2021 Volume 27 Issue 6 Pages 817-836
The physical properties of foods related to mastication and swallowing are explained. The maximum velocity, Vmax, obtained by the ultrasonic pulse Doppler method, and the time required for the bolus to flow in the pharynx, t2>, obtained by acoustic swallowing sound, are considered to be good predictors of aspiration risk. For thickener solutions, the viscosity, µ, especially that measured at a shear rate of 20–30 s−1 and above, can be a suitable index for liquid-type care foods for people with difficulties in swallowing. The ‘hardness’ is considered to be the most suitable index for semi-solid care foods among the parameters obtained by texture profile analysis (TPA). The Vmax and t2> can be interpreted as “cohesiveness (i.e., ease to form bolus)” of food from the viewpoint of the flow velocity distribution in the pharynx. On the other hand, ‘cohesiveness’ obtained by TPA does not reflect the ease of forming a bolus of food.
With the advent of an aged society, the incidence of dysphagia has been increasing. Patients with dysphagia often aspirate food, which is one of the factors causing pneumonia in elderly people (Igarashi et al., 2002; Crary and Groher, 2003; Nishinari, 2004).
Human beings chew foods, mix them with saliva, and swallow them as a bolus in the case of solid or semi-solid foods. Thin (low viscosity) liquids, which are swallowed with little mastication or mixing with saliva, flow readily with little cohesion between particles and, therefore, are easily deformed during swallowing. Thus, a thin liquid may penetrate the larynx and enter the trachea if laryngeal closure does not occur quickly enough to prevent liquid penetration. Many dysphagic patients are not capable of controlling laryngeal closure, and, therefore, often aspirate liquids of low viscosity (Palmer, 1997; Palmer et al., 2003). Thickeners or gelling agents are used in foods for dysphagic patients to enhance viscosity and cohesion between food particles, and many commercial foods and powders containing thickeners and gelling agents have been developed.
Care foods for people with difficulties in mastication and swallowing are required to be non-sticky and cohesive. Various standards and criteria have been set for the physical properties or texture of care foods, including stickiness and cohesiveness. In 1994, the Ministry of Health, Labour and Welfare in Japan defined the diet criterion for patients with difficulties in mastication and swallowing, where standards of the physical properties of sols and gels are indicated. In 2009, the Ministry of Health, Labour and Welfare in Japan set the standard criteria for foods authorised for people with swallowing difficulty (hereinafter referred to as “the Criterion”), where the standards of the parameters obtained from texture profile analysis (TPA) were determined. The criterion was placed under the control of the Consumer Affairs Agency, Government of Japan, from 2010. The Japan Care Food Conference, organised by various Japanese food companies, defined another criterion for universal design foods in 2003, setting voluntary standards for thickened foods. However, the appropriate physical properties of foods for patients with difficulty in swallowing have not yet been determined.
Because the physical properties of food change during mastication and swallowing, it is difficult to assess the risk of aspiration only by using the physical properties of foods obtained from mechanical measurements. Thus, non-invasive bioinstrumentation has attracted great interest among food scientists in recent years. Considering that a thin liquid flows through the pharynx too rapidly for patients with dysphagia to control, it is very important to analyse the velocity distribution of the liquid (or bolus) in the pharynx. Videofluorography (VF) is a well-known method for directly observing the process of mastication and swallowing (Palmer, 1997; Palmer et al., 2003; Takahashi et al., 2003; Nishinari, 2004). However, multiple collections of data on the same subject increase the risk of damage due to exposure to X-ray radiation. In addition, given that a contrast medium such as barium sulphate needs to be added to foods for VF examination, the physical properties of the ingested foods may be altered. Nakazawa et al. (2000) were the first to apply the ultrasonic pulse Doppler method to measure the velocity of a bolus of boiled rice through the pharynx, followed by Nagatoishi et al. (2001), who measured the velocity of various types of food in the pharynx. As the ultrasonic pulse Doppler method is safe and is without the need for a contrast medium for velocity measurement, it can quantitatively compare the velocity of various foods passing through the pharynx.
Acoustic analysis of swallowing sounds is an effective bioinstrumentation method to clarify swallowing profiles. However, this method has been mainly utilised for diagnostic purposes as a non-invasive test (Lazareck and Moussavi, 2004), but rarely in food texture studies (Nakauma et al., 2011). In the measurement of swallowing sound, not only the flow sound of the bolus itself at the pharynx, but also all the sounds emitted by human organs at the time of swallowing can be recorded. Nakauma et al. (2011) investigated the relationship between physiological response and sensory perceived scores in the swallowing of polysaccharide thickener solutions using acoustic and sensory evaluations. Ishihara et al. (2011) investigated the swallowing profiles of food polysaccharide gels obtained by acoustic analysis in relation to bolus rheology.
In this review, the physical and textural properties of foods related to mastication and swallowing are explained. In addition, the relationship between these properties and the information obtained from bioinstrumentation is discussed.
2.1. Viscosity (Kumagai et al, 2008; Kumagai and Kumagai, 2009)
2.1.1. The definition of the viscosity and the diet criterion of sol for patients with difficulties in masticating and swallowing in 1994
The viscosity, µ, which is a physical property and characterises the flow resistance of simple fluids, is defined as follows:
 |
is the shear rate. The fluid with a viscosity value that is constant irrespective of 
is referred to as a Newtonian fluid, and the fluid with a viscosity value that changes depending on 
is a non-Newtonian fluid (Bird et al., 1960).
In 1994, the Ministry of Health, Labour and Welfare in Japan defined the diet criterion of sol for patients with difficulties in masticating and swallowing as the value of viscosity (steady shear viscosity) µ of sol to be ≥ 1.5 × 103 mPa·s measured 2 min after the revolution of the rotor has started with a B-type (Brookfield-type) viscometer, a rotational viscometer with a vertical cylindrical rotor (rotor rotation rate, 12 rpm; 20 ± 2 °C) (Sakai and Kayashita, 2006); the Japan Care Food Conference, organized by various Japanese food companies, also defined a similar criterion for the Universal Design Foods for sol in 2003. However, this criterion was empirically determined. The B-type viscometer does not provide an accurate value of µ, except for Newtonian liquids, with a viscosity that is constant irrespective of 
. Although the approximated value of 
can be empirically calculated with the B-type viscometer, it depends on the type of rotor and size of the sample container, even at the same rotational rate. Kumagai et al. (2009) measured the viscosity of solutions prepared from widely used thickeners using two viscometers, a B-type viscometer and a cone-and-plate viscometer, with which an accurate value of µ even for non-Newtonian liquids can be obtained as a function of 
, and the B-type viscometer determined an approximated value of viscosity at a selected shear rate 
. In addition, the rate of rotor rotation (12 rpm) in the criterion set in 1994 was calculated to give a 
of 2 to 3 s−1 in the viscosity range used as thickener solutions for patients with dysphagia.
2.1.2. The solution viscosity and dynamic viscosity―empirical Cox–Merz rule (Tashiro et al., 2010a)
The complex viscosity, η*, is defined as follows:
 |
where G* (≡ G′ + iG″; G′ is the storage modulus, G″ is the loss modulus) is the complex elasticity, i is the imaginary unit, and ω is the angular frequency for the dynamic elasticity measurement.
The solution viscosity 
is related to the complex viscosity η*, according to the following empirical Cox–Merz rule (Cox and Merz, 1958):
 |
where η′ (=η′ (ω)) and η″ (=η″ (ω)) are the real and imaginary parts of η*(=η*(ω)), respectively (Cox and Merz, 1958). This rule states that the values of µ and |η*| should be identical on a comparable time scale of observation, as long as the structure that determines the mechanical properties of the material remains intact under steady flow (Ikeda and Nishinari, 2001). The parameters η′ (the dynamic viscosity) and η″ (the out-of-phase dynamic viscosity) are related to G′ and G″, as follows:
 |
 |
In this paper, the absolute value of the complex viscosity η*, | η* |, is referred to as η*, as is conventional in rheology.
Hydrocolloids are rheologically characterised according to ω, the dependence of G′, a measure of elasticity, and G″, representing viscous components, as follows (Nishinari, 2000):
(i) for a strong gel (an elastic gel or a true one), G′ is far larger than G″, and both moduli G′ and G″ are independent of frequency;
(ii) for a weak gel, G′ is slightly larger than G″, and the two moduli are only slightly dependent on the frequency. Note that “weak gel” is essentially a liquid and that another word to describe such a liquid as a “structured liquid” has been proposed (Picout and Ross-Murphy, 2003). However, the term “weak gel” will be used in the subsequent parts of this paper.
(iii) in the case of a true polymer solution, G′ is smaller than G″ at lower frequencies, but both moduli G′ and G″ increase with increasing frequency and show a crossover, G′ becoming larger than G″ at higher frequencies; and
(iv) for a dilute polymer solution, G′ is far smaller than G″ at all frequencies, and both moduli G′ and G″ are strongly dependent on the frequency, and at very low frequencies, the relations G′∼ω2 (the slope of log10G′ vs. log10ω plot is 2) and G″∼ω (the slope of log10 G′ vs. log10ω plot is 1) are known.
2.1.3. The viscosity of thickener solutions
Figure 1 shows the dependencies of the viscosity µ, dynamic viscosity η′, and complex viscosity η* of the thickener solutions on ω. The decrease in µ with increasing 
, the so-called shear thinning (Morris et al., 1981; Richardson et al., 1989; Rinaudo, 1993; Ross-Murphy, 1995), is observed for all the thickeners examined. The values of µ for the xanthan gum and guar gum solutions decreased significantly with an increase in 
, as previously reported (Richardson and Ross-Murphy, 1987; Milas et al., 1990). On the other hand, shear thinning for carboxymethylcellulose (CMC) was slight, and the CMC solutions showed a behaviour similar to that of the Newtonian fluid in the range of concentrations examined. Carboxymethylcellulose solutions have usually been reported to exhibit non-Newtonian behaviour, but shear thinning was not prominent at low concentrations, according to the studies by Ismail et al. (1980) and Tako and Nakamura (1984).
Shear rate (
) or angular frequency (ω) dependence of viscosity (µ), dynamic viscosity (η′), and complex viscosity (η*) for carboxymethylcellulose (CMC), xanthan gum, and guar gum solutions.
Open symbols (○, □, △, ◇), dynamic viscosity, η′; solid symbols (●, ■, ▲, ◆), complex viscosity η*
+, ×, steady shear viscosity, µ
The thickener solutions in Fig. 1 can be characterised from ω, the dependence of G′, and G″, as follows (data are not shown), and the CMC solution can be taken as the dilute solution in the concentration range examined. The xanthan gum solution can be taken as a weak gel above 0.5 % and as a true polymer solution below 0.2 %. The guar gum solution can be used as a true polymer solution in the examined concentration range (Tashiro et al., 2010a).
The ω dependency of η′ and η* is similar to that of µ, but small differences between µ, η′, and η* are observed. In particular, η* of the xanthan gum solutions was several times larger than µ and η′ at high concentrations. The values of µ and η* are similar at identical values of 
and ω. The empirical Cox–Merz rule states that the shear rate (
) dependence of viscosity, µ, is superimposable on the angular frequency (ω) dependence of the complex viscosity, η*, for many polymer solutions (Eq. [3]). In addition, since the values of G′ and G″ are of the same order of magnitude, the values of η″ and η′ also become of this same order, which makes the value of η* similar to those of η′ and η″ (Eq. [2]). Therefore, the plots of η′, η*, and µ would be similar to one another, as shown in Fig. 1. However, small differences between µ, η, and η* were observed. In particular, η* of the xanthan gum solutions was several times larger than µ and η′ at high concentrations. For the Cox–Merz rule, it is necessary that the structure that determines the mechanical properties of a material should remain intact under steady flow (Ikeda and Nishinari, 2001). At high extents of coil overlap, polymer samples fail to obey the Cox–Merz rule because of the formation and breaking of non-covalent bonds. Ikeda and Nishinari reported that a weak gel-type κ-carrageenan solution did not obey the Cox–Merz rule (Ikeda and Nishinari, 2001). The xanthan gum solution used in this study can be considered as a weak gel above 0.5 %. This is the reason why small differences between µ and η* were observed, especially for the xanthan gum solutions at high concentrations.
2.2. Texture profile analysis (Akima et al., 2018)
Texture profile analysis (TPA) is a double compression test for determining the textural properties of foods. During TPA, test samples are compressed twice using a rheometer equipped with a plunger, and TPA is therefore often called the “two bite test” (Fig. 2). Among the parameters observed in TPA, ‘hardness’, ‘adhesiveness’, and ‘cohesiveness’ have been widely used for comparison of the sensory attributes and rheological properties of various foods (Friedman et al., 1963; Nishinari et al., 2013). The ‘hardness’ is the peak force during the first compression cycle. The ‘adhesiveness’ is the negative force area B for the first bite. The ‘cohesiveness’ is the ratio of positive force area during the second compression to that during the first compression (A2/A1) (Fig. 2). According to “Overview of Texture Profile Analysis” shown on the homepage of Stable Macro Systems Ltdi)., ‘cohesiveness’ is how well the product withstands a second deformation relative to its resistance under the first deformation. Note that these three parameters obtained from TPA will be referred to as ‘hardness’, ‘adhesiveness’, and ‘cohesiveness’ by adding single quotation marks (′ ′) in this paper.
Texture profile analysis (TPA).
In “the Criterion” determined for foods of people with swallowing difficulty by the Ministry of Health, Labour and Welfare in 2009, the measurement of ‘hardness’, ‘adhesiveness’, and ‘cohesiveness’ should be conducted on samples poured into a cup of 40-mm diameter and 20-mm height to a depth of 15 mm. Using an apparatus that can measure uniaxial compression stress, the compression speed should be fixed at 10 mm/s, and clearance should be set at 5 mm. A cylindrical plastic plunger 20 mm in diameter and 8 mm in height should be lowered and raised twice. In “the Criterion”, even liquid samples that cannot keep their shape but flow under gravity are poured into a cup and subjected to uniaxial compression, without considering the validity or physical meaning of the TPA parameters. Table 1 shows the diet criterion of texture for foods prepared for individuals with difficulty in swallowing by the Consumer Affairs Agency, Government of Japan. Some critical comments have recently been published concerning the validity of TPA, especially the concept of ‘cohesiveness’ (Nishinari et al., 2019).
| Parameter obtained by TPA | Standard regulation | ||
|---|---|---|---|
| I | II | III | |
| Hardness (N/m2) | 2.5 × 103 ∼ 1 × 104 | 1 × 103 ∼ 1.5 × 104 | 3 × 102 ∼ 2 × 104 |
| Adhesiveness (J/m3) | < 4 × 102 | < 1 × 103 | < 1.5 × 103 |
| Cohesiveness | 0.2 ∼ 0.6 | 0.2 ∼ 0.9 | – |
Standard regulation I: serious swallowing difficulty
Standard regulation II: moderate swallowing difficulty
Standard regulation III: slight swallowing difficulty
TPA: texture profile analysis
Figure 3 shows the ‘hardness’, ‘adhesiveness’, and ‘cohesiveness’ values for four types of polysaccharide gels. The gels in Fig. 3 can be taken as strong gels based on the dynamic elasticity measurements (Akima et al., 2018). The figure also shows the range of the criterion set by “the Criterion”: I, II, and III are the standard value ranges for people with severe dysphagia, for people with moderate dysphagia, and for people with mild dysphagia, respectively. The data on the left (a) and the right (b) are measured at test speeds (plunger speed) in TPA of 1 mm/s and of 10 mm/s (set by “the Criterion”), respectively. There are no significant differences in the values of ‘hardness’, ‘adhesiveness’, and ‘cohesiveness’ between test speeds of 1 mm/s and 10 mm/s. The values of ‘hardness’ vary from about 500 to 20 000 N/m2 by changing the concentration. The values of ‘adhesiveness’ increase with increasing concentration of the gelling agent (Akima et al., 2018). The values of ‘cohesiveness’ vary from about 0.2 to 0.9 by changing the sample used and the concentration of the gelling agent (Akima et al., 2018). In this way, the ‘hardness’, ‘adhesiveness’, and ‘cohesiveness’ of the gel can be controlled by changing the type and concentration of the gelling agent.
The parameters (‘hardness’, ‘adhesiveness’, and ‘cohesiveness’) of gels obtained from texture profile analysis (TPA).
Test speed: (a) 1 mm/s; (b) 10 mm/s;
■□, AG (agar); ◆◇, LC (locust bean gum & κ-carrageenan);
●○, LX (locust bean gum and xanthan gum); ▲△, NG (native gellan gum)
I, II, III: standard criteria for foods authorised for people with swallowing difficulty by the Consumer Affairs Agency in Japan;
I, serious swallowing difficulty; II, moderate swallowing difficulty; III, slight swallowing difficulty.
The ‘hardness’, ‘adhesiveness’, and ‘cohesiveness’ can be measured for liquids by the method in “the Criterion”. The values of ‘hardness’ and ‘adhesiveness’ of thickener solutions such as xanthan gum increased with increasing concentration of the gelling agent. As for liquids, there were small deviations (within about 10 %) in the values of ‘hardness’, ‘adhesiveness’, and ‘cohesiveness’ between test speeds of 1 mm/s and 10 mm/s (Hasegawa-Tanigome et al., 2013). The ‘cohesiveness’ of water is almost unity, and decreases with increasing concentration of the gelling agent (Nishinari et al., 2013; Hasegawa-Tanigome et al., 2013). Referring to the TPA parameters for liquid foods contained in a cup is, however, a “misuse” that the founder of TPA feared in 1998 (Szczesniak and Bourne, 1998; Kumagai et al., 2011; Nishinari et al., 2013). Some review papers on TPA were published in the Journal of Texture Studies in 1975. All of these papers described the application of TPA for solid foods (Szczesniak and Hall, 1975) and various solid foods in Japan, where this method is the most widely used (Tanaka, 1975). Thickened liquids for elderly people with difficulty in mastication and deglutition are, however, widely evaluated by the TPA method (the standard of the Consumer Affairs Agency, Japan).
3.1. Measurement of velocity through the pharynx by the ultrasonic pulse Doppler method
An ultrasonic apparatus for diagnosis, ECCOCEE (SSA-340 A type; Toshiba Medical Systems Co., Tokyo, Japan) equipped with a linear scan probe (PLF-703NT) for pulse Doppler measurement was used at 6.0 MHz, a frequency suitable for measuring a relatively shallow part of the body. All subjects were female, healthy, and free from mastication or swallowing problems. The subject sat on a chair, at room temperature (20 to 25 °C), with her head fixed and her back straightened. An ultrasonic probe was set against the neck (Fig. 4 [A]) so that the ultrasonic pulse was directed at an upward angle of 60°. The subject scooped up a sample of 6 g, put it into her mouth, and swallowed it in a single swallowing movement. Immediately thereafter, measurements were taken in the B-mode and colour Doppler mode. The B-mode was used to identify a region of the pharynx, based on the fact that the reflection of ultrasonic waves is different among organs. In the colour Doppler mode, the signal obtained by the pulse Doppler method is illustrated in colour. A velocity distribution spectrum, brightness, as a function of the passage time and velocity through the pharynx (Fig. 4 [B] [a]), was obtained, where the brightness is proportional to the volume of “food particles” (small portions of food detectable with the ultrasonic probe) at a given passage time and velocity. In Fig. 4 (B) and (b), a three-dimensional velocity spectrum is shown, and the volume fraction of the bolus in the pharynx is expressed as a function of the passage time and velocity. Velocity distribution spectra (n = 20–30), measured as described above, were averaged using Image-Pro Plus ver. 5.0 (Nippon Roper Co., Ltd., Tokyo, Japan) to obtain a colour velocity distribution spectrum (Fig. 4 [B] [c]). In the colour spectrum, the volume fraction of “food particles” increases in the order of red, green, and blue.
Measurement of the velocity through the pharynx by the ultrasonic pulse Doppler method.
(A) The direction of the ultrasonic pulse during the measurement of velocity distribution through the pharynx. The direction of the ultrasonic pulse is shown by the large arrow.
(B) Spectra of velocity distribution through the pharynx
(a) Velocity distribution figure; (b) three-dimensional velocity spectra of swallowed food particles slightly above the epiglottis; and (c) colour velocity spectrum, which is the average of 20 to 30 times the distribution figure (a).
3.2. Analysis of velocity distribution spectra (Nakazawa et al., 2000; Hasegawa et al., 2005; 2008a; 2008b)
In Fig. 5, the spectrum for water, which is often aspirated into the trachea of patients with dysphagia, is distributed over a wider velocity range than that of yoghurt, which is rarely aspirated. In other words, water scatters into smaller portions in the pharynx, while yoghurt does not. From the spectra, the maximum velocity, Vmax (m s−1), and mean velocity, vm (m s−1), were calculated. To reduce background noise, we defined the maximum velocity, Vmax, as the velocity at which the brightness value was 12 dB lower than that of the brightness in the statistical mode of velocity. In Fig. 5, the maximum velocity Vmax of water is approximately a few times higher than that of yoghurt (Subject X).
The velocity spectra of water and yoghurt of the Subject X.
The spectra are an average of 20–30 times: (b) relationship between the velocity, time, and particle volume fraction in the velocity spectra shown in (a), which shows the highest particle volume at each time. The dB (about the lower) line was deduced from the mode line; (c) maximum velocity of water (■) and yoghurt (●).
4. 1. Relationship between the rheological properties of thickener solutions and Vmax (Kumagai et al., 2009; Tashiro et al., 2010a; 2010b)
Figure 6 shows the typical velocity distribution spectra of the thickeners and that of water, shown for comparison. The spectrum for water is distributed over a wider velocity range than those of the other solutions. Similarly, the spectra for the thickener solutions at low concentration are distributed over a wider velocity range than those of the thickener solutions at high concentration.
Typical colour spectra showing velocity through the pharynx of (a) carboxymethylcellulose (CMC), (b) xanthan gum, and (c) guar gum.
The velocity spectra for water, which dysphagic patients often aspirate into their trachea, are also shown for comparison.
The data were obtained from Subject X
Figure 7 shows the dependencies of the mean (vm) and the maximum (Vmax) velocity of the thickener solutions through the pharynx calculated from the velocity spectra (Fig. 5) on the thickener concentrations; data on water (0 % concentration on the abscissa) and yoghurt are shown for comparison. In addition, thickener concentration dependencies of µ of the thickener solutions at shear rates of 2.5 and 25 s−1 are also shown. The decrease in the mean velocity, vm, with increasing concentrations of all thickeners is slight. However, the value of Vmax for the thickeners decreased as the thickener concentration increased. The values of Vmax for the thickener solutions approach that of yoghurt, although the asymptotic values seem to be slightly larger than the value for yoghurt (0.20–0.06 m s−1). Given that Vmax is more concentration-dependent than the mean velocity, vm, it is better to use Vmax as an index for care foods for dysphagic patients than the mean velocity. The values of µ at a shear rate of 2.5 s−1, which is within the values 2 to 3 s−1 calculated from the rotor revolution rate, 12 rpm, in the diet criterion defined by the Ministry of Health, Labour and Welfare of Japan in 1994, increased to 0.39 Pa·s (2.5 %), 4.5 Pa·s (1.5 %), and 6.4 Pa·s (1.0 %) for CMC, xanthan gum, and guar gum, respectively, with increases in the thickener concentration. Note that these values of µ are larger than 1.5 Pa·s in the criterion in 1994, although the values of Vmax of these thickener solutions are slightly larger than that of yoghurt. This indicates that the value of µ = 1.5 Pa·s at 2–3 s−1in the criterion is inappropriate as an index for care foods for dysphagic patients.
Dependencies of velocities of passage for thickener solutions through the pharynx and of viscosity, µ, on the thickener concentration.
Data for water (concentration, 0 %) and yoghurt are also shown for comparison.
●, maximum velocity, Vmax, of thickener solutions; ■, Vmax, of yoghurt;
◆, mean velocity, vm, of thickener solutions; ▲, vm of yoghurt;
○, µ at 
= 25 s−1; □, µ at 
= 2.5 s−1
The data were obtained from Subject X
Figure 8 shows the relationships between the maximum velocity, Vmax, and the viscosities (µ, η′, and η*) at ω, 
= 2.5 rad/s, or 25 rad/s (or s−1). The maximum velocity Vmax of the thickener solutions correlates well with µ, η′, and η* on semi-logarithmic plots, as shown by the regression curves and the values of the correlation coefficient, R. From this aspect, µ, η′, and η* can be suitable indices for care foods of the liquid type for dysphagic patients. Threshold values of µ, η′, and η* of thickener solutions for the diet for dysphagic patients were evaluated in the 
and ω ranges examined. The order of magnitude of the threshold values of µ, η′, and η*, which give the maximum velocity of yoghurt, Vmax = 0.2 m s−1 is 1 Pa·s, at 
or ω of 20–30 s−1 (or rad/s) (Tashiro et al., 2010a).
Correlation between maximum velocity through the pharynx, Vmax, and viscosity (µ), dynamic viscosity (η′), and complex viscosity (η*), at 
= 2.5 s−1 or ω = 2.5 rad/s (three figures at left) and at 
= 25 s−1 or ω = 25 rad/s (three figures at right) for the various thickener solutions.
●, carboxymethylcellulose (CMC); ■, xanthan gum; ◆, guar gum.
The data were obtained from Subject X
The magnitude of the shear rate related to swallowing is unclear. Shear rate of 50 s−1 is widely used but not so well established (Nishinari et al., 2019). In the Classification of Modified Diet for Dysphagic Persons in 2013 of the Japanese Society of Dysphagia Rehabilitation, µ at 
of 50 s−1 is recommended (Fujitani et al., 2013). In Criteria for Labelling Permission for “Foods for Special Dietary Uses” by the Consumer Affairs Agency, Government of Japan, µ at 
of 50 s−1 is also recommendedii). The appropriate shear rate is considered to be one order of magnitude larger than 2–3 s−1.
4. 2. Relationship between texture of gels and the maximum velocity Vmax (Hasegawa-Tanigome et al., 2013; Akima et al., 2014; 2017; 2018)
Figure 9 shows the relationships between the three parameters of the polysaccharide gels in Fig. 3 and Vmax (Subject Y). The plots of Vmax vs. the three textural parameters measured at a plunger speed of 10 mm/s were similar to those measured at 1 mm/s. The ‘hardness’ is correlated best with Vmax, among the three textural parameters. On the other hand, the ‘cohesiveness’ was not correlated with Vmax. The ‘hardness’ is, therefore, considered to be the most suitable index for care foods for dysphagic patients among these three parameters. The value of Vmax for gels decreased as the ‘hardness’ increased and approached that of yoghurt: the value of Vmax is similar to that of yoghurt above 10 000 N/m2 of the ‘hardness’. In “the Criterion” in 2009, the upper limit and lower limit values of ‘hardness’ are set for food for people with swallowing difficulty: the lower limit values of levels I, II, and III are 2.5 × 103 N/m2, 1 × 103 N/m2, and 3 × 102 N/m2, respectively. These values of the lower limit of ‘hardness’ are of the same order at which gives the Vmax of yoghurt.
Relationship between the parameters obtained from the texture profile analysis (TPA) of gels and maximum velocity, Vmax, for Subject Y.
Test speed: (a) 1 mm/s; (b) 10 mm/s;
■□, AG; ◆◇, LC; ●○, LX; ▲△, NG
The ‘hardness’, ‘adhesiveness’, and ‘cohesiveness’ are measured in a cup for both liquids and solid semi-materials. Liquid samples cannot keep their shape but flow under gravity. As mentioned before, referring to the TPA parameters for liquid foods contained in a cup is a “misuse”. In a cup, the ‘cohesiveness’ of water is almost unity, and decreases with increasing concentration of the thickening or gelling agent. In the measurement without a cup, the cohesiveness of water is almost zero. If the cohesiveness were measured without a cup, such cohesiveness would be correlated with Vmax. The erroneously defined ‘cohesiveness’ would have led to an illogical conclusion.
The apparent viscosity of the boluses of polysaccharide gels at a shear rate of 25 s−1 and Vmax of the bolus (Subject Y) is shown in Fig. 10. The apparent viscosity of the boluses of Subject Y increased with increasing gel concentration for each gel. In addition, Vmax decreased with an increase in the apparent viscosity of the boluses. As shown in Fig. 3, the ‘hardness’ of the gels increased with increasing gel concentration. From these aspects, the high apparent viscosity accompanied by large ‘hardness’ of gels at high concentrations is considered to cause a decrease in Vmax.
Apparent viscosity of boluses of gels for Subject X at a shear rate of 25 s−1 and maximum velocity, Vmax, of the bolus (Subject Y).
■, AG; ◆, LC; ●, LX; ▲, NG
4. 3. Can ‘hardness’ be used as an index for thickened care foods for people with dysphagia instead of viscosity?
The Japan Care Food Conference set a voluntary standard for thickened food: ‘hardness’ is used as an index of “thickness strength” (four levels), but the solution viscosity µ is not included in the index. Funami et al. (2009) reported that the ‘hardness’ of thickener solutions was highly correlated with the solution viscosity µ, that is, plots of µ vs. ‘hardness’ were well fitted with a straight line. If so, ‘hardness’ could be used as an index for care foods for people with dysphagia, not only for gels but also for thickener solutions (sol).
Figure 11 shows the relation between the viscosity µ and ‘hardness’ for liquid samples with various flow properties (Akima et al., 2017). Carboxymethylcellulose and locust bean gum solutions were considered to be Newtonian fluids, while xanthan gum solutions and commercial thickening powder solutions showed similar shear thinning (Akima et al., 2017). In Fig. 11, the plots of µ at a shear rate of 3 s−1 vs. hardness are well fitted with a straight line. However, plots of µ at shear rates of 25 s−1 and 50 s−1 vs. hardness were fitted with two different straight lines for solutions with different flow properties. Funami et al. (2009) measured the viscosity at a rotor rotation rate of 12 rpm with a B-type viscometer, according to the diet criterion of sols for patients with difficulties in masticating and swallowing set by the Ministry of Health, Labour and Welfare in Japan in 1994. The rotor rotation rate of 12 rpm corresponds to a shear rate of 2–3 s−1 (Kumagai et al., 2009). As shown in Fig. 5, the viscosity µ at 
of 20–30 s−1 and above correlates well with the maximum velocity through the pharynx Vmax, which is a suitable index for the liquid type of care foods for dysphagic patients. As mentioned before, the appropriate shear rate is considered to be one order of magnitude larger than 2–3 s−1. Viscosity at a shear rate of the order of 101 s−1 vs. ‘hardness’ cannot be fitted with one straight line, as shown in Fig. 11.
Relationship between the viscosity and hardness for liquid samples.
▲, xanthan gum; ◆, locust bean gum; ●, carboxymethylcellulose (CMC); ▲▼ ◆ ■, commercial thickening agents (▲, T1; ▼, T2; ◆, T3; ■, T4) ; ○, honey; ◇, PG; □, GE; △, CA
Figure 12 shows the relationships between the ‘hardness’ of the gels and thickener solutions (sol) and Vmax (Subject Y). The values of Vmax for both gels and thickener solutions decreased as the ‘hardness’ increased and approached that of yoghurt. However, the value of the ‘hardness’ at which Vmax is similar to that of yoghurt is about 1 000 N/m2 for thickener solutions and about 10 000 N/m2 for gels. The value of ‘hardness’ for most thickener solutions changed only to approximately 300 to 1 000 N/m2, but it is considered that the viscosity increased and Vmax decreased. In the case of gels, if the value of the ‘hardness’ was below 1 000 N/m2, Vmax did not decrease as much as that of yoghurt: if it was above 10 000 N/m2, Vmax was considered to decrease to that of yoghurt. Therefore, the ‘hardness’ is useful as a parameter for evaluating the risk of aspiration, but in the case of samples with different properties, such as liquid samples and gels, it is desirable to set the criteria according to the differences in properties. From these aspects, it is difficult to use hardness as a substitute for viscosity at high shear rates related to human swallowing. When using hardness for liquid foods, it is necessary to consider the rheological characteristics such as flow properties.
Relationship between hardness obtained from the texture profile analysis (TPA) and maximum velocity, Vmax, for sols and gels.
▲, xanthan gum; ◆, locust bean gum; ●, carboxymethylcellulose (CMC);
Commercial thickening agent solution ▲, T1; ▼, T2; ◆, T3; ■, T4
◇ □ △ ; gel (◇, Psyllium seed gum & deacylated gellan gum (PG); □, deacylated gellan gum GE; △, κ-carrageenan)
The data were obtained from Subject Y
Acoustic analysis of swallowing sounds is an effective method for clarifying swallowing profiles. However, this method has been mainly utilised for diagnostic purposes as a non-invasive test (Lazareck and Moussavi. 2004), but rarely in food texture studies (Nakauma et al., 2011). In the measurement of swallowing sound, not only the flow sound of the bolus itself at the pharynx, but also all the sounds are emitted by human organs at the time of swallowing. Nakauma et al. (2011) investigated the relationship between the swallowing analysis and rheological behaviour of polysaccharide thickener solutions. From the profiles of the swallowing sounds of each subject, they evaluated the time required for the epiglottis to close (t1), the time required for the bolus to flow (t2), and the time required for the epiglottis to open (t3). The duration t2 decreased as the concentration of the gel increased, and t1 and t3 were independent of the concentration. Ishihara et al. (2011) investigated the swallowing profiles of food polysaccharide gels obtained by acoustic analysis in relation to bolus rheology.
We investigated the relationship between the velocity of the bolus in the pharynx and acoustic sounds during swallowing using the method described by Nakauma et al. (2011). All subjects were female, healthy, and free from mastication or swallowing problems. Subjects were asked to put the specimen into the mouth and to swallow the whole amount (10 g) once without mastication by teeth or compression between the tongue and the hard palate. The swallowing sound was recorded using a throat microphone (i.e., a vibration sensor). The microphone was placed temporarily into a position approximately 5 mm from the vocal cords and fixed with a neck band. Data were processed using the software CSL-4400 (Kay Elemetrics Corp., NJ, USA) to determine the frequency distribution pattern.
Figure 13 shows a typical profile of swallowing sounds for (a) water and (b) yoghurt (Akima et al., 2016). The time required for the bolus to flow t2 of yoghurt is shorter than that of water. The value of time t2 of gel tended to decrease as the concentration increased, approaching to that of yoghurt (Akima et al., 2016, data are not shown). In Fig. 14, the t2 of gels (Subject Z) is shown: t2 tended to decrease with the increase in the ‘hardness’, which is caused by the apparent steady viscosity µ of the gel bolus. The value of t2 is smaller as that of Vmax became smaller, with t2 correlating well with Vmax. Therefore, acoustic analysis of sounds could also be a simple method for evaluating velocity.
Typical profile of swallowing sound: (a) water, and (b) yoghurt.
The data were obtained from Subject Z
The time required for the bolus of gels to flow through pharynx, t2.
(A) Dependence of duration time for flow of bolus through the pharynx t2 on the parameters obtained from texture profile analysis (TPA).
(B) Relationship between t2 and maximum velocity Vmax.
◇, PG; △, PS; □, GE; ●, Yoghurt; ▼, Water The data were obtained from Subject Z
6. 1. Physical meaning of Vmax
The maximum flow velocity Vmax was considered to evaluate the width of the velocity distribution of the bolus in the pharynx (Fig. 15). When the bolus slowly and coherently flows through the pharynx (as in the yoghurt), the flow velocity distribution would be narrow, causing Vmax to be smaller. When the bolus flows in pieces (such as water), the flow velocity distribution becomes wider because particles with low flow velocity and particles with high flow velocity coexist, causing Vmax to be larger (note that the change in the mean velocity vm is slight, as shown in Fig. 7). Therefore, Vmax is regarded as a parameter that quantifies the width of the flow velocity distribution and could be a good predictor of aspiration risk.
Relationship between cohesiveness (cohesiveness in the pharynx) of bolus and maximum velocity Vmax (or t2).
6. 2. Relationship between Vmax obtained from flow velocity measurement and t2 obtained from swallowing sound measurement
The relationship between the maximum flow velocity, Vmax, obtained from pharyngeal flow velocity measurement using ultrasound and the duration of bolus flow through the pharynx, t2, obtained from swallowing sound measurement will be discussed here. As shown in Fig. 7, the pharyngeal flow velocity is approximately 10−1 to 100 m s−1 from the pharyngeal mean velocity vm and the maximum velocity Vmax. Because the length of the pharyngeal part of a human being is 10−2 to 10−1 m, the bolus passage time of the pharyngeal part can be estimated to be several tens to several hundreds of milliseconds. Given that the t2 shown in Fig. 14 is approximately 101 to 102 ms, it is considered reasonable to regard as the time for the bolus to pass through the pharynx. The maximum velocity Vmax represents the width of the pharyngeal flow velocity distribution, and in the bolus with a large value of Vmax, there are portions with high flow velocity and those with low flow velocity. The passage time through the pharyngeal part is shorter than the part with a high flow velocity (sample piece) and the part with a low flow rate (Fig. 15). As a result, when Vmax is large, t2, which is the passage time of the entire sample, is considered to be large. In other words, as shown in Fig. 14, there is no contradiction in that the value of t2 decreases as the value of Vmax decreases. From these aspects, it is considered that t2 and Vmax are different physical quantities, but their trends of increase and decrease are similar.
6. 3. Cohesiveness of bolus and ‘cohesiveness’ obtained from TPA (Nishinari et al., 2013; Hasegawa-Tanigome et al., 2013; Kumagai et al., 2019)
When t2 or Vmax is large, the flow velocity distribution is wide in the pharynx; that is, when the bolus passes “without collecting” if thickener solutions or bolus flow coherently through the pharynx, t2 or Vmax is smaller than that of yoghurt (Nakauma et al., 2011; Akima et al., 2016). From this aspect, t2 or Vmax can be interpreted as "cohesiveness (i.e., ease to form bolus in the mouth)" of food (if t2 or Vmax is smaller, the bolus is more easily formed and flows more coherently) from the viewpoint of the flow velocity distribution in the pharynx.
When liquids such as water are poured in a cup and subjected to TPA, the calculated ‘cohesiveness’ becomes almost united. With an increasing concentration of thickener solutions,' hardness' and ‘adhesiveness’ increased and ‘cohesiveness’ decreased (Nishinari et al., 2013; Hasegawa-Tanigome et al., 2013). It is recognised that water has a high probability of inducing aspiration because it is the least cohesive and tends to scatter into smaller portions. Yoghurt is widely recommended for individuals with difficulties in mastication and swallowing because its texture is believed to be cohesive and does not scatter into small portions in the oropharyngeal cavity. The ‘cohesiveness’ of the thickener solution obtained by conventional TPA is far smaller than that of water (Nishinari et al., 2013; Hasegawa-Tanigome et al., 2013). Thus, extending the TPA concept is erroneous.
(i) The maximum velocity Vmax of water, which is often aspirated into the trachea of patients with dysphagia, is about a few times higher than that of yoghurt, which is rarely aspirated. The maximum velocity Vmax of the thickener solutions approached that of yoghurt as the concentration increased. Therefore, Vmax is considered a good predictor of the aspiration risk of foods.
(ii) For thickener solutions, the maximum velocity through the pharynx, Vmax, correlates well with the viscosity, µ, the dynamic viscosity, η′, and the complex viscosity, η*, especially those measured at 
or ω of 20–30 s−1 (or rad/s) and above. From this aspect, µ, η′, and η* can be suitable indices for the liquid type of care foods for dysphagic patients.
(iii) The order of magnitude of the threshold values of µ, η′, and η* of thickener solutions, which give the maximum velocity of yoghurt, is 1 Pa·s at 
or ω of 20–30 s−1 (or rad/s).
(iv) The parameters obtained by TPA, the ‘hardness’, ‘adhesiveness’, and ‘cohesiveness’ of the gel, can be controlled by changing the type and concentration of the gelling agent.
(v) The ‘hardness’ is correlated best with Vmax among the three textural parameters. On the other hand, the ‘cohesiveness’ was not correlated with Vmax. The ‘hardness’ is, therefore, considered to be the most suitable index for care foods for dysphagic patients among these three parameters. The value of Vmax for gels is similar to that of yoghurt, above 10 000 N/m2 of the ‘hardness’. Referring to the TPA parameters for liquid foods contained in a cup is a “misuse”. In a cup, the ‘cohesiveness’ of water is almost unity. The erroneously defined ‘cohesiveness’ would have led to an illogical conclusion.
(vi) The time required for the bolus to flow, t2, by analysing the acoustic swallowing sound tends to decrease as the concentration increases, approaching that of yoghurt. The value of t2 is smaller as the value of Vmax decreases, and t2 correlates well with Vmax. Therefore, t2 is also considered a good predictor of the aspiration risk of foods.
(vii) The parameters t2 and Vmax can be interpreted as “cohesiveness (i.e., ease to form bolus in the mouth)” of food (if t2 or Vmax is smaller, the bolus is more easily formed and flows more coherently) from the viewpoint of the flow velocity distribution in the pharynx. On the other hand, ‘cohesiveness’ obtained by TPA does not reflect “cohesiveness (i.e., ease to form bolus in the mouth)” of foods.
