods. They are widely used because their cooking is simple and inexpensive. In addition, surimi products can easily change the shape, taste, and texture of the final product as per consumer preferences (Kim et al., 2005). In South Korea, the most popular surimi product is the fried type, which is traditionally mixed with vegetables (Park, 2005a; Choi et al., 2012). Currently, most surimi products are prepared using starch or wheat flour as an ingredient. Unfortunately, the price of wheat flour has increased and consumers were required high-quality products. Therefore, there is a need to come up with a cheaper alternative to replace wheat flour. Recently, rice flour was highlighted as a possible alternative to wheat flour (We et al., 2011).

Rice is the major staple crop of nearly half of the world’s population, and is particularly important in Asia, where approximately 90% of the world’s rice is produced and consumed (Zeigler and Barclay, 2008; Khush, 2004; Lee, 2012). However, rice consumption in South Korea has decreased sharply owing to westernization of diet and increasing popularity of instant foods (Cha, 2010; Statistics Korea, 2011). Most rice are consumed after simple cooking and therefore the South Korean food industry is exploring the use of rice in new foods, both for domestic consumption and for export purposes (Shin, 2009). The rheological properties and manufacturing process of surimi products containing different starches or functional powders, such as potato (Yang and Park, 1998; Tabilo-Munizaga and Barbosa-Canovas, 2005; Chung and Lee, 1996), waxy corn (Alvarez et al., 1997), wheat (Kong et al., 1999), red ginseng (Shim et al., 2012), Lycii fructus (Shin et al., 2008), and Poria cocos (Shin et al., 2009) have been reported. Currently, there exists few papers on the use of rice flour as a gel-enhancing ingredient in surimi products studied (Jung and Yoo, 2005; Jung et al., 2007; Hur et al., 2011). However, the relationship between rice flour and surimi with regard to rheological properties and sensory evaluation has not been fully elucidated. In particular, no attempt has been made to optimize the content of rice flour in surimi products. In a previous study, we evaluated the effect of rice flour addition on the rheological and sensory properties of surimi products to check for the feasibility of using rice flour as an alternative to wheat flour for the production of premium surimi products (Cho et al., 2013). Therefore, the objective of the present study was to determine the optimum formulation of surimi products containing rice flour by using response surface methodology.

Golden threadfin bream (Nemipterus virgatus) surimi (grade A; Khan Hoang Seaprex Co., Ltd., Soc Trang, Vietnam) was used in this study. (This surimi contained the usual cryoprotectants (e.g., sucrose, sorbitol, sodium polyphosphate, and sodium pyrophosphate)). Randomly selected surimi blocks (10 kg) were cut into 1 kg portions and kept at -20℃ until further use. Rice flour (roll-mill, 80 mesh) was purchased from Daesun Flour Mills Co., Ltd. (Hampyeong, Korea). The other ingredients (Table 1) were provided from Hansung Enterprise Co., Ltd. (Seongnam, Korea).

[Table 1.] Ingredients (？？) of the preparation of surimi products containing rice flour

Ingredients (？？) of the preparation of surimi products containing rice flour

Frozen surimi was partially thawed for 1 h at room temperature, to allow the core temperature to reach approximately -5℃. The thawed surimi was then cut into cubes (5 × 5 × 5 cm) and placed in a chopping bowl. The sample was chopped further at 1500 rpm for 5 min by a silent cutter (OMF-500, Ohmichi Co., Ltd., Maebashi, Japan). Two percent salt was added and the sample chopped for another 5 min at 1500 rpm. Rice flour was added slowly until homogenization of the mixture. The other ingredients were then added to the homogenized mixture according to the amounts listed in Table 1 and 168mixed for 3 min by the silent cutter. The formulation of Table 1 was selected because it is a general combination of ingredient for the production of commercial surimi products. During the chopping and mixing procedure, the temperature was maintained at approximately 5℃ by crushed ice bath. After molding of the mixture, it was fried in soybean oil for 3 min at 150℃. All samples were then vacuum-packed and stored at 4℃.

Breaking force and deformation of the surimi product was measured by a puncture test using a TA-XT2 texture analyzer (Stable Micro Systems, Surrey, UK) equipped with a 0.25-inch spherical stainless probe. When the gel sample lost its strength and ruptured at a crosshead speed of 1.0 mm/s, the breaking force and depth of depression were recorded. Ten cylindrical samples with a length of 2.5 cm were subjected to the test. Gel strength was calculated using the following formula: Gel strength (g·cm) = breaking force (g) × depth of depression (cm).

A trained panel comprising 20 members of the Korea Food Research Institute determined the organoleptic characteristics of the sample. The outward appearance, color, flavor, taste, chewiness, hardness, and overall acceptance were evaluated by a 7-point Likert scale (1 = extremely dissatisfied, 7 = extremely satisfied).

Central composite design (CCD) (Box and Wilson, 1951) was adopted in the optimization of rice flour and water content proportions in surimi products (Fig. 1). CCD matrix consists of 2² factorial points, 4 axial points (α = 1.414) and 3 center points. Rice flour content (X₁, %) and water content (X₂, %) were chosen as the independent variables (IVs). The range and center point values of the two IVs were based on the results of preliminary experiments (Table 2). Commonly, the minimum content of wheat flour was 5% in the commercial surimi products. In addition, when the rice flour added more than 15%, the prototype surimi products was not forming gel. Therefore, we chose the range of rice flour content from 5% to 15%. Gel strength (Y₁, g∙cm ) and overall acceptance (Y₂) were selected as the dependent variables (DVs) for the different combinations of IVs (Table 3). To minimize the effects of unexpected variability in the observed responses, the experimental runs were randomized.

[Fig. 1.] The example points of a central composite design (CCD) with two input parameters.

[Table 2.] Experimental range and values of the independent variables in the central composite design for preparation of surimi products containing rice flour

Experimental range and values of the independent variables in the central composite design for preparation of surimi products containing rice flour

[Table 3.] Central composite design and responses of the dependent variables for preparation of the surimi products containing rice flour

Central composite design and responses of the dependent variables for preparation of the surimi products containing rice flour

For the response surface regression procedure, MINITAB software version 14 (Minitab Inc., Harrisburg, PA, USA) was used to fit the following quadratic polynomial equation:

where Y is the dependent variable (gel strength and overall acceptance), β₀ is a constant, βi, βii and βij are regression coefficients and Xi, Xj are the levels of the independent variables. The DVs, Y₁ (gel strength) and Y₂ (overall acceptance), were considered as parameters for optimization in order to prepare a sample with the highest gel strength and overall acceptance. The response surface plots were developed using Maple software version 7 (Waterloo Maple Inc., Ontario, Canada) and represented a function of two IVs.

Statistical analyses of data were performed using SPSS software Version 13 (SPSS Inc., Chicago, IL, USA). Results were presented as average and standard deviations. Tukey’s multiple range test was carried out with one-way analysis of variance (ANOVA) to determine the significance of the differences in the mean among treatments. The significance level was set at P < 0.05 for all analyses.

Response surface methodology (RSM) is a useful statistical technique that has been applied to optimize formulations in various food products (Murphy et al., 2004). This methodology reduces the number of experimental runs required to generate sufficient information in order to provide a statistically acceptable result (Fogaça et al., 2013). The response surface regression procedure was employed to fit the quadratic polynomial equation to the experimental data. All the coefficients of linear (X₁, X₂), quadratic (X₁X₁, X₂X₂) and interaction (X₁X₂) were calculated for its significance using the t-test, and the estimated coefficients of each model are presented in Table 4. In all models, the constant coefficients and linear terms were highly significant (P < 0.01). In contrast, the X₁X₁ term of quadratic coefficients and interaction coefficients of Y₁ (gel strength) and Y₂ (overall acceptance) were not significant. However, the X₂X₂ term of Y₁ (gel strength) and Y₂ (overall acceptance) were significant (P < 0.05). To develop the fitted response surface model equations, all insignificant terms (P > 0.05) were estimated and the corresponding fitted models are shown in Table 5. Values of the determination coefficient (R²) for Y₁ and Y₂ were 0.946 and 0.965, respectively, and significant at the 99% probability level. The R² values indicate that the model equations described the experimental designs adequately (Cho et al., 2005). This result confirmed that the experimental design based on the preliminary test was adequately performed.

[Table 4.] Estimated coefficients of the fitted quadratic polynomial equation for responses based on t-statistic for preparation of surimi products containing rice flour

Estimated coefficients of the fitted quadratic polynomial equation for responses based on t-statistic for preparation of surimi products containing rice flour

[Table 5.] Response surface model for preparation of surimi products containing rice flour

Response surface model for preparation of surimi products containing rice flour

The statistical significance of the quadratic polynomial model equation was evaluated by analysis of variance (ANOVA). Table 6 shows ANOVA results for the models that explain the response of the two dependent variables, Y₁ (gel strength) and Y₂ (overall acceptance). The linear terms (X₁, X₂) for all the DVs were significant (P = 0.001 and P = 0.000, respectively) at the 99% probability level, whereas cross-product terms for all the DVs were not significant (P = 0.146 and P = 0.134, respectively) at the 95% probability level. The quadratic terms (X₁X₁, X₂X₂) of all the DVs were not significant (P = 0.062 and P = 0.077, respectively) at the 95% probability level. In the total regression model, all the DVs were significant (P < 0.01) at the 99% probability level. If the lack-of-fit test is not significant, then the model is fit (Isa et al., 2011). The results of the lack-of-fit test show that the F values of Y₁ and Y₂ were 11.62 and 2.91, respectively, and the related P-values were not significant (0.080, and 0.266, respectively) at the 95% probability level. These results indicate the suitability of the models to accurately predict variation.

[Table 6.] Analysis of variance (ANOVA) for response of dependent variables (Y1 and Y2)

Analysis of variance (ANOVA) for response of dependent variables (Y1 and Y2)

The estimated response function and the effect of the IVs on the DVs are shown in Fig. 2. When X₁ increased from 5 % (-1.414) to 15 % (+1.414), Y₁ and Y₂ increased. However, when X₂ increased from 7 % (-1.414) to 14 % (+1.414), Y₁ and Y₂ were decreased. In a study by Chen et al. (1993), starch addition up to 8% increased gel strength while lowering the expressible moisture value. This result is similar to our result where we show that the gel strength and moisture values have an inverse relationship. In addition, in a previous study, we suggested that an effective amount of rice flour to add was 10-15% (Cho et al., 2013). This result supported that rice flour content up to 15% increased gel strength and overall acceptance. As the coded values of IVs X₁ and X₂ were reached 13.5% (+1) and 9.25% (-0.5), respectively, Y₁ (gel strength) and Y₂ (overall acceptance) were increased. Generally, 4-12% of starch or wheat flour is added to surimi product (Park, 2005b). However, up to 20-26% has been used for reducing the cost of surimi products (Hunt et al., 2009; Shim et al., 2012). Therefore, we suggest that rice flour can be more useful than wheat flour for reducing the costs of premium surimi products.

[Fig. 2.] 3D plot for preparation of surimi products containing rice flour for gel strength (A) and overall acceptance (B). X1, rice flour content (%, w/w); X2, water content (%, v/w); Y1, ？？el strength (？？·cm); Y2, overall acceptance.

Optimal conditions included both coded and uncoded values of the dependent variables (Y₁, and Y₂), which are shown in Table 7. According to the canonical analysis of the response surface regression procedure, the eigenvalues of Y₁ was positive and the eigenvalue of Y₂ was negative. Therefore, the stationary points were target values. To obtain the optimal proportion of rice and water in surimi products, each target value, Y₁ (gel strength) and Y₂ (overall acceptance), were considered as a parameter during the optimization. The multiple optimal conditions (coded and actual values) of X₁ (rice flour content) and X₂ (water content) were 1.15 (14 %) and -0.55 (9.1 %), respectively. It has been reported that starch increases the gel strength of surimi products at lower concentration of 3% instead of higher concentrations of 6-10% (Yoon et al., 1997). In contrast, our results with rice flour show the opposite trend. It is considered that the gel strength was dependent on the inherent properties of starch and other ingredients used in the surimi product formulation (Hunt et al., 2009). The predicted values of Y₁ (gel strength), Y₂ (overall acceptance) at optimal conditions were 656.4 g∙cm and 6.34, respectively, and under optimal conditions, the experimental values were similar to the predicted values of Y₁ and Y₂ (647.8 g∙cm and 6.21, respectively). Therefore, the estimated response surface model was adapted for optimization of rice flour content in surimi products. Surimi products produced under optimum conditions were similar in strength to the gel strength of commercial products. However, its sensory evaluation score was higher than that of commercial products (Fig. 3). In particular, typical fishy smell, which is a major cause of evasion of surimi product (Lee and Lee, 2002), was reduced by adding rice flour. Therefore, rice flour enhances the quality of surimi products and is considered a valuable ingredient in the preparation of premium surimi products.

[Table 7.] Multiple response optimization for the preparation of surimi products containing rice flour of predicted and experimental values

Multiple response optimization for the preparation of surimi products containing rice flour of predicted and experimental values

[Fig. 3.] Sensory evaluation of the prototype surimi products (PSP) and the commercial surimi product (CSP).