What is the Electronic Tongue?
The electronic tongue (or e-tongue) is an analytical device with the ability to discriminate, identify, and quantify taste and aftertaste. Through its taste sensors, it can interact with all molecules responsible for a given taste. When the interaction happens, the taste sensors send signals to an analysis unit and are treated by a data processing software. The data is processed using individual sensor responses to give a profile of the product taste. This mechanism allows the obtention of a digital fingerprint of product gustatory characteristics that highly correlates with human taste perception. The first models of electronic tongue appeared in the 90' and have now evolved and into more sophisticated systems. The leading example is the TS-5000Z made by the Japanese company Insent, used by New Food Innovation, agent for the company and consultant in taste analysis. Currently, over 500 systems have been sold and are used throughout the world. They find most of their applications in the food and pharmaceutical sectors in both research, development and manufacturing. Through this article, you will better understand the value of an electronic tongue and what motivated its development. Also, you will get a much deeper understanding of what it is and the way it works.
Why using an electronic tongue?
You might well be a food adventurous, appreciating trying out all sorts of new foods and discovering a whole world of odors, taste, and textures. With all the effort taken by chefs and industries to deliver new recipes and new products on the shelves, choices are endless. Whether you liked or not your latest discovery, to what extent are you able to precisely describe it? Even if you face the simple question, “how does it taste like?”, would you be able to answer? And even if you do so, will your answer ever be precise enough? Physical, psychological factors and individual preferences influence taste evaluation. Your food habits, food history, social, physical environment, and interaction with other senses are some other factors that also have an impact. In summary, food taste evaluation is not objective. However, the food industry relies on delivering products that are consistent in taste and flavour.
To get a satisfactory and reliable answer to this question, there were traditionally two options. The first one consists in asking a sensory panel to try the product for you and ask them the question. In this scenario, you will have to ask people to actually taste the product and give you feedback. Having a trained taste panel is expensive to train and maintain and you are limited in the quantity of products a palate can taste in one sitting. Even in standardised conditions, you will realize that the spread of the results hardly gives clear conclusions and they may only be applicable to that territory. What’s salty to someone in the UK may not be described as such by someone in Japan. You should then get the evaluations from 30, 50, 100, or even more people to get an average close to reality. Only then, you may be able to get to some sort of consensus. Otherwise, asking an expert on this product is also a good approach. Who can best accurately describe wine than an oenologist with 30 years of experience? It’s an expensive investment that is why some of the best have their tongues and noses insured for millions of pounds. While their expertise applies to wine it cannot be readily applied to coffee or another foodstuff. In the case you can find one, can the average food eater understand what he describes?
Overall, using humans is still an acceptable option and is often the way to go. Yet, there are situations when human expert panels cannot, or should not be used. Here are a few examples:
Repetitive quality control on an industrial processing line with a high volume of samples and where precise and consistent results are wanted
Samples that could potentially be toxic such as tasting of drugs, pharmaceuticals, evaluate food spoilage, or alcoholic drinks
Unappealing and unappetizing samples such as dogs or cats food
Economic reasons, and financial expenses but also a waste of time for recruitment, training, running tests and supervising the session.
If you want a globally understandable taste representation of your product.
If you decide not to use humans then there is analytical equipment that can help. Using chemical analyses, such as high-performance liquid chromatography (HPLC). This precisely indicates what are the chemical substances found in a product as well as their quantity. With the data obtained from the chemical analysis, you could try to predict taste. Unfortunately, that would be almost impossible. There is a large number of taste substances in a product and so many interactions that can occur. Because of synergistic and suppression effects, the taste of a chemical can be enhanced or reduced by another one. And those are very hard to guess. Adding sugar to a cup of coffee makes it sweeter yes, but it also makes it less bitter. Chemical quantification alone could be an option but is far to be the straightest way to measure taste.
To resolve this problem, a few researchers developed a taste sensing technology. This is an analytical device for discriminating and quantifying the taste of foods and other substances called the electronic tongue.
The expended definition
Among the academic community, the widely accepted definition of an e-tongue system is: “The electronic tongue is an analytical instrument comprising an array of nonspecific, low-selective, chemical sensors with high stability and cross-sensitivity to different species in solution and an appropriate method of PARC and/or multivariate calibration for data processing” [1].
Let’s give additional clarification.
“an array of nonspecific, low-selective, chemical sensors”
The name of the electronic tongue was given for its similarity to the taste sense of humans. Other chemical sensors would generally detect a target chemical substance specifically. Yet, the taste receptors that are on our tongue do not recognize individually every chemical substance. The taste sensors used on the e-tongue work in a similar way to those on our tongue.. They both are nonspecific and low-selective, which is also called global selectivity. This means that all the substances that are responsible for a given taste are taken into account to give the final reading.
“with high stability and cross-sensitivity to different species in solution”
Those sensors also have high stability. They will be able to give the same output for the same sample a day, a month, a year after as long as the testing conditions stay the same. The chemical part of people's taste perception might be the same a month after the other. Yet, physiological, psychological, and environmental factors will make the results much more volatile.
Cross-sensitivity relates to the fact that similar substances will have similar tastes. This is one way for the sensor to perceive taste. For example, molecules demonstrating a salty taste will generally be of a simple atomic constitution, showing a strong electrical conductivity.
In food, taste compounds have to migrate into the saliva before reaching the taste buds during mastication. The e tongue is exactly the same all measurements are done in solutions. Solid materials will have to be dissolved or extracted before reaching the taste sensors.
“and an appropriate method of PARC and/or multivariate calibration for data processing”
Once sensors have collected information the crude data is convrted into something more understandable. The determination of the taste quality of a product uses pattern recognition and classification (PARC), multivariate analysis methods, and other actions of artificial neural networks.
The first concept of a taste sensor appeared around 1990. Toko et al. applied for a patent of their taste sensor in 1989. Then, they developed a taste sensor equipped with multichannel electrodes using a lipid/polymer membrane for converting the chemical information to an electrical signal. This taste sensor with global selectivity composes part of the electronic tongue. Since then, systems based on various sensor array designs and chemometric strategies have been successfully exploited. Application examples are ranging from foodstuffs, environmental control, pharmaceutical, medical diagnostics, as well as process monitoring. It finds most of its applications in the pharmaceutical and food industry.
How does it work?
We can now spend some extra time understanding how it works. How does the magic happen? Starting from a sensor or an electrode submerged in the sample solution, how does the technology manage to give scores of the different tastes, aftertastes, and sometimes other indicators such as taste persistence or sharpness?
Let's come back to what inspired the electronic tongue, our own human tongue. The sense of taste consists of the five basic tastes that are saltiness, sourness, bitterness, umami, and sweetness. The way we perceive taste first starts on our taste buds. When eating a beverage or food, it triggers those sensory organs located on the tongue. Each of them is composed of 50-100 cells and are receptors for a very wide range of chemical substances. If activated by one of those chemicals, it enables the perception of the associated taste. To this end, the taste information is converted into a change in electrical potential that is transmitted to taste nerves. The signal conveyed by the nerve enables the release of neurotransmitters in the brain. It finally reaches the gustatory area in the brain that processes and converts the information received which judges the taste and its intensity.
If taken alone, chemical substances have no taste or smell. Knowing their composition or properties is not enough to predict their sensory characteristics. It's only when we receive, convert, and interpret the information they carry that we can understand their flavor profile. The taste sensors were conceived taking nature as a source of inspiration. Indeed, the working principle of a taste sensor and the electronic tongue follow the same path of processing the information. When we present a sample to the sensors their non-specific, semi-selective properties allow them to generate an overview of its gustatory traits. As inspired by biological recognition, this vast volume of data is analyzed by the "brain" of the machine. The electronic tongue software can apply chemometric, artificial intelligence and statistical methods. As a result, it can discriminate, identify, or quantify the taste profile of the sample. If the software was well pre-trained and calibrated, it is possible to ensure a high correlation with human perception. Yet it is using only objective data and thus, producing objective and quantitative results.
The machine also has greater potential than humans. As it can access all the data and not only the final decrypted taste values, dynamic factors are also recorded. The data signal measure can be done over time, and the evolution of the intensity followed. The taste active molecules take time to reach the sensors as do with taste buds in the mouth. This information can be exploited to get an increase in the precision of the predictive model. As well, it opens the doors for experiments that could not be possible otherwise. Thanks to the taste sensor, we determine the kinetic of taste diffusion. Well exploited it could be used to explore further some addition in food. Salt, sugar, mono sodium glutamate (MSG), or other polemical taste enhancers could be challenged by the way food is consumed. On the other side, it could help give guidelines on how to get the most (or the least) taste release. The machine can outperform human senses. It can “taste” very dilute solutions that untrained people will consider bland and thus have a wider operative spectrum. It can sense off notes before the human can detect the products quality deteriorating.
Among the five, the taste and smell are very different from the others as they only exist through human senses. Taste and smell are very linked to the person's perception and interpretation capacity to translate the information received. In this context, taste sensors and electronic tongue can be the route to go for a non-biased and quantified evaluation of taste.
Florian Woisel
Sources:
Ciosek, P. (2016). Milk and dairy products analysis by means of an electronic tongue. In Electronic Noses and Tongues in Food Science (pp. 209-223). Academic Press.
Nakamoto, T. (Ed.). (2016). Essentials of machine olfaction and taste. John Wiley & Sons.
Podrażka, M., Bączyńska, E., Kundys, M., Jeleń, P. S., & Witkowska Nery, E. (2018). Electronic tongue—A tool for all tastes?. Biosensors, 8(1), 3.
Toko, K. (Ed.). (2013). Biochemical sensors: mimicking gustatory and olfactory senses. CRC Press.
