Introduction

Thermoplastic extrusion is a technique, used in food production for more than 60 years, in which a raw material is subjected to high temperature and pressure in a short time and combines several simultaneous unit operations, such as: mixing, baking, kneading, molding and shearing. The characteristics and quality of extruded products depend on process variables (temperature, residence time, pressure, screw rotation speed, feed rate and die diameter) and on the raw material used (composition, granulometry and moisture) (Conti-Silva et al. 2012; Fiorda et al. 2015; Prabha et al. 2021). During extrusion, chemical changes occur in the food, such as protein denaturation, starch gelatinization, fiber solubilization, minimal nutritional losses and decreasing in levels of antinutritional factors, as well as changes to the physical properties, like the expansion and texture of the product (Fiorda et al. 2015; Leonard et al. 2020). This process is widely used in the food industry, including for breakfast cereals, snacks, baby and animal foods, confectionery and protein-based products (Leonard et al. 2020).

Among these products, corn-based expanded snacks stand out for their large consumption, different flavors and practicality. Consumer acceptance is directly related to value and convenience, as well as physical and sensory characteristics, expansion, density, hardness, texture, flavor and color of extruded snacks (Saldivar 2016; Ziena and Ziena 2022) and any small change during extrusion modifies both the structural, physical, chemical and sensory characteristics of the product. However, it is known that these products have low nutritional value and high lipid content, going against market trends, since consumers are looking for a practical and, above all, a healthy diet (Menis-Henrique et al. 2019; Neder-Suárez et al. 2021; Ziena and Ziena 2022).

Determining factor for the caloric value of snacks is their flavoring method, the post-extrusion flavoring that is the most used in the food industry. In this technique, vegetable oil or hydrogenated vegetable fat is sprayed onto the extrudates and subsequently flavored with other ingredients (salt, additives and seasonings). There are some disadvantages to this, such as an increase of up to 25% in the initial weight, non-uniform distribution of ingredients, high concentrations of sodium and lipids as well as greater risks of oxidation since this aromatization is carried out in an external environment in contact with oxygen (Bhandari et al. 2001). An alternative to produce healthier extruded foods with lower lipid content is the pre-extrusion flavoring method using aroma precursors, highlighting butyric acid and cysteine for obtaining odor and flavor cheese (Menis-Henrique et al. 2019, 2020), in which no lipid vehicle is needed and greater uniformity distribution of compounds and stability to the oxidation of volatile compounds responsible for the odor and flavor of foods occur. However, lipids are of great importance in the perception of sensory attributes, influencing appearance, texture, consistency and melting; thus, their reduction can result in low acceptance of the products (Bhandari et al. 2001; Saldivar 2016).

Still from this perspective of healthier products, those for vegans and vegetarians are on the rise, due to their sustainability and the growing concern about the impacts that livestock cause on the environment, especially the climate and water resources (Lee et al. 2022; Zhang et al. 2021). Plant-based products are a possible solution to these problems since they require fewer natural resources (Lee et al. 2022) and meet these market trend (Chiang et al. 2021). However, the acceptance of meat analogs or meat substitutes is a challenge in terms of appearance, texture and flavor, and the more the product resembles meat, the greater its success (Lee et al. 2023).

Soy protein is one alternative for producing meat analogs and thermoplastic extrusion is the most used technology for it, transforming soy protein into a product that really resembles meat texture. For this, hydration after extrusion is needed for it to become soft and juicy (Baune et al. 2022; Lee et al. 2022). On the other hand, the textured soy protein may have an undesirable flavor (Lee et al. 2022), and thus thiamine, or vitamin B1, is a relevant ingredient to be used as aroma precursor of meat (Milani and Conti 2024; Milani et al. 2022), helping to solve this issue.

In the light of this, extrusion conditions and the use of lipids are factors that have a great influence on the properties of extruded products. Although this importance is well known, studying their interactions and especially their effects depending on the food matrix used are still relevant and innovative for both academic and food industries, especially considering two food matrices so different in terms of composition and application as an extruded product. Therefore, this study aimed to evaluate the effects of the extrusion conditions, added oil and the food matrix itself on the physical properties and sensory characteristics of corn snacks and a meat analog obtained using pre-extrusion flavoring.

Material and methods

Material

Corn grits (Zanin, Ibiporã, Brasil) and soy protein concentrate (SPC) from Arcon® SM, with a minimum of 70 g protein/100 g in dry basis (d.b.) (supplied by ADM Foods & Wellness, Decatur, USA) were used as the food matrices for extrusion. Cysteine (L-cysteine HCL anhydrous, purity > 98.6%, Lepuge Insumos Farmacêuticos Ltda., São Bernado do Campo, Brazil) and food-grade butyric acid (purity > 99%, Sigma-Aldrich, Milwaukee, USA) were used as cheese aroma precursors. Thiamine (vitamin B1), supplied by Sigma-Aldrich (Milwaukee, USA) as thiamine hydrochloride (purity > 98%), was used as the meat aroma precursor. Salt (Cisne, Cabo Frio, Brazil), monosodium glutamate (Ajinomoto, Limeira, Brazil) and sunflower oil (Liza, Mairinque, Brazil) were purchased at the local market.

Extrusion of food matrices and obtaining products

Extrusion of corn grits was performed according to the method proposed by Menis-Henrique et al. (2020), with adaptations. The moisture of the corn grits was adjusted to 10%, 15% and 20% (d.b.) and butyric acid (0.4 g/100 g) and cysteine (0.2 g/100 g) were added. Corn grits were extruded in an RXPQ Labor 24 single screw extruder (INBRAMAQ, Ribeirão Preto, Brazil) with five independent heating zones: helicoidally grooved barrel; screw with a large step, one exit, with a compression ratio of 3.3:1 and length-to-diameter ratio of 15.5:1; pre-die extruder with holes of 2.7 mm; extruder die with a diameter of 2.9 mm (round hole); feed rate of 265 g/min; screw speed at 192 rpm; temperatures in zones 1–4: 40 °C, 70 °C, 90 °C and 15 °C lower than zone 5, respectively. Three combinations of corn grits moisture (d.b.) and temperature in zone 5 were used: 10% moisture and 140 °C (most severe conditions); 15% moisture and 120 °C (intermediate conditions); 20% moisture and 100 °C (least severe conditions). The intermediate condition results in corn extrudates with adequate physical properties (Menis-Henrique et al. 2019, 2020). The other conditions were determined to evaluate the effects of opposite conditions in terms of severity on the characteristics of the products, as also performed by Conti-Silva et al. (2012).

After extrusion, samples of approximately 5 cm in length were cut using a mold and then salt (1.4 g/100 g of extrudate) and monosodium glutamate (0.6 g/100 g of extrudate) were added (Menis-Henrique et al. 2020). Subsequently, one portion had no sunflower oil added while another one was sprinkled with enough sunflower oil to reach 6 g/100 g of snack (Menis-Henrique et al. 2019). Due to oil losses, 8.28 g of sunflower oil was sprinkled on 100 g of extrudate.

Extrusion of SPC was performed according to the method proposed by Milani and Conti (2024), with adaptations. SPC moisture was adjusted to 30%, 34% and 38% (d.b.) and thiamine (1.5 g/100 g) was added. SPC was extruded in the same extruder used for the corn grits: pre-die extruder with holes of 5.8 mm; extruder die with a diameter of 3.6 mm (round hole); feed rate of 170 g/min; screw speed at 216 rpm; temperatures in zones 1 to 4: 40 °C, 60 °C, 80 °C and 15 °C lower than zone 5, respectively. Three combinations of SPC moisture (d.b.) and temperature in zone 5 were used: 30% moisture and 180 °C (most severe conditions); 34% moisture and 160 °C (intermediate conditions); 38% moisture and 140 °C (less severe conditions). The intermediate condition was previously determined as optimal for the extrusion of SPC added with thiamine (Milani et al. 2022) and the other conditions were determined with the same purpose for corn grits, i.e., to evaluate the effects of opposite conditions in terms of severity on the characteristics of the products.

After extrusion to obtain the meat analog samples, the extrudates were crushed in small portions, in pulse mode for 10 s in a blender, hydrated in preheated water at 100 °C at a solid–liquid ratio of 1:4 and kept immersed in water for 15 min without heating (Milani and Conti 2024). Then, samples from each set of extrusion conditions were separated into two groups for finalizing the aromatization: (i) in a heated pan, 100 g of hydrated extrudate, 1.0 g of salt and 0.4 g of monosodium glutamate were mixed for 2 min; (ii) in another heated pan, 100 g of hydrated extrudate, 7.0 g of soy oil, 1.0 g of salt and 0.4 g of monosodium glutamate were mixed for 2 min. The amounts of salt and glutamate were those used by Milani and Conti (2024).

Analysis of physical properties of corn snacks and meat analog

Expansion ratio, density, instrumental texture and color of the corn snacks were evaluated. The expansion ratio was determined by the ratio between the diameter of the snack and the diameter of the extruder die, with ten random measurements. Density was determined from ten random measurements of the diameter (D, cm) and length (L, cm) of the snacks using an IP54 digital caliper (Digimess, São Paulo, Brazil) and samples were also weighed (W, g). Density (ρ, g/cm3) was calculated using the following equation: ρ = 4W/πD2L. The force required to completely break the snacks was determined using a TA.XT/Plus/50 texture analyzer (Stable Micro Systems, Godalming, UK) and Texture Exponent 32 software (Stable Micro Systems, Godalming, UK), using a blade set with guillotine probe. Ten samples were cut perpendicularly using a pre-test speed of 2 mm/s and test speed of 1 mm/s and the acoustic signals generated by the breakdown of the snacks were picked up using the Acoustic Envelope Detector (Stable Micro Systems, Godalming, UK) coupled to the texture analyzer and with the Exponent 32 software. The acoustic analysis conditions were the same as those described by Dias-Faceto et al. (2020). Force (N) versus time (s) and sound pressure level (dB) versus time (s) curves were plotted simultaneously and data collected were the peak force (Paula and Conti-Silva 2014) and sound pressure level (Dias-Faceto et al. 2020). Color analysis was performed using a Color Flex colorimeter (HunterLab, Reston, USA) with illuminant D65 and a 10° observer angle, as described by Menis-Henrique et al. (2020). The system CIE-L*C*h (L* = luminosity; C* = chroma; h = hue) was used.

Instrumental texture of five replicates of the meat analogs was determined using the same texture analyzer with a cylindrical acrylic probe of 45 mm diameter. Samples from each set of conditions, immediately after their preparation and maintained at a temperature in the range 46–50 vC, were added to a depth of 20 mm in a round plastic container of 50 mm-height and 45.6 mm-internal diameter. The analysis was performed using a pre-test speed of 2 mm/s, a test speed of 1.5 mm/s and 50% compression (Milani and Conti 2024). The peak force, in Newtons (N), was recorded. Color analysis was performed in the same way as described for corn snacks.

Sensory analyses of corn snacks and meat analog

Sensory analyses were performed at the Sensory Analysis Laboratory of the institute. Tests were performed in complete block (one session for each product and six samples at each session), using individual booths at 22 °C and under white light. The snacks were presented in napkins and the meat analog in plastic cups, all coded with three random digits. Samples were balanced (Macfie et al. 1989) and presented in a monadic manner. A glass with water at room temperature was given to consumers to drink between samples.

Corn snacks and meat analogs were submitted to (i) a RATA (Rate-All-That-Apply) test with four categories (not applicable, low intensity, medium intensity and high intensity) (Meyners et al. 2016) and (ii) an acceptance test using the nine-point hedonic scale (9 = liked extremely, 5 = neither liked nor disliked, 1 = disliked extremely) for appearance, odor, texture, flavor and overall acceptance (Meilgaard et al. 2007) in the same session. The sensory attributes for RATA of corn snacks were those studied by Menis-Henrique et al. (2020), with the addition of ‘adhesive’ and ‘burnt flavor’ attributes. For RATA of the meat analog, attributes were the same as those studied by Milani and Conti (2024). For a better understanding of terms, ‘odor’ refers to orthonasal olfactory, while ‘flavor’ refers to aroma (retronasal olfactory). RATA attributes were balanced in the evaluation forms for each sample and consumer (Macfie et al. 1989).

Eighty-two consumers evaluated corn snacks. They were aged from 18 to 54 years old (86% from 18 to 30 years old), 55% female, that like cheese-flavored snacks a little (35%) or very much (65%). Regarding the frequency of consumption of cheese-flavored snacks, 24% consume at least once a week, 28% consume fortnightly and 48% consume once a month.

Sixty consumers evaluated meat analogs. They were aged from 18 to over 54 years old (70% from 18 to 25 years old), 58% female, that like soy-based foods a little (52%) or very much (48%). Regarding frequency of consumption of soy-based foods, 3% consume daily, 17% twice a week, 7% once a week, 25% fortnightly and 48% at least once a month. The soy-based foods consumed were texturized soy protein (63%), soy milk (33%), soy juice (22%), tofu (17%), sauces (5%), hamburger (3%) and snacks (2%). Moreover, 87% of them consume meats and 97% consider a non-meat product with a meat odor and flavor attractive.

Statistical analyses

Physical properties data were submitted to the General Linear Model having extrusion conditions, oil addition and their interaction as factors. The same model was applied to sensory data, including consumers as factor, but without interactions between consumers and the other factors. All analyses were followed by the Tukey test when pertinent and performed using a significance level of 0.05, using the XLSTAT software for Microsoft Excel (Addinsoft, USA).

Principal component analysis was performed to correlate the results. For this, the mean variables were entered in the columns and the samples in the rows and data was standardized in the columns before analysis. The extraction of factors was performed from the correlation matrix, without rotation of factors. This analysis was performed using the Statistica 7.0 software (StatSoft Inc., USA).

Results and discussion

Physical properties and sensory characteristics of corn snacks

Corn snacks showed adequate expansion ratios, densities and peak force (Supplementary Table 1) like those of a previous study performed with the same aroma precursors (Menis-Henrique et al. 2020). All the physical properties of the corn snacks were affected by the extrusion conditions, while expansion ratio, density and sound pressure level were not influenced by the addition of oil neither by the interaction between factors (Supplementary Table 2). Corn snacks from 15% M/120 °C had the highest expansion ratio (result not shown), an expected result since such conditions are used by the research group as optimal. Corn snacks from 20% M/100 °C had the highest density (result not shown). When using high moisture and low temperature, water vaporization during and at the end of processing is compromised, reducing the expansion of the products, increasing their density (Leonard et al. 2020). The peak force was higher for the snacks from 20% M/100 °C and with oil, even in relation to the same extrusion conditions but without oil (result not shown), results that follow the highest density of this sample. The snacks from 10% M/140 °C had lower sound pressure level when compared to the other conditions. In more severe extrusion conditions, there is a greater shearing of the mass inside the extruder, resulting in a product with a more fragile structure, reducing the intensity of the sound generated during the rupture of the product (Conti-Silva et al. 2012; Leonard et al. 2020).

Snacks had a yellow color, since their hues were between 70º and 100º (Ramos and Gomide 2017). Nevertheless, luminosity, chroma and hue increased from the extrusion conditions 10% M/140 °C to 20% M/100 °C (results not shown). Color is affected by the application of high temperatures, causing pigment changes due to pigment degradation, the Maillard reaction and caramelization (Leonard et al. 2020; Yi et al. 2022). The effect of the oil was different depending on the color attribute, reducing the values of luminosity and hue, and increasing chroma (result not shown). The oil acts as a 'cover' over the snack, reducing the lightness; in addition, it has a slightly yellowish color, contributing to a higher chroma (Menis-Henrique et al. 2020).

Intensities of attributes ranged from 0.1 (burnt flavor) to 2.6 (hardness) on the scale from 0 (not applicable) to 3 (high intensity), values for snacks from 20% M/100 °C (Table 1). The acceptance ranged from 3.5 (flavor) to 7.4 (appearance) in the nine-point hedonic scale (Table 1). All the sensory characteristics were affected by the conditions studied (Supplementary Table 3). The corn snacks from 10% M/140 °C had the lowest intensity of crispness (Fig. 1A) and the highest intensity of adhesiveness (Fig. 1B) when compared to the other snacks, as well as higher intensity of oil odor (Fig. 1C) when compared to the snacks from 15% M/120 °C. Similar results to crispness were found for the yellow color, hardness and cheese flavor, while oil odor, burnt flavor and oil flavor had the same behavior as adhesiveness (graphs not shown). The use of low moisture and high temperature led to a higher shearing of the dough during extrusion, resulting in snacks with a color driven to brown (less yellow color), with stronger burnt flavor and a more fragile structure (less crispness and hardness) (Conti-Silva et al. 2012; Leonard et al. 2020). Concerning the cheese flavor, the lowest intensity for snack 10% M/140 °C is probably due to the highest burnt flavor reducing the perception of the cheese flavor.

Table 1 Sensory characteristics of the corn snacks (mean ± SD; n = 82)
Full size table
Fig. 1
figure 1

Attribute intensities and sensory acceptance of the corn snacks. Values are represented as mean ± 95% confidence interval (n = 164 for A, B, C, F and G; n = 82 for D and E; and n = 246 for H). Legend: M = moisture of the corn grits (dry basis) before extrusion. Temperature at zone 5 of the extruder barrel. Different letters indicate significantly different means by the Tukey test (p ≤ 0.05)

Full size image

Regarding the interactions between extrusion conditions and oil, the addition of oil increased the intensity of the cheese odor only for the snack from 20% M/100 °C (Fig. 1D), while it increased the intensity of the salty taste in all samples (Fig. 1E; identical results for the umami taste). The increase in the intensities of the salty and umami tastes in all samples after adding oil is due to the importance of lipids in taste perceptions. Moreover, the intensities of cheese odor, salty taste and umami taste were higher for 20% M/100 °C when compared to 10% M/140 °C only when oil was added (Fig. 1D, E), results that may be attributed to the structure of products that allowed the oil to penetrate.

In general, the acceptance of appearance, odor and flavor, as well as overall acceptance, were higher for snacks from 15% M/120 °C (Fig. 1G for flavor). The snack obtained under 20% M/100 °C had lower acceptance of texture (Fig. 1F) because those samples showed higher density that affects the hardness of the products. All attributes had higher acceptance when oil was added, as well as the overall acceptance (Fig. 1H). The addition of oil also enhanced the yellow color, oil odor and cheese flavor while it decreased the cereal flavor and burnt flavor (graphs not shown). Regarding the cheese flavor, the presence of lipids leads to a delay in the release of lipophilic volatile compounds (Menis-Henrique et al. 2019), thus the release of volatile compounds lasted longer, resulting in a higher perception of the cheese flavor intensity, decreasing the intensities of cereal and burnt flavor intensities and increasing the acceptance.

Physical properties and sensory characteristics of meat analog

Meat analog showed low peak force (Supplementary Table 4), which was expected because it was milled and hydrated. Although this property was not affected by any of the conditions studied, all color attributes were affected (Supplementary Table 5). The meat analog from 38% M/140 °C (with and without oil) had the highest values for luminosity and hue, while those from 30% M/180 °C, with and without oil, presented the highest chroma values. High temperatures cause changes in color due to the Maillard reaction and caramelization, in addition to producing reddish-brown colors for the most severe conditions (Leonard et al. 2020; Yi et al. 2022). Although there are differences, all the meat analog samples may be indicated as having orange in the solid color, since their hues were between 25º and 70º (Ramos and Gomide 2017).

Intensities of attributes ranged from 0.3 (white color) to 2.5 (caramel color and wet texture) in the scale from 0 (not applicable) to 3 (high intensity), while acceptance ranged from 4.6 (texture) to 6.7 (texture) on the nine-point hedonic scale, for the meat analog samples from 30% M/180 °C without and with soy oil, respectively (Table 2). In contrast to the corn snacks, not all attributes were affected by the conditions and no interaction between factors was found (Supplementary Table 6). Many of the attributes not affected by the factors are related to volatile compounds. According to Yuliani et al. (2004), aromatization using pre-extrusion may present losses of volatile compounds or even the development of others due to the influence of high process temperatures, resulting in oxidation and thermal degradation, especially in the case of thiamine.

Table 2 Sensory characteristics of the meat analogs (mean ± SD; n = 60)
Full size table

Meat analog from 38% M/140 °C showed the highest intensity of white color (Fig. 2A), as well as a rubbery texture (graph not shown), while those from 30% M/180 °C exhibited the highest intensities for wet texture (Fig. 2B), caramel color, crumbly/sandy texture and burnt aftertaste (graphs not shown). The addition of oil masked the rubbery texture (Fig. 2C) and the residual burnt flavor (graph not shown), while promoting greater intensities for the seasoning flavor (Fig. 2D) and the aromatic and tasty attributes (graphs not shown). Again, the primordial role of the lipid vehicle is verified, retaining the volatile compounds and reducing the perception of undesirable flavors and textures (He et al. 2022). The acceptance of odor (Fig. 2E), texture, flavor and overall acceptance (graphs not shown) was higher when vegetable oil was used, validating the results found by RATA. Regarding the extrusion conditions, consumers showed lower acceptance for texture of the meat analog from 30% M/180 °C (Fig. 2F), an expected effect, since this sample presented higher intensities for wet and crumbly/sandy texture, although the 38% M/140 °C conditions presented the highest intensity for rubbery texture. All these results explain the lowest overall acceptance for the meat analog from 30% M/180 °C when compared to the other two sets of conditions (Fig. 2G).

Fig. 2
figure 2

Characteristics of sensory profile of the meat analogs. Values are represented as mean ± 95% confidence interval (n = 120 for A, B and C; n = 180 for D and E). Legend: M = moisture of the corn grits (dry basis) before extrusion. Temperature at zone 5 of the extruder barrel. Different letters indicate significantly different means by the Tukey test (p ≤ 0.05)

Full size image

Effects of food matrices on physical and sensory characteristics of products

Figure 3 shows the correlations between physical properties and sensory characteristics for corn snacks and for meat analog. For corn snacks (Fig. 3A, B), three groups of samples are identified: (1) one of corn snacks from 10% M/140 °C, regardless of the addition of oil or not, (2) samples without oil, regardless of extrusion conditions and (3) samples with oil, also independent of extrusion conditions. Samples from the 10% M/140 °C conditions, with and without oil, were described by adhesiveness and burnt flavor. Corn snacks from 20% M/100 °C without oil were described by density, peak force, all color instrumental properties, yellow color, hardness, cereal flavor and acceptance of appearance, although cereal flavor did not correlate with the other variables. The other snacks with oil were described by sound pressure level, cheese odor, crispness, salty taste, umami taste, cheese flavor, acceptance of odor and flavor, although salty taste and umami do not correlate with the other variables.

Fig. 3
figure 3

Correlations between physical properties and sensory characteristics of corn snacks (A – projection of variables; B – projection of samples) and the meat analogs (C – projection of variable; D – projection of samples). Graph A (1- expansion ratio, 2- density, 3- peak force, 4- sound pressure level, 5- luminosity, 6- chroma, 7- hue, 8- yellow color, 9- cheese odor, 10- oil odor, 11- hardness, 12- crispness, 13- adhesiveness, 14- salty taste, 15- umami taste, 16- cereal flavor, 17- cheese flavor, 18- burnt flavor, 19- oil flavor, 20- acceptance of appearance, 21- acceptance of odor, 22- acceptance of texture, 23- acceptance of flavor, 24- overall acceptance). Graph C (1- peak force, 2- luminosity, 3- chroma, 4- hue, 5- caramel color, 6- white color, 7- uniform granules, 8- meat flavor, 9- soy odor, 10- wet texture, 11- crumbly/sandy texture seasoning, 12- rubber texture, 13- salty taste, 14- meat odor, 15- noodle flavor, 16-seasoning flavor, 17- burnt aftertaste, 18- aromatic, 19- tasty, 20- acceptance of appearance, 21- acceptance of odor, 22- acceptance of texture, 23- acceptance of flavor, 24- overall acceptance). Graphs B and D (M = moisture of the corn grits in dry basis before extrusion. Temperature at zone 5 of the extruder barrel)

Full size image

For the meat analogs, in the same way as for the corn snacks, three groups of samples are identified (Fig. 3C, D): (1) those from 30% M/180 °C, regardless of the addition of oil, (2) samples without oil, regardless of extrusion conditions and (3) samples with oil, regardless of extrusion conditions. Samples from 30% M/180 °C without oil were described by caramel color, chroma, crumbly/sandy texture and burnt aftertaste; on the other hand, samples from these conditions and with oil were described by wet texture and acceptance of appearance, although these did not correlate to each other. The meat analog from 38% M/140 °C and without oil was described by white color, luminosity, rubbery texture and noodle flavor, in addition to uniform granules and soy odor, although the latter two do not correlate with the other variables. Finally, samples from 34% M/160 °C and 38% M/140 °C with oil were described by meat flavor, seasoning flavor, tasty, acceptance of odor, texture and flavor and overall acceptance, although flavor acceptance and overall acceptance do not correlate with the other variables.

As already described, there is a similarity in the separation of samples, independent of the food matrix used. Samples were separated regarding the most severe conditions, i.e., 10% M/140 °C and 30% M/180 °C for corn snacks and the meat analog, respectively, and then in function of adding oil. Moreover, for sensory acceptance, both corn snacks and the meat analog from intermediate conditions (15% M/120 °C and 34% M/160 °C for corn snacks and the meat analog, respectively) and less severe conditions (20% M/100 °C and 38% M/140 °C for corn snacks and the meat analog, respectively) and with added oil stood out in terms of acceptance of odor and flavor, while acceptance of texture and overall acceptance described the meat analog only.

Physicochemical and structural properties of the food matrix, as well as chemical properties of aroma compounds, are considered the main factors affecting aroma release (Ployon et al. 2017), and then flavor. Moreover, food oral processing depends on several dimensions, being food matrices (liquid, solid, degree of structure) an important one (Devezeaux De Lavergne et al. 2021). Studies show that different food matrix structures influence the perception of texture and flavor (Matsuyama et al. 2021; Tournier et al. 2017; Wilson and Brown 1997). By contrast to the literature, the present study shows little effect of food matrix (corn grits or soy protein concentrate), being the results driven mainly by the extrusion conditions and the addition or not of oil.

Conclusion

Extrusion conditions affected all physical properties and most of the sensory characteristics of corn snacks, while they had less influence on the properties of the meat analogs. The same behavior (more influence for corn snacks than for the meat analogs) was found for the addition of vegetable oil, as well as the interactions between extrusion conditions and addition of oil. Corn snacks from 10% M/140 °C were adhesive and had a burnt flavor and a good texture acceptance. Snacks from 15% M/120 °C had a higher expansion ratio and stood out for their sound pressure level and crispness, in addition to the acceptance of texture and flavor. Snacks from 20% M/100 °C had higher density, peak force and color intensities, as well as stood out in terms of sound pressure level, crispness, hardness and lower acceptance of texture. The addition of oil intensified those attributes related to flavor. Meat analogs from 30% M/180 °C had greater intensity for caramel color, wet texture, rubber texture and burnt aftertaste, while those from 38% M/140 °C showed greater intensity of white color. The addition of oil improved the sensory acceptance, although had few effects on the intensities of sensory attributes. When studying samples regarding correlations between physical and sensory variables, there was a similarity in the discrimination of samples, independent of the food matrix used. Samples were separated regarding either the most severe conditions or the addition of oil, especially for the sensory acceptance, for which the samples from intermediate and less severe conditions and with added oil stood out. In conclusion, the extrusion conditions and the addition of oil have an impact on the physical properties and sensory characteristics, with little effect from the matrix itself.