Robot”s “taste sensor” : Visualizing Taste with the Eyes

Is it possible to communicate “taste” using your other senses? Is it possible to express it numerically? Is it possible to objectively represent it using a machine?

Electric signals obtained from the sensor are converted to taste quality based on the Weber-Fechner law which gives an approximately accurate generalization of the intensity of sensation. The base of logarithm is defined as 1.2. For example, 12.5 units means 10 times higher concentration than that of the original sample, and 25 units is 100 times higher concentration. After the simple conversion, we can visualize the taste as “taste map”. The axis shows taste and each unit represents the intensity of tastes.

There is an infinite variety of answers to those questions. “It’s impossible to quantify taste or communicate it to others because taste is very individual.” “Taste is subjective, and therefore cannot be measured numerically.” “I can appreciate taste, and so why does it need to be quantified using a machine?”

These answers were all made by self-claimed academic experts during the early stage of the development of a “taste sensor” (about 10 years ago). A “taste sensor” is “equipment, apparatus or a device that measures taste”.

The question that needs to be asked before exploring the nature of “taste” is whether or not “taste” is actually measurable, with the answer to that being “Yes”. This can be explained for physiological reasons. “Taste” is determined at the level of the nerve fibers, which then connect with gustatory cells. “Taste” is made up of five basic taste sensations: sour, bitter, sweet, salty, and umami. The fifth taste sensation, “umami” or savory, was discovered as an independent taste sensation by the Kiyoshi Toko & his teams, and is therefore known as “Umami” in English too.

“I could not differentiate the orange juice from the apple juice when I drank it with my nose pinched shut, which made me realize that we actually differentiate taste by smell.” The sweet “taste”, in the context of “increased sweetness when a pinch of salt is added to sweet red-bean soup”, is “sensed by the brain”, and therefore not easily analyzed.

Just to keep the record straight, the beginning of the development of a taste sensor was quantifying or objectifying “taste sensed by the tongue” or the so-called chemical taste, not the “complicated taste sensed by the brain” (which perhaps could be called subjective taste) described in the above example.

In making a taste sensor electrodes are attached to a membrane produced from a lipid mixture, a relative of soap, with polymer molecules. The sensor indicates the result with the help of a computer. A device based on Japan’s original technology has been patented in various countries around the world. Its mechanism is that an interaction between the lipid membrane and chemical substances changes the latter into electric signals, which are transferred along a wire to a computer, where the data obtained is analyzed. The lipid membrane outputs information on the taste, thus quantifying the basic taste sensations of sour, bitter, sweet, salty, and umami.

The mechanism is rather similar to that of the actual biological system. Our tongues have small raised protrusions called papillae formed from flower bud-like organs called taste buds. Taste buds are comprised of a set of gustatory cells. Each individual gustatory cell is covered with a biological membrane made up of lipids and proteins. The biological membranes are where a variety of chemical substances get processed. The information received here is transformed into electric signals, which travel to the brain via the gustatory nerves.

If this mechanism were to be viewed in terms of the taste sensor, the lipid membrane, wire and computer would be the biological membrane, gustatory nerve and brain, respectively.

The use of this taste sensor allows taste to be viewed at a glance. For example, regular beer or low malt beer can be quantified as human sensations with the taste of the malt being plotted on the horizontal axis and the taste of the sharpness/dryness on the vertical axis. The device can immediately tell you that beers with a strong malt taste such as Yebisu and Kirin Lager emphasized the sharpness/dryness axis after Asahi Super Dry was released on the market, and that recently released low malt beers and other quasi-beers now follow this tendency. And therefore taste can be viewed by the eyes.

Music has scores written for it, which is the reason why we can reproduce and listen to the works of Beethoven and Bach today in the 21st century. A database produced by taste sensors could be called a “score for reproducing food” or a “food score”. A “food score” would make it possible to transfer your mother’s cooking, traditional dishes and secret recipes beyond time and space. Traditional dishes could be reproduced. Food with the same taste as that enjoyed on earth could also be provided at space colonies and satellite stations in the future.

Let’s return to the question posed at the beginning: “Can taste be measured?” The characteristic taste of hot-selling food products could be expressed by clarifying the correlation between the human senses and numerical values representing the basic taste sensations of sour, bitter, sweet, salty, and umami obtained from a taste sensor (or an objective and subjective match). In fact, it is possible to objectify taste. Needless to say, taste is not universal and is significantly affected by age, location, history, race, culture etc. However, assessing a taste that suits a specific race or culture etc can be realized merely by setting appropriate values for it as the basic taste sensations of sour, bitter, sweet, salty, and umami are capable of being measured.

At present taste sensors are being used by more than 200 organizations that include food manufacturers, pharmaceutical companies, public experimental laboratories and centers, and universities. Intelligent Sensor Technology Inc. (abbr.: INSENT; head office: Atsugi), a company developing, manufacturing and marketing taste sensors, and the Taste & Aroma Strategic Research Institute (head office: Yokohama), a company distributing and providing a taste database, have both been successful with increasing sales.

The National Museum of Emerging Science and Innovation (Miraikan) holds events where “Taste Measuring Robots” are constructed. It is very popular with participants because of the educational electricity, chemistry and biology aspects of the fabrication process, which is just one example why a taste sensor would be a good scientifically educational object for use by elementary and junior high school students.

Recently the problem of students “avoiding science” is often debated. That terminology, however, should only be used after deliberating on the real meaning. The problem is not students “avoiding science” but rather “avoiding education”. What is surprising is the surge in number of students who cannot write or speak Japanese. This phenomenon stems from the incorrect style of education they received while at elementary and junior high schools. Teachers yell nothing but “let’s participate in sport clubs and enjoy sports together”.

There is something horribly wrong going on there. Although humans as animals include the instinct of acting as groups in protecting themselves, education that only emphasizes sports is way off the mark. Here of course we don’t even need to refer to the misleading policy of “relaxed education”. As part of its very nature education is a lonely activity, and the youth of today cannot abide loneliness. They always have cell-phones with them and feel uneasy if they cannot immediately contact someone.

1. What Is  Taste Sensor?

A sensor is a device which changes various types of physicochemical properties of target to a different state ( usually electrical signals ). Our feeling body parts
are also sensors.The stimulus received by ears and eyes is changed into
an electrical signal which is sent to the brain.

A taste sensing system has been developed at Kyushu University. Transducers of the sensor are composed of lipids immobilized with polyvinyl chloride. The multichannel electrode is connected to a channel scanner through high-input impedance amplifiers. The electric signals are converted to a digital code by a digital voltmeter and then transferred to computer. The sensor output is not the amount of specific taste substances but the taste quality and intensity. The sensor has a concept of “global selectivity” which is the ability to classify enormous kinds of chemical substances into basic taste groups such as saltiness, sourness, bitterness, umami and sweetness.

Let’s consider the situation we sell or buy something.

The one who wants to sell it says that it contains more contents (more than
it does),

whereas the buyer estimates it as less.

If the both have a measure, i.e., sensor, to quantify the weight,

this kind of argument will go well.

Yes, it weights 125 g.

Then, it is 12500 Yen because it is 100 Yen per 1 g.

As you can see from the above example,

sensors provide us a quantified, objective scale.

In selling or buying foodstuffs,

our opinions do not agree,

as one says it is delicious while another says it is not so good.

So, it is very convenient if we have a sensor to measure the taste.

Then, what do we have to do if we want to compare tastes?

Suppose one eats some delicious food, how can one tell another person about
the taste.

In that case, if there is a sensor which can measure the taste and convert
the taste to figures,

we can all imagine what the taste is like and it will be easy to compare

How are we able to measure tastes? Let’s think about it.

Suppose you know the quantities of all the ingredients, then can you tell
the taste of the food?

To begin with, there are tens of thousands of different substances

which produces taste and hence measuring all of them is very difficult.

Also, concerning taste, there is a known phenomenon called a mutual effect.

The quantity of a substance cannot easily decide the taste.

Concerning the mutual effect, if you put some salt on

a water melon, it becomes more sweet, isn’t it so?

How do we perceive taste then?

So if we mimic our taste reception mechanism, we might be able to measure

2. Taste Reception Mechanism

Our bodies are made up of cells.

Of course, the tongue which perceives taste is also made up of cells.

The cell’s skin, biological membrane, consists of a double layer of lipid
molecules and pruteins.

Lipids are oil-like substances contained in our bodies.

Lipids are made up of the water loving (hydrophilic) part and the water
hating (hydrophobic) part.

There is a lot of water in both the inside and outside of the cell

and hence lipid molecules make up a double layer on the biological

with the hydrophilic part facing the inside and outside of the cell.

In water, the membrane is electrically charged

because the hydrophilic parts of lipid melecules are ionizing.

Also, the inside and outside of the cell are full of

different concentrations of salt-like substances.

K+ ions flow from the inside to the outside of cell,
because the membrane permeates to K+ and the concentrations is higher in the
inside.This electric current due to K+ ions flow causes

the electric potential difference between the cell interior and exterior.

It is called a membrane potential,

which is affected by some kinds of chemical substances contained in the
cell exterior.

The nature of taste reception cells differs from cell to cell and also
nerve to nerve.

Also, foods have different physicochemical properties.

The electric potential of nerve (nerve excitation) and the membrane
potential of taste cell depend on the food.

For ourselves, we can judge if the taste is sweet or sour from electric
potential change pattern in the brain.

3. Taste Sensor

Then, let’s make the imitation of taste reception mechanism and construct a
taste sensor.

We saw that the tongue’s cell membrane receive the taste.

Let’s mix polyvinyl chloride (PVC) and a lipid

to make an artificial membrane and see if we can’t use it as a biological

The so-formed membrane (lipid/polymer membrane) is pasted on a small tube,
which contains

potassium chloride solution (i.e., the inner side solution of the cell).

Various foodstuffs change the membrane potential of the cell.

The tube where we pasted the lipid membrane is immersed into orange juice,

and then the juice’s main ingredient causes the membrane potential change.

By investigating the electric potential changes,

isn’t it possible that we can measure taste elicted by chemical substances?

In case of ourselves,

we can judge if something tastes sweet or sour from the electric potential
change patterns

in the brain due to various cells.

The taste sensor utilizes plural kinds of membranes with different lipids.

Lipid membranes have different response properties for chemical substances.

The electric potential change of lipid membranes depends on the taste.

The electric potential change pattern of the lipid/polymer membrane

is investigated and the taste is quantified Using the comupter.

Then, let us measure taste using the taste sensor which uses lipid

The graphs below show the results for two kinds of sour ingredients and

three kinds of “umami” ingredients.

Here, we used seven different kinds of lipid membranes.

The sour tastes of lemon and vinegar are due to different ingredients,

but the seven membranes’ electric potential change pattern are similar.

The pattern is sightly different and the difference comes from

the difference of the sour taste between lemon and vinegar.

The “umami” tastes of a kelp, a dried bonito and a shiitake mushroom

(these are Japanese typical broth base) are due to different ingredients;

as expected, however, the electric potential change patterns are similar.

Furthermore, the group of sour foodstuffs and that of “umami” foodstuffs

electric potential change patterns which are completely different.

We found that foodstuffs with the same taste had the same graph shapes

and those with different tastes had different patterns.

As you can see, the taste sensor using lipid membranes can measure tastes.

Please read Biomimetic Sensor Technology (Cambridge University Press) for
the detail.

Kiyoshi Toko
Cambridge University Press
First published:

This book deals with the development of what are known as
the electronic nose and the electronic tongue, that is to say odor and
taste sensors. Of all sensor technologies, these have been widely
considered as the most difficult to realize.

These sensors have been developed on the basis of
mechanisms found in biological systems, such as parallel processing of
multidimensional information or by the use of biomaterials. The author
begins with an introductory section dealing with the principles of
measurement and multivariate analysis. Reception mechanisms in biological
systems are briefly reviewed. It is shown that a taste sensor using lipid
membranes can reproduce a sense of taste and provide clues for the
reception mechanism in gustatory systems. Several types of biosensor,
including enzyme-immobilized membranes, SPR, the quartz resonance
oscillator and IC technologies are explained in detail. Generally a sensor
is designed to reproduce or surpass the human senses. For this purpose, it
is not enough simply to detect a single input signal. Instead, it is
necessary to be able to perform adequate processing of diverse kinds of
information. The expansion of sensor technologies into everyday life will
require the fusion of sensor technologies in various fields. This book is
dedicated to the development of intelligent sensors and systems.

This book will be of value to researchers from a wide
range of fields, including electrical engineering, food science,
electronics, mechanical engineering, medical science, biochemistry,
biophysics and biotechnology.