Making Sensory Tests Instrumental

Making Sensory
Tests Instrumental
May 1998 -- QA/QC

By: Scott L. Hegenbart
Contributing Editor

  The way a food product stimulates the senses directly relates to a consumer's eating pleasure. Every food product designer's goal is finding the combination that creates that sensation for a majority of individuals. To test for these attributes, designers often use such tools as sensory evaluation and consumer-acceptance testing. Unfortunately, such procedures aren't known for providing rapid results, and often carry a hefty price tag. Consequently, product designers tend to use these tests sparingly during a product's developmental stage, and rarely recommend them for quality-control evaluations during production. One possible way around this limitation is to take sensory and consumer data and correlate them to instrumental measurements.

  "As a sensory person, I am very much in favor of this approach," says Gail Vance Civille, president, Sensory Spectrum, Chatham, NJ. "It can allow us to more easily do shelf life and quality control on the hundreds of thousands of products that are produced every year."

  Instrument-based tests offer several advantages. Results from instrumental tests are faster and more cost-effective compared with sensory paneling. During a product's developmental stage, using the objective, numerical result from an instrumental analysis makes it easier to compare successive formulas and to track progress toward the established goal. Another advantage is that instrumentation can be used as a quality-control tool where sensory panels often are impractical.

  Using instrumentation to perform sensory analysis isn't a brand-new idea. Alina Szczesniak, Ph.D., a researcher at General Foods in the early 1960s, defined texture profile analysis, and correlated descriptors with instrumental readings. In more recent years, electronic noses have become increasingly useful developmental tools. Although such tools and techniques are readily available, their usefulness often suffers from improper correlation with sensory data. By taking the time to properly establish a good correlation between sensory and instrumental data, food product designers can reduce developmental time and improve the quality of the products they create.

Notes on flavor

  The basic tastes of sweet, salty, bitter and sour are only part of the flavor experience. Much of it results from the interaction of human sensory organs with the volatile and semi-volatile organic chemical constituents foods contain. Product designers have been able to test for specific volatiles with gas chromatography/mass spectroscopy (GC/MS) methods for many years. Unfortunately, some foods may contain dozens, or even hundreds, of volatile flavor-contributing chemicals. Analyzing the flavor profile with GC/MS would simply be too time-consuming to fit into the typical project schedule.

  An electronic nose differs from standard chromatographic techniques because it doesn't attempt to separate every individual volatile component. Rather, it possesses an array of sensors. Each responds to various volatile chemicals in a slightly different way, and the instrument records an overall impression of these combined responses. Using the transmission from individual receptors to formulate an overall olfactory experience is pretty much the same way the human nose functions. Although an electronic nose contains far fewer sensors than the approximately 10,000 contained within the human nose, it uses a computer and specially designed software to interpret the sum of the various signals much like the human brain interprets olfactory sensor signals.

  As sophisticated as these instruments are, the various charts and data interpretations they generate are not the same as having a panelist describe a flavor experience.

  "The electronic nose has to be taught what to look for," Civille says. "That means that somebody with a good nose has to smell the samples to determine what is acceptable in the context of the project target. Not only do you have to define what's within the target, but also what lies outside the target range."

Biting into texture

  Although product designers have far more options for measuring texture compared with flavor, using these various instruments to accurately measure texture and mouthfeel still presents a major challenge. Texture is a composite property that embodies a number of physical properties - such as viscosity and elasticity - in a complex relationship. It is impossible to describe texture or mouthfeel by a single value obtained from an instrument. Furthermore, mouthfeel involves a food's entire physical and chemical interaction in the mouth - from initial perception on the palate, to first bite, through mastication and, finally, the act of swallowing. It provides a physical form of sensory gratification that goes far beyond simple rheology.

  Fortunately, the field of texture profile analysis (TPA) provides a set of texture terms and procedures for sample manipulation. This is essential because texture involves more than just initial biting force. The terms for describing solid oral texture are divided into seven general categories that follow, in chronological order, the events comprising mouthfeel/texture - from initial lip contact up through swallowing. Depending on the features of a particular product, these sensation categories can be expanded or eliminated depending on what factors are applicable.

  Surface texture is how the product's surface feels to the lips and tongue. It includes surface geometry, such as the overall degree of roughness and whether the rough sensation is large and bumpy or fine and grainy. Also included is the amount of wetness or oiliness on the surface.

  Partial compression without breaking evaluates how the sample returns to its original shape to determine the degree of springiness.

  First bite is an evaluation by the incisors, and covers several eating qualities. First hardness is determined by the force required to bite through. Cohesiveness is the amount of sample that deforms rather than ruptures, and fracturability is the force with which the sample breaks.

  Also important in first bite is the uniformity of the force throughout the bite, the amount of moisture released from the sample, and the amount of particles resulting from the bite or that are detected in the sample's center.

  First chew is an evaluation by the molars and includes many of the qualities found in first bite. Additional elements include adhesiveness and denseness. Adhesiveness is the force required to remove the sample from the molars. Denseness is evaluated as the compactness of the sample's cross section.

  Chew down evaluates the mouthfeel after several chews when the sample and saliva have mixed to form a mass. Qualities to evaluate here include the product's moisture absorption; cohesiveness, judged by the degree to which the mass holds together; and adhesiveness, described as the degree to which the mass sticks to the teeth or roof of the mouth.

  Rate of melt describes the amount of product melted after a certain number of chews. Along with this is an evaluation of the particles in the mass, how moist or oily the mass is, and the number of chews it takes to completely disintegrate the sample.

  Residual describes the sensations after swallowing or expectorating. Chalkiness, for example, is determined by the amount of particles left in the mouth. Also included is the amount of oil left on the mouth surfaces. Sticky mouth coating is evaluated by the degree of tackiness observed when tapping the tongue on the roof of the mouth. Last, "tooth packing" is evaluated by the amount of product left on the teeth.

  Semisolid products aren't chewed so much as they are "manipulated" in the mouth. Consequently, this sort of product requires other terms to account for the different oral handling.

  First compression is judged by pressing the sample between the tongue and the palate. Slipperiness is measured by the degree to which the product slides across the tongue. Firmness is the force required to compress the mass. Cohesiveness is the degree that the sample deforms, rather than shears. Denseness, as with solid samples, is the compactness of the cross section.

  Manipulation involves further compression of the sample to determine the amount of particles and their general size in the mass. Afterfeel is the amount of film or mouth coating remaining after swallowing or expectoration.

Making texture instrumental

  Current texture analyzers can perform complete TPA calculations, and often have various probes and rigs available, including various sizes of needles, cones, cylinders, punches, knives and balls. The various instruments apply these probes in several ways. This combination of probe and mode of action means that product designers can pretty much find an instrument that can determine: adhesion, bloom/gel strength, breaking point, cohesion, creep, crispiness, density, elasticity, extrudability, film strength, hardness, lumpiness, rubberiness, slipperiness, smoothness, softness, spreadability, springback, tackiness, tensile strength, viscosity and nearly every other known rheological property of foods.

  Among the various instruments, some specialize in measuring compression and elasticity. Others specialize in biting force. Most are equipped with specialized software so that after the instrument performs the test, the attached computer can perform the calculations to generate the final numbers and a graphic representation of the texture curve.

  As sophisticated as the technology is, the resulting data is meaningless unless the product designer can achieve a proper correlation to sensory testing. Like teaching an electronic nose what to look for in a product's aroma profile, product designers must examine the product to determine what texture-analysis methods will yield the most meaningful results.

A relevant record

  Correlating instrument readings for flavor and texture with sensory and consumer data is the only way to transform these results from esoteric numbers into a relevant record of the consumer's experience eating the product. This correlation doesn't begin after the instrumental analysis is complete. Rather, product designers need to follow some key guidelines from the get-go:

  Define the product. The process begins with a well-trained sensory panel to define the product, using TPA and descriptive flavor analysis. To be successful, the panelists must be able to analyze the product and create a lexicon for descriptive analysis. In addition, the panel also must be familiar with the texture classification system, the use of standard rating scales and the correct procedures for texture profile analysis. Panelist training should start with a clear definition of each attribute.

  Furthermore, the test operator must specify the techniques used to evaluate the food product. He or she also must explain how the food product is placed in the mouth, whether it is acted upon by the teeth (even the particular teeth in question) or by the tongue, and what particular sensation is to be evaluated. Panelists should be given reference standards for evaluation so they can practice their sensory evaluation techniques and the use of scales.

  Once the panel is trained, use it to test samples of the product and any similar products on the market. From this, the project manager can determine the attributes characteristic to the product category in question.

  Determine key attributes. Once the trained sensory panel has determined the product's sensory characteristics, it's time to determine which of these are most important to the product. Most foods offer a wide range of flavor and texture experiences. This step helps single out the features that are most critical in defining the product to make the correlation process more manageable.

  How relevant the various attributes of a product are should be determined by qualified consumer testing. This must not be neglected, because it is easy to assign importance to certain product characteristics that consumers may not find all that critical. The trained sensory panel determines the attributes defining the product. The consumer test determines which of these is most important for optimum product quality.

  Determine attribute scale. After a product designer determines the key attributes of a product, the next step is to assign intensity levels to these various flavor descriptors and texture attributes. Some sort of intensity scale is required for the sensory attributes to be meaningfully correlated with the numerical output of testing instruments. Perform this step by conducting a second round of sensory evaluation in which the trained panelists assign intensity levels on various descriptors/texture attributes. To evaluate the texture of a snack chip, for example, crispness most likely will be identified as an important attribute. Here, the panelists might be asked: "On a scale where 1=extremely soft and 9=extremely crisp, how would you rate the firmness of chip samples A, B and C?"

  Identify relevant instrumental tests. After taste paneling is complete, it's time to take instrument readings of the food product. But first, the product designer must determine what tests need to be conducted and what instruments should be used.

  "I see a lot of instrument misuse," Civille says. "I've observed people doing instrumental measurements, and I ask what it is they're measuring. Often, what they tell me they're measuring isn't at all related to the product's actual physical properties."

  If, for example, you're using an instrument that compresses the sample between two plates when a human would be biting with the incisors, the results will not correlate. Keep in mind, also, that people manipulate different types of products differently in the mouth. Pudding, unlike a cookie, is not bitten and chewed, but is compressed and spread around the mouth by the tongue. Still other products may require a combination of techniques. If that same pudding contains nuts, consumers will have to chew; they can't just manipulate.

  "If consumers will bite it and shear it, you have to get a knife and shear it," Civille explains. "If you're concerned about how a product sticks to the package, you can't measure how it sticks to a stainless steel rod. You have to simulate exactly how the product is used."

  Do your best to determine the right instrument and the right probe or rig to use to simulate how the product is consumed. Experiment with simulating what happens in the mouth by changing the shear rate of the instrument you're using. Foods tend to get thinner as they're chewed and consumers chew faster accordingly. You might try performing successive compressions on the same sample with an increased shear rate. Do your best, but always keep in mind that even the best-designed experimental procedure can encounter limitations.

  Perform the tests. With the instruments and procedures lined up, it's time to run the tests. Some details to consider during this process include testing the samples at the temperature they will be consumed. Product temperature affects both a product's texture and how flavor volatiles will be given off. Be prepared to run several tests at different temperatures if a product is used in multiple ways. Also keep in mind that multiple preparation methods - microwave vs. conventional oven, for example - also will affect flavor and texture. Test samples prepared using each method.

  Show statistical correlation. After collecting and compiling the data, it's time to attempt to show statistical correlation between sensory data and instrumental measurements. The most straightforward approach is to perform a statistical analysis to determine the relationships between the objective measurements and the sensory and consumer acceptance data. This will reveal which product features will yield the most acceptable product and, therefore, be the attributes that will be the most critical to monitor in the product's QC program. A word of caution: Some product attributes may seem to correlate with sensory and consumer data, but actually don't.

  "Sometimes it's just an artifact of the test method," Civille says. "There might not be a direct cause-and-effect relationship, but in this particular situation and this particular case, it happens to correlate."

  Occasionally, product designers can find no good correlation between instrument readings and taste panel scores. This is really no one's fault, but the available instrumentation just isn't capable of manipulating a food product in precisely the same way as the human mouth during mastication. Even if this should happen, the effort will not have been wasted, as you will have a clear sensory and consumer acceptance history on the product to guide the development process.

  Correlating instrumental analysis with sensory data can be a very useful product-development tool. This process is not only faster than running repeated sensory panels, but more economical. Thanks to this economy, product designers can use instrumental correlations to more effectively guide the product-development process, and help plan a quality-assurance program. After all, it is often said that consumers won't continue to buy a product if it isn't enjoyable. Correlating sensory data with instrumental testing helps designers optimize the eating pleasure to make products more successful.

  Scott Hegenbart is multimedia production specialist with the Department of Food Science and Technology at the University of Nebraska-Lincoln, where he develops methods for teaching food science using computer-based multimedia and the World Wide Web. During his nearly 14 years in the food industry, he has authored numerous articles on food product development for a wide variety of publications.

Back to top