Consumer perception and sensory properties of bakery products fortified with chicken protein for older adults

Loss of appetite is common in older adults and as an adequate protein intake is recommended to support the maintenance of muscle mass and strength during ageing, a sufficient intake of dietary protein is essential. One strategy is to develop protein fortified foods for older adults. This study developed rye bread and chocolate cake fortified with chicken protein powder. Two independent trained panels were recruited to evaluate the sensory properties of the products. Seventy-five Danish older adults aged between 60 and 83 years were recruited to rate their overall liking on the 9-point hedonic scale; and their emotional response to each product was collected by check-all-that-apply (CATA). Results showed that fortification with chicken protein powder caused sensory changes especially in texture such as less firm , elastic texture and stronger floury mouthfeel for rye bread; less spongy, crumbly and stronger dense, moist texture for chocolate cakes. The sensory changes resulted in small yet significantly lower consumer acceptance of rye bread, but no significant differences in liking of chocolate cakes were found. All samples were above the acceptability level by Danish older consumers. Emotions pleased, interested, sad, disappointed and disgusted significantly discriminated the rye bread samples; disappointed and bored emotions significantly discriminated the chocolate cake samples. Moreover, significant correlations were found between the familiarity with protein fortified food products, appropriateness of rye bread or chocolate cakes for protein fortification and willingness to purchase of protein-fortified rye bread or chocolate cakes. This study was a first application of chicken protein powder in fortification of cereal products for older consumers.


Introduction
Between 2015 and 2050, the proportion of the world's population over 60 years will nearly double from 12% to 22% (WHO, 2021). The relative loss of muscle mass, the so called sarcopenia, is a highly prevalent clinical feature among older adults. Sarcopenia is partly due to a declining anabolic response to protein intake and to the offset of inflammatory and catabolic conditions associated with chronic and acute diseases that occur commonly with aging (Keller, 1993;Paddon-Jones and Rasmussen, 2009) and has been associated with an increased risk of falls, reduced functionality, and even higher mortality. Adequate dietary protein intake, next to exercise, has been shown to play a positive role in preventing and managing sarcopenia. While it has been shown that older persons have less appetite and cannot easily increase the portion sizes of their meals, replacing regular products by protein enriched substitutes may be a viable solution to increase their protein intake (Stelten et al., 2015).
Previous research has been conducted to develop fortified foods by utilizing different protein sources such as whey protein and soy protein.
When adding whey protein powder into muffin, negative sensory properties occurred such as high astringency mouthfeel and very dry texture (Höglund, 2017;Wendin et al., 2017). Friedeck, Karagul-Yuceer and Drake (2003) found that soy protein isolate contributed to green/grassy and doughy/fatty flavours, albeit it did not change the texture acceptance. Song, Perez-Cueto and Bredie (2018) developed rye bread and cream cheese prototypes fortified with whey protein hydrolysate, whey protein isolate, and/or soy protein isolate, results showing that different proteins had various effects on the sensory characters of the products. Meat proteins, on the other hand, have received less attention in the literature, despite having high nutritional value and application potential.
A few studies have used meat protein in food protein fortification (Farouk et al., 2018;Kessler et al., 2019;Meinert et al., 2016). Moreover, all of them have utilized protein from livestock meat but no study has reported on poultry meat. Currently chickens are produced from slaughterhouse where some of the remaining parts of the meat are used for pet food. Exploiting the side streams for human consumption may be a future contribution to a more sustainable poultry production. However, this requires understanding of acceptability of chicken protein applications in food production. In food development, assessment of consumer acceptance and preference has been widely used in order to measure consumers' hedonic response of products (Liu et al., 2021). A recent study investigated the consumer valuations of carbon labelled protein-enriched burgers in European older adults and provided insight into older adults' relative preferences, intentions, and willingness to pay for protein-enriched burgers (Broeckhoven et al., 2020). Several other latest studies have worked on consumer acceptance of fortified foods with alternative protein sources and confirmed that it is important to consider differences in consumer motivation (Grahl et al., 2020;Saint-Eve et al., 2021). In addition, the interest in measuring emotional responses to foods and beverages has been growing rapidly during the last decade. Previous studies suggest that the measurement of emotions to foods and beverages give new information beyond classic food acceptance testing (King and Meiselman, 2010;Thomson, 2008). The findings by Yang et al. (2020) in food product fortified with alternative sustainable ingredient highlighted the value of measuring emotional response for novel products to understand what may drive purchase behaviour when products are matched for liking. Therefore, combining the measurement of consumer acceptances and emotions can be a valuable tool to support product development.
The aim of the study was to investigate the consumer acceptance and emotion responses to bakery products fortified with chicken protein powder by Danish older adults. Two bakery product categories, rye bread and chocolate cakes were selected. Rye bread is one of the most commonly consumed staple foods on a daily basis primarily among Danish adults (Song et al., 2018). Chocolate cake is more usually consumed after meal or between meals.

Experiment design
The study was performed at the Future Consumer Lab of University of Copenhagen. The tests consisted of sensory descriptive analyses of rye bread and chocolate cake by two independent trained panels as well as a consumer test by older adults. All samples of rye bread and chocolate cakes were freshly baked the day before each test day. The present study was performed according to the principles established by the Declaration of Helsinki. The study protocol was approved by the Research Ethics Committee of University of Copenhagen (Case no: 5040140/20-5000). Participation of all subjects was by informed consent.

Ingredients and preparation of rye bread
The ingredients used in the preparation of rye bread were: sunflower rye bread mix (Valsemøllen, Denmark), dry yeast (Malteserkors, Denmark), sunflower seeds (Valsemøllen, Denmark), tap water, and one sample contained wheat gluten (Nutty vegan, Denmark). The added chicken protein powder was purchased from Henningsen, the Netherlands (product no: 26,017 PC chicken C-200K AR). The protein was produced by spray drying cooked mechanically separated chicken meat in an approved factory, and it contained chicken meat (99.9% w/ w) and antioxidant (extract of rosemary, 0.1%w/w). The chemical and nutritional contents of the chicken protein were (w/w): protein 53%, fat 40%, moisture 5%, salt (natural presence) 2%.
Five different formulations were prepared in which a gradual increase in the amount of chicken protein powder was added from 0 to 25 w/w% of the total dry ingredients along with a sample containing gluten in order to stabilise the matrix ( Table 1). All of the dry ingredients were mixed with water (35 ± 0.6 • C) using a K-spade at speed 2 for 5 min in a Kenwood (Chef XL, model KVL40). The dough was placed in a 31 × 8.25 × 8 cm cm rectangular form and rest for 60 min at room temperature (23.5 • C, ± 0.3 • C) before baking for 60 min in a preheated oven (Rational, model SCC WE 101) at 170 • C with hot air function. After cooling upside down, the bread were stored at 21 • C overnight in a plastic bag until the evaluation on the following day.

Ingredients and preparation of chocolate cakes
The ingredients used in the preparation of chocolate cakes were: butter (Lurpak, Denmark), sugar (Dansugar, Denmark), egg (Danaeg, Denmark), wheat flour (Amo, Denmark), cocoa powder (Gestus, Denmark), baking powder (Dr. Oetker, Denmark), baking soda (Dr. Oetker, Denmark), salt (Salina, Denmark), and one sample contained wheat gluten (Nutty vegan). The same chicken protein powder was used in the same doses as described for the bread. All forms were sprayed with a baking spray (DECOCINO).
Five different formulations were prepared using the creaming mixing method (Gisslen, 2004) using a Kenwood (Chef XL, model KVL40) mixer. The day before the baking the protein powder were mixed with the water according the recipe (Table 2), using a handhold blender (bamix, model G200) at maximum speed for 5 min, to prevent lumps. The day of the baking, the dry ingredients including cocoa powder, baking powder, baking soda, salt, flour and gluten were mixed ( Table 2). The butter and sugar were whisked at speed 4 for 5 min; eggs were added and whisked at speed 2 for 1 min. Then the dry and the wet (protein and water) ingredients were partially added twice and whisked at speed 1 for 10 s and then at speed 3 for 5 s. The cake batter was placed in a 24 × 10.5 × 5.6 cm rectangular silicone form and baked for 35 min in a preheated oven (Rational, model SCC WE 101) at 170 • C with hot air function. The cakes were allowed to cool and stored in the silicone form, covered with a plastic bag, at 21 • C overnight until the evaluation.

Sensory analysis of rye bread
A sensory panel consisting of 10 assessors (4 females and 6 males; mean age = 26.2 ± 4.5) was used for sensory profiling of rye bread. The assessors were recruited from the external trained sensory panel at the Future Consumer Lab and they received regular training for descriptive analysis. All assessors were compensated for their participation.
The assessors participated in three 2-h training sessions prior to the sensory evaluation. During the first training session, the panel was instructed to generate their own attributes and later supplemented with a list of potential attributes. In subsequent sessions, the panellists were provided by the panel leader with some reference standards. The assessors selected the attributes that were most suitable for describing the sensory differences among the samples and for which it was possible to use actual physical objects as reference standards or at least a precise written definition. The final set of attributes along with their reference materials are reported in Table 3. After the training phase, the assessors participated in the final evaluation in triplicates (approximately 45 min each). All of the evaluations were conducted in standardised individual sensory booths at room temperature (20 ± 1 • C). Each sample was served in plastic cup, blind labelled with three-digit numbers, and covered with plastic lids. To minimize carry-over effects, samples were presented in a randomized order to the panel. The assessors were asked to rate each attribute for all the samples in a 15-cm unstructured line scale using FIZZ software (Biosystems, Courtenon, France). Water (both still and sparking) and milk were provided for palate cleansing between samples. There were two 10min breaks between each session to limit fatigue.

Sensory analysis of chocolate cakes
A second panel consisting of nine assessors (4 females and 5 males; mean age = 25.4 ± 2.9) was recruited for sensory profiling of chocolate cakes. A similar procedure was performed as for rye bread and the final sensory attributes along with definitions and reference standards are shown in Table 4.

Consumer test
Seventy-five older Danish adults (52 females and 23 males; aged between 60 and 83 years; mean age ± SD = 71.0 ± 5.3 years; mean BMI ± SD = 26.1 ± 5.0) were recruited from the external consumer database at University of Copenhagen. They were independently home-living and active healthy individuals. The demographic details of participants are displayed in Table 5. All participants were compensated for their participation.   Degree of coating in the mouth after mastication of a food After-taste Sweet After taste quality characteristic for sucrose (24 g/L in water) Salty After taste quality characteristic for sodium chloride (2 g/L in water) Bitter After taste quality characteristic for caffeine (0.54 g/ L in water) The consumer test was conducted in individual sensory booths, in which two sessions were included: rye bread tasting and chocolate cake tasting with a 15-min break in between. During the tasting, participants were required to fill out questionnaires for rye bread and chocolate cakes, respectively. The first part of the questionnaire enquired about the overall liking using the 9-point hedonic scale from 1-"dislike extremely" to 9-"like extremely" (Peryam and Girardot, 1952). Next, participants were asked about their willingness to purchase the product if it was introduced to the market on a 7-point scale from 1-"not at all" to 7-"definitely". Participants were then asked to express their emotional responses to the product using the check-all-that-apply (CATA) method (Reinbach et al., 2014). A total of 12 emotional terms were included: happy, desired, satisfied, pleased, interested, calm, neutral, sad, disappointed, disgusted, bored, and guilty. These emotion terms were preselected from previous studies (King et al., 2013;Orjuela-Palacio et al., 2014;Poonnakasem et al., 2016;Wardy et al., 2018), and decided through research group discussions. In the end of the test, participants were asked to fill out a questionnaire with questions including basic demographics, consumption frequency of rye bread and chocolate cakes as well as their attitudes towards protein-enriched foods. The questionnaire was translated back and forth between English and Danish and provided with a Danish version to the participants.

Data analysis
The sensory descriptive data were analysed by analysis of variance (ANOVA) using a mixed model with product and replicate as fix factors and assessor as random. Where significant product effects were found, the ANOVA was followed by post-hoc comparison using Tukey's Honest Significant Difference (HSD) test (p < 0.05). To enable a visual exploration of the sensory results, the data with significant sensory attributes were analysed by Canonical Variate Analysis (CVA) under discriminative analysis with canonical correlation functions. The ellipses show 95% confidence intervals in each direction of variation.
Agglomerative hierarchical clustering (AHC) analysis was performed to identify homogeneous clusters of consumers with similar acceptance of rye bread or chocolate cakes, respectively. One-way ANOVA was conducted on consumer liking data followed by the Tukey HSD test if applicable. External preference mapping (PREFMAP) was conducted to investigate relationships among consumer acceptance and sensory attributes across rye bread and chocolate cakes, respectively. The CATA responses where rendered as dichotomous data where a value of 1 indicated that an attribute had been checked and a value of 0 indicated the opposite. Differences between samples with respect to frequency of mention on each individual CATA attribute were assessed using Cochran's Q Test, as customary for this type of data (Meyners et al., 2013). Besides, the Pearson's correlation between the familiarity, appropriateness and willingness to purchase protein fortified product were calculated. All data analyses were performed using XLSTAT (version 2020.1.2, Addinsoft, France).

Sensory analysis of rye bread
The results of the ANOVA analyses on the sensory descriptive data of rye bread are shown in Table 6. All attributes except hay, sour, salty taste in crumb and malty in crust were found to significantly (p < 0.05) discriminate between the samples. Fig. 1 shows the CVA plots of samples projection and significant sensory attributes of rye bread, accounting for 98.3% of the total variance (87.9% and 10.4% for PC1 and PC2, respectively). As observed in Fig. 1, the samples spanned mainly the first dimension, contrasting the sample RB0 (control) and RB10 (addition of 10% protein) with sensory attributes crumbly, floury to samples with addition of 20% and above protein with firm, elastic and chewy. Specifically, addition of more than 10% chicken protein resulted in significantly less firm, elastic, chewy texture as well as stronger crumbly texture and floury mouthfeel ( Table 6). Addition of more than 20% protein caused significantly lighter brown crumb, small but significantly higher bitterness and lower sweetness (Table 6).

Sensory analysis of chocolate cakes
The results of the ANOVA analyses on the sensory descriptive data of chocolate cakes are showed in Table 7. All attributes except nutty and sweet were found to significantly (p < 0.05) discriminate between the samples. Fig. 2 shows the CVA plots of samples projection and significant sensory attributes of chocolate cakes, accounting for 96.0% of the total variance; 91.8% in PC1 and 4.3% in PC2, respectively. The samples thus spanned mainly in the first dimension, contrasting the sample CC0 (control) and CC10 (addition of 10% protein) with sensory attributes spongy, crumbly and soft to samples with addition of 20% and above protein with dense, moist and oily (Fig. 2). Specifically, addition of more than 20% chicken protein resulted in significantly lower cacao but higher chocolate odour and flavour; higher darkness, dense and moist appearance; less spongy, soft, crumbly but higher dense, moist, oily texture, as well as stronger sticky, floury, and mouth coating mouthfeel (Table 7).

Consumer acceptance to rye bread and chocolate cakes
Four samples of rye bread and chocolate cakes, respectively, were included in the consumer test. Mean hedonic scores of all samples by older adults are displayed in Table 8. Consumers who were homogenous in their acceptance towards different samples were grouped through agglomerative hierarchical clustering (AHC). For rye bread, reduced liking was found in the sample RB10 (addition of 10% chicken protein powder) compared to the control sample RB0, but the difference was not significant. The higher addition of chicken protein (RB25) resulted in significantly lower liking. Combining with the external preference mapping (Fig. 3a), it shows that three consumer clusters were identified. Some 47% of consumers (cluster 1) showed significantly higher liking towards the control sample which was characterised with elastic, firm and chewy texture. Cluster 2 (29%) of consumers showed the equally high liking towards the RB10 and the RB0. Cluster 3 (24%) displayed the highest liking to RBGLU, which was added with gluten and exhibited with dry texture and bitter taste.
For chocolate cakes (Table 8 & Fig. 3b), no significant differences were observed in the overall liking of all four samples. Similarly, three consumer clusters were identified. In cluster 1 (19%), the sample CCGLU was mostly liked, followed by the sample CC10. Cluster 2 (43%) of consumers equally liked the sample CC25 and CCGLU, which were linked with most sensory attributes such as moist, oily and dense texture. Another 39% of consumers (cluster 3) showed higher liking to the sample CC0 and CC10, which had strong cocoa powder flavour and odour, spongy and soft texture. Fig. 4 displays the total number of elicitations for rye bread and chocolate cakes by older adults after eating, and significance stars are given to those significant emotions at p < 0.05 level. For rye bread (Fig. 4a), satisfied, pleased, interested, calm, neutral and disappointed were mostly cited emotions by older adults. Emotions pleased, interested, sad, disappointed and disgusted significantly discriminated the samples. The sample RB0 evoked the strongest pleased emotion than other samples, and the sample RB25 provided more disappointed elicitation. Other large differences can be observed for interested, sad and disgusted. Regarding the chocolate cakes (Fig. 4b), happy, satisfied, pleased, interested, calm, neutral and disappointed were mostly cited emotions. Emotions disappointed and bored significantly discriminated the samples. Surprisely, the sample CC0 was considered as a significantly more bored sample compared to all fortified samples. Addition of protein evoked more positive emotions such as happy, desired, and pleased elicitations, however, these differences were not significant among cake samples.

Consumes attitudes towards fortified foods
For rye bread, significantly positive relationships were found in all correlations between the familiarity with protein fortified food products, appropriateness of rye bread for protein enrichment and willingness to purchase of protein fortified rye bread (Fig. 5). Particularly the correlation between the latter two measures was relatively strong (Pearson's correlation coefficient = 0.724, p < 0.001). Regarding chocolate cakes, the correlations between different measures were also positive, nevertheless, the correlation between the familiarity and the willingness to purchase was weak and not significant (Pearson's correlation coefficient = 0.143, p > 0.5). The positive correlations indicate that increasing the familiarity and knowledge to protein fortified product for  older adults could be a feasible way to increase their purchase intension of related protein enriched products.

Discussion
This study has evaluated the sensory perception, consumer acceptance and emotional responses of older adults for fortified bakery products. Addition of chicken protein powder led to less elastic, chewy texture for rye bread; less spongy, soft and greater dense, moist texture for chocolate cakes. For both matrices, a stronger sticky, floury mouthfeel was obtained. The sensory changes might be because the chicken protein conferred a protective network on partial gluten structure which increase the dense and sticky texture of the dough and baked products (Song et al., 2018;Ziobro et al., 2013). When the protein addition was above 10%w/w of the total dry ingredients, fine particles were perceived (floury mouthfeel) due to insolubility of the chicken protein powder. This may thus inhibit the application of chicken protein powder   in some products such as soup and sauces. The sensory changes had a negative impact in consumer acceptability of fortified rye bread but not for chocolate cakes. Cocoa seems have a positive masking of chicken protein powder in the bakery products (cocoa and chocolate flavour). Chocolate flavour has been previously reported to have a greater impact on masking soy flavour than vanilla in ice cream (Friedeck et al., 2003). It is worthwhile noting that all samples were acceptable by older consumers (average liking scores above 5 on a scale from 1 to 9), indicating the sensory perceptions were not beyond the accepted level among senior consumers. However, one should bear in mind that in comparison to younger adults, older adults may use a higher part of the 9-point hedonic scale when making their evaluations (Duizer and Field, 2015). The measurement of emotions could help food developers understand consumer enjoyment of food products. For rye bread, the addition of chicken protein resulted in less pleased and interested emotion and stronger sad, disgusted, disappointed emotion. For chocolate cakes, the fortification with chicken protein did not produce many differences between samples on the emotional responses, with only two significant terms (disgusted and bored). When looking at the two product categories in general, chocolate cakes evoked higher positive emotional responses such as happy, desired and pleased emotions than rye bread. This finding is in line with a recent study showing that consuming sweet foods triggered more positive emotions than consuming savoury foods (Yang et al., 2020). It is worth noting that the consumer's evaluation was conducted in individual booths, which would not elicit the same emotion associated with tasting the products at home or other locations. As noted by Liu et al. (2019) and Piqueras-Fiszman and Jaeger (2014), consumers' hedonics and emotional responses to food products can be affected by the consumption context. Future studies could be conducted in a real-life condition that may lead to different results, and such findings would be valuable as a comparison with the present work.
In general, participants were optimistic about the carrier formats, by expressing high perceived appropriateness of rye bread and chocolate cake and willingness to purchase. Furthermore, a strong relationship was obtained between perceived appropriateness and willingness to purchase. This result is in line with previous research which showed that the appropriateness of enriched foods affects consumers' willingness to trial purchase (van der Zanden et al., 2015). In addition, this study showed that familiarity to protein fortified foods had a positive influence to perceived appropriateness and willingness to purchase. Older adults' food behaviour has been reported to be largely habit driven (Doets and Kremer, 2016). A recent study confirmed that older adults' intentions to consume alternative protein-fortified products are associated with familiarity, knowledge of protein and the environmental impact of different protein sources (Broeckhoven et al., 2020). It therefore, seems warranted to increase the familiarity and knowledge towards enriched foods as a modulator of seniors' food perception, liking and intake of such kinds of foods.
Some studies have explored the application of meat-based protein powder from beef and pork in food fortification. Kessler et al. (2019) reported that snack sausages enriched with meat protein hydrolysates had minor or no change on the perceptual quality. In an earlier study by (Meinert et al., 2016), Danish meat product saveloy enriched with meat protein hydrolysates were assessed by the sensory panel as having the   strongest off-flavours (chemical flavour) that were not perceived by consumers. The value of the present work was to exhibit the potential of using protein from poultry in food products. This may reduce the environmental and climate impact of meat production by making use of poultry slaughterhouse side stream in human consumption. Addition of chicken protein powder contributed to an acceptable sensory quality and higher content of nutrients from the chicken, in particular protein.
In future research, combination of different proteins that have different physical and chemical properties could be attempted to develop optimally appealing protein fortified food products targeting specific subgroups. One should also be cautious that this work has been done blindly without informing about the ingredients to consumers before running the test. The acceptance of cereal products fortified with meat-based protein should be evaluated with advance informing in the future research.

Conclusions
This study explored the potential application of new ingredient chicken protein in food products. It contributed to new knowledge regarding the sensory evaluation, consumer acceptance and emotional responses of older adults for protein-fortified bakery products. Addition of chicken protein powder mainly affect texture properties of bakery products, such as less firm, elastic texture and stronger floury mouthfeel for rye bread; less spongy, crumbly and stronger dense, moist texture for chocolate cakes. The sensory changes resulted in slightly lower consumer acceptance of rye bread with 10% addition of chicken protein, and when increasing the addition of chicken protein the liking was reduced; no significant differences were obtained in liking of fortified chocolate cakes. It should be noted that all samples were above the acceptability level by Danish older consumers. This study supports that measuring both consumer acceptance and emotional response could be a valuable tool for developing novel functional products. The study also indicates the potential of using protein powder from poultry slauterhouse sidestream for human consumption. The results provide useful knowledge to the kitchen of elderly homes or hospitals. However, the acceptance of cereal products fortified with meat-based protein should be evaluated with advance informing in the future research.

Implications for gastronomy
This work was a first application of poultry protein powder in fortification of cereal products. The recipes of rye bread and chocolate cakes fortified with different dosages of chicken protein were developed. The sensory properties of products were evaluated with trained panels; the acceptability and emotional responses were evaluated with older adults. The results contribute to the development of functional foods for target consumers and provide useful knowledge to restaurants, kitchens and related organizations.