Masking the Detection of Taste Stimuli in Rats: NaCl and Sucrose

Abstract

While psychophysical and neurophysiological assessments of taste sensitivity to single chemical compounds have revealed some fundamental properties of gustatory processing, taste stimuli are rarely ingested in isolation. Arguably, the gustatory system was adapted to identify and report the presence of numerous chemicals ingested concurrently. To begin systematically exploring the detectability of a target stimulus in a background in rodents, we used a gustometer to train rats in a 2-response operant task to detect either NaCl (n = 8) or sucrose (n = 8) dissolved in water, and then tested the sensitivity of rats to the trained NaCl stimulus dissolved in a sucrose masker (0.3, 0.6, or 1.0 M, tested consecutively) versus sucrose, or the trained sucrose stimulus dissolved in a NaCl masker (0.04, 0.2, or 0.4 M) versus NaCl. Detection thresholds (EC₅₀ values) were determined for the target stimulus dissolved in each concentration of the masker. Except for 0.04 M NaCl, all masker concentrations tested increased the target stimulus EC₅₀. Target stimulus detectability decreased systematically as masker concentrations increased. The shift in liminal sensitivity for either target was similar when the threshold for the masker was considered. At least for these prototypical stimuli, it appears that the attenuating impact of a masker on the detection of a target stimulus depends on sensitivity to the masking stimulus. Further study will be required to generalize these results and extend them to more complex maskers, and to discern neural circuits involved in the detection of specific taste signals in the context of noisy backgrounds.

Introduction

While eating, animals sample and ingest complex stimuli that have diverse chemical compositions and sensory properties. During this action so critical to survival, the chemosensory systems must provide the brain with an accurate analysis of the contents of foods and fluids to optimally generate adaptive behavioral responses. For example, ingestion of essential nutrients (e.g., sugars and amino acids) is generally promoted, while ingestion of poisonous substances is generally discouraged. Signals arising, in part, from the gustatory system, provide the animal with information on an expedient temporal scale allowing for appropriate responses to occur early after eating or drinking has begun.

While investigators typically study taste using a single chemical compound dissolved in water, purposefully simplifying the experimental design in efforts to reveal the basic response properties of the gustatory system, humans and animals rarely encounter pure stimuli in the environment. Rather, they encounter complex mixtures and, as the literature attests, taste stimuli interact when sampled together (see Keast and Breslin 2003; Wilkie and Capaldi Phillips 2014). Thus, the response of the system to a single taste compound could be expected to be affected by the presence of other tastants. Arguably, the gustatory system evolved to detect beneficial and dangerous compounds that are presented in mixtures.

Experimentally speaking, the simplest combination of taste stimuli is a binary mixture, containing only 2 taste stimuli (usually dissolved in water). Taste mixture studies with humans are commonly conducted with subjects tasked with reporting the intensity, or the quality, of 1 component. For instance, a person may be given a sample of NaCl + sucrose and asked to report the magnitude of saltiness. When presented with heterogeneous mixtures, containing 2 qualitatively different stimuli such as NaCl and sucrose, the reported intensity of either component is often decreased by the presence of the other. For example, in the case of NaCl + sucrose mixtures, the intensity of the target quality (saltiness) of the mixture may be less than that of the relevant component in water (i.e., the same concentration of NaCl in Water; e.g., Stevens 1996; Green et al. 2010). This reduction in reported intensity, relative to that component by itself in water, is a common result in taste mixture studies and is referred to as mixture suppression. The degree of suppression is affected by the quality, concentration, and chemical properties of each component (see Keast and Breslin 2003; Wilkie and Capaldi Phillips 2014). Mixture suppression helps to explain why adding sugar to coffee makes it seem less bitter, and has ramifications for human health—for instance, adding sweeteners to bitter pediatric formulation increases the likelihood of acceptance (Mennella et al. 2013).

Taste mixture studies conducted with rodents have yielded results similar to those in humans. Generalization studies typically use a conditioned taste aversion paradigm, whereby an animal is first trained to avoid either a mixture or one of its components by the presentation of the stimulus immediately followed by the administration of an agent that causes visceral malaise. The animal is then tested for its generalization of the learned aversion to the mixture and/or the components. These studies have shown that an animal trained to avoid a mixture will sometimes avoid the components to a lesser degree than animals trained to avoid the individual components alone in water (e.g., Smith and Theodore 1984; Frank 1985; Frank et al. 2003). This result suggests that while the component qualities are still identifiable in the mixture, each is less intense than when presented in isolation, similar to the results of human studies.

Although the conditioned taste aversion procedure provides potentially useful information regarding the perceived quality of mixtures and their components, it is not without limitations. It is possible that some compounds (e.g., bitter ligands) are more vulnerable to association with visceral malaise. If so, the actual intensity of the component in the mixture could be conflated with its saliency in the task. Moreover, the conditioned avoidance expressed to a given concentration of the test stimulus is related to the specific concentration of the conditioned stimulus in training. Thus, the greater the disparity between the concentration of the test stimulus and the conditioned stimulus, the lesser the conditioned response. This is referred to as stimulus generalization decrement and because taste aversion conditioning usually involves a single concentration of the conditioned stimulus, the types of responses observed in the test are in some part influenced by the concentration of the training stimulus (Nowlis 1974; Harder et al. 1984; Scott and Giza 1987; Spector and Grill 1988).

Thus, when interested in studying component detectability within mixtures, a different methodology may be required. To that end, here we developed a paradigm that would allow the systematic study of the ability of rats to identify the presence of a specific stimulus when presented in a constant chemical background (i.e., in a binary mixture with another tastant) that would minimize such interpretive concerns. The 2-response taste discrimination procedure has great utility in this regard. In this task, a thirsty rat is given a sample of stimulus and receives water for correct responses, continually reinforcing performance across all test sessions, rather than testing in extinction as with conditioned taste aversion, which will eventually lead to a loss of the conditioned effect. This consistent reinforcement across all concentrations also reduces the likelihood of generalization decrements. Since the stimulus serves as a cue for behavior (responding), its hedonic properties are less likely to influence responding, reducing the issues of unconditioned avoidance that occur with certain stimuli when used in intake tests. The presentation of small volumes and the measurement of immediate responses increase confidence that the behavior is based on the orosensory properties of the stimulus. This method therefore provides a tool for quantifying the detectability of a stimulus in a constant background. Note that this approach differs from a methodology designed to study the detectability of a mixture itself, which is sometimes used in the literature (e.g., Stevens 1995). The stimuli used in this study were NaCl and sucrose, chosen based on their clear discriminability when combined in rodent generalization mixture studies (Smith and Theodore 1984; Frank 1985; Grobe and Spector 2008), indicating that the presence of each chemical is detectable when appearing together in sufficiently high concentrations.

Materials and methods

Subjects

Twenty male Sprague-Dawley rats, approximately 10 weeks old at the beginning of the study, were used as subjects. Animals were individually housed in polycarbonate cages with environmental enrichment (Rattle-A-Round, Otto Environmental) in a room with a 12-h light/dark cycle and automatically controlled temperature and humidity. Food (Rodent Diet 5001, PMI) was available ad libitum throughout the experiment. Water (reverse-osmosis deionized water; DW) was available contingent upon training and testing (see Operant training for descriptions, Table 1 for summary). Testing occurred Monday through Friday. Water bottles were removed the afternoon before the first session, and replaced after the last session, of the testing week. Animals had ad libitum access to water over the weekend. Procedures were approved by the Animal Care and Use Committee at Florida State University.

Table 1.

Schedule of experimental phases

Phase	Number of Daysa	Stimuli	Stimulus Presentation
Operant Training
Stationary Training	3	Water	Ad libitum
Side Training	4	0.4 M NaCl or water	Constant
Alternation	4	0.4 M NaCl and water	Semirandom
Random Training	5	0.4 M NaCl and water	Semirandom
NaClb in Water I	22	NaCl vs. water	Semirandom
NaClc in 0.3 M sucrose	18	NaCl+0.3 M Sucrose vs. 0.3 M Sucrose	Semirandom
Masking concentration series
NaClb Masked by Sucrose
NaCl in 0.6 M sucrose	15	NaCl+0.6 M Sucrose vs. 0.6 M Sucrose	Semirandom
NaCl in 1.0 M sucrose	16	NaCl+1.0 M Sucrose vs. 1.0 M Sucrose	Semirandom
NaCl in Water II	20	NaCl vs. Water	Semirandom
Sucroseb Masked by NaCl
Sucrose in Water	14	Sucrose vs. Water	Semirandom
Sucrose in 0.4 M NaCl	12	Sucrose+0.4 M NaCl vs. 0.4 M NaCl	Semirandom
Sucrose in 0.2 M NaCl	12	Sucrose+0.2 M NaCl vs. 0.2 M NaCl	Semirandom
Sucrose in 0.04 M NaCl	14	Sucrose+0.04 M NaCl vs. 0.04 M NaCl	Semirandom

a Days were sequential but not necessarily consecutive.

b When used as the target stimulus, the first NaCl concentration tested was always 0.4 M, and the first sucrose concentration tested was always 0.3 M. One concentration of the target stimulus was used in a single session with the concentration decreased across sessions.

c The target stimulus was dissolved in the listed masker, filled to an appropriate volume. Thus, the discrimination for each phase is as listed as: Target+Masker vs. Masker.

Stimuli

Stimuli were made daily using reagent-grade chemicals (Macron Fine Chemicals) dissolved in DW. NaCl detection thresholds were determined using 0.00015–0.4 M NaCl. The concentrations of NaCl used to mask detection of sucrose ranged from 0.04 to 0.4 M NaCl. Sucrose detection thresholds were determined using 0.0025–0.3 M sucrose, while the concentrations of sucrose used to mask detectability of NaCl ranged from 0.3 to 1.0 M sucrose.

In masking phases, the masking stimulus was dissolved in DW to create the masking solution. Binary mixtures were made by dissolving the appropriate mass of the target stimulus (here, the solute) in the premade masking solution (i.e., the solvent). This strategy created a solution with the intended concentration of target stimulus, plus the intended concentration of masking stimulus, as needed for the phase of testing.

Apparatus

All training and testing were conducted in a specialized apparatus called a gustometer, described in detail elsewhere (Spector et al. 2015). Briefly, the testing chamber consists of a rectangular plastic cage with a wire mesh floor and a stainless-steel front panel. Three vertical access slots are positioned in the stainless-steel panel. Taste stimuli are deposited onto a borosilicate glass sample ball, which spins around a horizontal axis behind the central access slot, via Teflon tubing connected to syringes mounted to stepping motors. When an animal licks the sample ball twice within 250 ms, a preload of ~10 µl of the sample is deposited to coat the sample ball, and an additional ~5 µl of sample fluid is delivered with each lick. A force transducer connected to the sample ball registers licks that are then recorded by a computer executing trial contingencies. Between trials, the sample ball is retracted into a washing well, cleaned with DW, and dried with pressurized air. The reinforcer is delivered through a tube penetrating the center of a fixed response ball positioned behind each access slot to either side of the sample ball. Each ball registers licks as described for the sample ball. When necessary, access to any ball can be manually blocked by stainless-steel shutters. The cage and fluid delivery system are housed within a sound-attenuating chamber. Background noise is presented during the session to reduce extraneous auditory signals. A stainless-steel shield blocks everything from view except the sample ball during sampling, to limit potential visual cues. Air is drawn away from the sample ball and out of the chamber via a plastic duct connected to an exhaust fan to limit extraneous olfactory cues.

Trial structure

All detection training and testing phases were conducted in 30-min sessions using a 2-response operant task with a modified method of limits procedure similar to that previously described (Blonde et al. 2006). In this task, a thirsty animal samples a small volume of stimulus and must quickly determine how to respond based on which stimulus is presented. Correct responses result in a water reward. The rat initiates a trial by licking the clean, dry sample ball twice within 250 ms to ensure active engagement in the task. Following a sample phase of 10 licks or 3 s, the sample ball is retracted. The rat then has a short duration (limited hold) to respond by licking 1 of the 2 response balls. Correct responses are reinforced with 20 licks (within 5 s) of DW, and incorrect or no responses result in a time-out with no water delivered. The durations of the limited hold and the time-out vary during training (see descriptions in Operant training). The sample ball is cleaned and dried with DW during a ~7-s intertrial interval and then repositioned in preparation for a new trial.

Operant training

All rats were first trained and tested in a NaCl detection task, requiring the animal to discriminate a single concentration of NaCl from DW, beginning with a training concentration of 0.4 M NaCl. Once the animals were trained using a high concentration of the stimulus, the concentration of the target stimulus (dependent on phase) was decreased until the animal could no longer discriminate the stimulus from its background. See Table 1 for a summary of training and testing.

Lick training

In the first phase of training, animals were given ad libitum access to DW from 1 source for a full 30-min session. Access to the other 2 fluid sources was blocked with shutters. The available source rotated across sessions (sample, left, right, in that order).

Side training

The rats were then trained to associate 1 response ball with 1 sample stimulus (e.g., to lick the left response ball if the sample is NaCl). In each session, the rat was presented with a sample (10 licks or 3 s) of a stimulus and had up to 180 s (limited hold) to lick the response ball associated with that stimulus. The other response ball was blocked by a shutter, and the stimulus associated with the other response ball was not presented as a sample. During side training no time-out was delivered if the rat failed to respond. Stimulus assignments were counterbalanced so that approximately half of the animals were assigned NaCl to the left response ball, and the other half were assigned NaCl to the right response ball. The stimulus presented was alternated across sessions, with 2 sessions for both 0.4 M NaCl and DW.

Alternation

Both stimuli were presented within a single session for the first time during this phase of training, and the rat had access to both response balls. One stimulus (e.g., 0.4 M NaCl) was presented as the sample until the rat had responded correctly to a criterion number of nonconsecutive trials, whereupon the other stimulus (e.g., DW) was presented until the criterion was met. The criterion was subsequently lowered each day (criteria: 8, 6, 4, and 2). In this phase, the limited hold was lowered to 15 s, and rats were given a 10-s time-out if they responded incorrectly or failed to respond.

Random training

In the final phases of training and for all testing phases, stimuli were presented in randomized (without replacement) blocks of trials (block size: 4–6) with a probability of stimulus presentation of 0.5. Trial parameters were progressively changed to the final values. The limited hold was decreased to 10 s for 2 sessions. Then, the time-out was increased to 20 s for 2 sessions. At this point, all rats were performing at ≥80% overall on trials with a response, indicating a high level of performance (proportion of trials with correct responses). The first phase of operant testing (NaCl in Water I; see Detection testing: NaCl in water, Table 1) then began.

Interphase training

When the task was changed from one stimulus set to another (see Table 1), the rats were first given 2 sessions of side training (1 session with each stimulus for the new task) and then as many sessions of Random Training as necessary for every rat to perform ≥80% correct overall on trials with a response. This occurred in 2–4 sessions, whereupon the testing phase for that stimulus set began. The limited hold and time-out durations remained at the final reported values (10 and 20 s, respectively) for all remaining phases of training and testing because the rats were familiar with the task parameters.

Detection testing: NaCl in water

All rats were first tested in a NaCl detection task, which required the animal to detect the presence of NaCl when dissolved in water. Once testing began, the concentration of NaCl used as a stimulus was systematically lowered by approximately 0.33 log₁₀ units from the training concentration (0.4 M NaCl), across sessions, until all rats had reached ~50% correct overall (i.e., chance levels of performance for this task) on trials with a response. As long as every rat had performed ≥80% on trials with a response for the concentration tested, the test concentration decreased for the next test session. Once any rat performed below 80% correct on a concentration, all rats were given stimulus control sessions, interspersed between test sessions. During control sessions, each rat was tested with the lowest concentration to which it had consistently performed at least 80% correct overall on trials with a response; in this way, stimulus control was tailored to the sensitivity of the individual animal. All session parameters were set as described for Random Training.

Detection testing: masking concentration series

Once all rats had reached chance levels of performance with NaCl in water, the stimulus set was changed to determine the detectability of NaCl (the target stimulus) in 0.3 M sucrose (the masking stimulus), beginning with 0.4 M NaCl as in the previous detection task. That is, instead of being tasked with detecting NaCl in water, the rat was given either 0.4 M NaCl dissolved in 0.3 M sucrose as a sample (0.4 M NaCl + 0.3 M sucrose), or the sample was 0.3 M sucrose. This concentration of sucrose was chosen because previous work has shown that rats will respond to a mixture of similar NaCl and sucrose concentrations as sharing similar qualitative characteristics to NaCl or sucrose presented separately (Grobe and Spector 2008), indicating that neither chemical is fully masked by the other at these suprathreshold concentrations.

As the target stimulus, NaCl (in 0.3 M sucrose) remained assigned to the same response ball as in the detection task. The response ball previously assigned to DW was assigned to 0.3 M sucrose in this phase. Animals were trained as described above for Interphase Training. Testing was conducted as described above for NaCl in water.

After testing the detectability of NaCl in 0.3 M sucrose, the rats were split into 2 groups. One half of the rats continued being tested for the detectability of NaCl, with increasing concentrations of sucrose as the masker (0.6 and 1.0 M sucrose). Following those phases, NaCl sensitivity (in water) was reassessed (Table 1). The same concentrations of NaCl, beginning with 0.4 M NaCl, were used throughout the experiment, with each phase ending when every rat had reached chance levels (~50% correct overall on trials with a response) on at least 1 concentration of NaCl.

The remaining rats were instead tasked to discriminate sucrose (now the target stimulus) from water in a detection task as done with NaCl. Sucrose was assigned to the response ball previously associated with NaCl (the previous target stimulus). This sensitivity assessment was followed by phases detecting sucrose masked by 0.4 M NaCl, 0.2 M NaCl, and finally 0.04 M NaCl (Table 1). The training concentration of sucrose was 0.3 M, and it was used to begin every phase wherein sucrose was the target stimulus. Sucrose concentrations were decreased during testing by approximately 0.3 log₁₀ units, and the same sucrose concentrations were used in each phase until every rat had reached chance levels of performance (~50% correct overall) on trials with a response.

Negative control tests

Following the last phase of testing for each set of rats, a negative control test was conducted wherein 6 reservoirs were filled with the vehicle used in the last phase (see Table 1). Half of the reservoirs were assigned as the vehicle, and the other half were assigned as the target. For the rats being tested with NaCl as the target stimulus, water was the single stimulus in this control test because the previous phase was detection of NaCl in water. For the rats being tested with sucrose as the target stimulus, 0.04 M NaCl was the single stimulus in this control test because the previous phase was detecting sucrose in 0.04 M NaCl. The purpose of this test was to verify that the performance was based on the chemical cues of the stimulus rather than others that may have been present during testing (e.g., visual or auditory cues from the machine). It was expected that all rats would perform at chance levels (50% correct overall).

Data analysis

Overall proportion correct on trials with a response (i.e., performance) was collapsed across sessions for each concentration for each animal. Data from all test and control sessions were included in analyses. Performance between phases was compared using 2-way analyses of variance (ANOVAs; phase × concentration).

A psychometric function was fit to the performance of each rat in each phase, using the logistic equation:

$${\displaystyle \begin{array}{l}{\displaystyle f(x)=\left[\frac{a-0.5}{(1+{10}^{(\mathrm{l}\mathrm{o}{\mathrm{g}}_{10}x-c)b})}\right]+0.5}\end{array}}$$

where x = the molar concentration, a = the performance asymptote, b = the slope, and c = the log₁₀ molar concentration for the target at ½ of asymptotic performance. The c-value, or the EC₅₀ value, was operationally defined as the detection threshold. When comparing the effect of the masker concentration (across masking phases), curve parameters were analyzed using paired-sample t-tests.

We calculated an area under the curve (AUC) value using overall proportion correct for trials with a response at each concentration, for each rat for each phase. To calculate the AUC we summed the difference between the proportion correct at each concentration and 0.5 (chance performance). Negative values (i.e., when performance was less than 0.5) were set to 0.5 for these calculations. Proportional change in AUC was calculated by subtracting the AUC when the target was dissolved in water from the AUC calculated in each masking phase and then dividing the difference by the AUC of the target in water:

$${\displaystyle \begin{array}{l}{\displaystyle \frac{(\mathrm{A}\mathrm{U}\mathrm{C}\mathrm{i}\mathrm{n}\mathrm{m}\mathrm{a}\mathrm{s}\mathrm{k}\mathrm{e}\mathrm{r}-\mathrm{A}\mathrm{U}\mathrm{C}\mathrm{i}\mathrm{n}\mathrm{w}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r})}{\mathrm{A}\mathrm{U}\mathrm{C}\mathrm{i}\mathrm{n}\mathrm{w}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}}}\end{array}}$$

These AUC values and proportional changes in AUC were compared across phases by paired-sample t-test. Proportional changes in AUC were also compared with a mean of zero change (e.g., no difference from when the target was dissolved in water) using 1-sample t-tests.

In an effort to check for the influence of the adaptation state of the receptors on behavioral performance, sometimes reported in mixture studies (Kroeze 1979, 1982; Gillan 1982; Lawless 1982), the subset of trials with a response in each session that followed a correct response (and therefore, a water reinforcement) was compared with trials with a response that followed an incorrect response (and therefore, no water “rinse”). In those cases, overall proportion correct was calculated for each subset of trials, collapsed across sessions for each concentration for each rat. These proportions were used to calculate the AUC for trials after reinforcement, and a proportional change in AUC for trials after reinforcement and for trials after no reinforcement was calculated as described for all trials. In the context of the methodology used here, the water reinforcement following correct responses acted as a rinse that might reduce adaptation. The proportional shifts in AUC from trials after reinforcement (relative to AUC from trials after reinforcement during target detection in water) were then compared with the proportional shifts in AUC from trials after no water reinforcement (relative to AUC during target detection in water). The AUC values were then compared using paired-sample t-tests.

Four rats were removed from study during training due to illness. Their data are not included in any analyses. The final group sizes are: rats tested with NaCl masked by sucrose, n = 8; rats tested with sucrose masked by NaCl, n = 8. Statistical significance was considered to be P ≤ 0.05.

Results

Negative control tests

For all rats, the performance of the negative control test was not significantly different from chance performance (50%) based on a binomial distribution. For rats tested with NaCl as a target, this test was conducted with 6 reservoirs filled with water (mean: 50.7 ± 1.3%). For rats tested with sucrose as the target, this test was conducted with 6 reservoirs filled with 0.04 M NaCl, the masker used in the last phase (mean: 51.2 ± 0.7%). These results were expected and indicate that the performance during masking phases was based on chemical signals from the stimuli, and not extraneous cues.

NaCl masked by sucrose

The second phase of NaCl detectability in water (NaCl in Water II; Table 1) was conducted because, in some cases, repeated testing can result in improved performance in this task and a significant difference in curve parameters (e.g., Blonde et al. 2006). There was no significant effect on curve parameters between the 2 phases of NaCl in Water (Table 2), indicating little change in sensitivity as a function of time or experience. However, ANOVAs on the proportion correct revealed a significant effect of phase, concentration and an interaction, with better performance in NaCl in Water II, indicating some effect of experience albeit not enough to shift psychometric functions. For this reason, data from all phases with sucrose as a masking background stimulus were compared with those from the NaCl in Water I phase.

Table 2.

Logistic function parameters by phase, compared with target detection in water

NaCl masked by sucrose
	a a		b a		c a
	Mean (SE)	t-Test	Mean (SE)	t-Test	Mean (SE)	t-Test
NaCl in Water I	0.96 (.01)	N/A	−1.04 (.09)	N/A	−2.62 (.13)	N/A
NaCl in 0.3 M sucrose	0.95 (.01)	t 7 = 0.26	−1.10 (.14)	t 7 = 2.10	−2.34 (.13)	t 7 = 5.21
		P = 0.80		P = 0.07		P < 0.01
NaCl in 0.6 M sucrose	0.91 (.04)	t 7 = 1.22	−13.74 (6.44)	t 7 = 2.02	−2.42 (.12)	t 7 = 4.71
		P = 0.26		P = 0.08		P < 0.01
NaCl in 1.0 M sucrose	0.97 (.01)	t 7 = 1.27	−1.65 (.33)	t 7 = 2.14	−2.20 (.09)	t 7 = 5.50
		P = 0.25		P = 0.07		P < 0.01
NaCl in Water II	0.96 (0.1)	t 7 = 0.34	−0.85 (.09)	t 7 = 1.24	−2.92 (.15)	t 7 = 2.10
		P = 0.74		P = 0.26		P = 0.07
Initial phases for rats later tested with Sucrose masked by NaCl
NaCl in Water I	0.97 (.01)	N/A	−6.28 (5.11)	N/A	−2.53 (.12)	N/A
NaCl in 0.3 M sucrose	0.93 (.01)	t 7 = 2.55	−1.77 (.56)	t 7 = 0.87	−2.35 (.07)	t 7 = 2.48
		P > 0.05		P = 0.41		P = 0.04
Sucrose masked by NaCl
	a a		b a		c a
	Mean (SE)	t-Test	Mean (SE)	t-Test	Mean (SE)	t-Test
Sucrose in Water	0.91 (.01)	N/A	−3.35 (.82)	N/A	−1.83 (.04)	N/A
Sucrose in 0.04 M NaCl	0.96 (.01)	t 7 = 3.68	−2.17 (.33)	t 7 = 1.35	−1.69 (.05)	t 7 = 2.02
		P < 0.01		P = 0.22		P = 0.08
Sucrose in 0.2 M NaCl	0.96 (.02)	t 7 = 1.97	−1.43 (.21)	t 7 = 2.22	−1.49 (.05)	t 7 = 5.29
						P < 0.01
		P = 0.09		P = 0.06
Sucrose in 0.4 M NaCl	0.96 (.02)	t 7 = 2.32	−1.56 (.18)	t 7 = 2.02	−1.16 (.06)	t 7 = 15.21
		P > 0.05		P = 0.08		P < 0.01

Bolded values indicate a significant difference from Target in Water (P < 0.05) SE, standard error.

^a a = the performance asymptote, b = the slope, and c = the log₁₀ molar concentration for the target at ½ of asymptotic performance. The c-value of the logistic function is also the EC₅₀ value, operationally defined as the detection threshold.

In general, NaCl sensitivity was decreased when tested in the context of a sucrose background as indicated by a rightward shift in the psychometric function (Figure 1). The ability of rats to perform the discrimination task was not affected, with no difference in the asymptotic performance (a-parameter) for any mask concentration (Table 2). However, the EC₅₀ (c-parameter), operationally defined as the detection threshold, was significantly increased relative to the baseline detection phase (NaCl in Water I; Table 2). Thus, while the animals were able to discriminate the training concentration of the target from the masker in each phase, sensitivity to the target was blunted such that thresholds were higher when a background stimulus was present. The slope of the functions (b-parameter) was not different from NaCl in Water I (Table 2). Overall, the addition of a background masking concentration affected detectability, with ANOVAs revealing significant main effects of both mask and target concentration, as well as significant mask concentration × target concentration interactions, when comparing the performance under each mask concentration to that under no mask (i.e., NaCl in Water; Table 3).

Figure 1.

Psychometric functions by masking phase. Mean (± standard error) performance and psychometric functions are calculated for each phase for rats tested with NaCl as the target, masked by sucrose (left panel), and for rats tested with sucrose as the target, masked by NaCl (right panel). Data are presented for the target in water (white; solid lines), and the target in low (light gray symbols; short dashed lines), medium (medium gray symbols; medium dashed lines), and high (dark gray symbols; long dashed lines) masker concentrations. Note the different scale range on the horizontal axis for the left versus the right panel because the dynamic range of sensitivity for NaCl and sucrose differ.

Table 3.

ANOVAs comparing performance across phases: NaCl masked by sucrose

	Compared with
	In 0.3 M sucrose	In 0.6 M sucrose	In 1.0 M sucrose	NaCl in Water II
NaCl in Water I
Phase	F 1,7 = 35.70	F 1,7 = 22.39	F 1,7 = 5.80	F 1,7 = 9.45
	P < 0.01	P < 0.01	P < 0.05	P = 0.02
Concentration	F 10,70 = 75.08	F 9,63 = 61.31	F 7,49 = 49.69	F 11,77 = 84.71
	P < 0.01	P < 0.01	P < 0.01	P < 0.01
Phase × Conc	F 10,70 = 3.17	F 9,63 = 7.98	F 7,49 = 8.10	F 11,77 = 4.93
	P < 0.01	P < 0.01	P < 0.01	P < 0.01
NaCl in 0.3 M sucrose
Phase		F 1,7 = 0.50	F 1,7 = 2.76	F 1,7 = 133.35
		P = 0.50	P = 0.14	P < 0.01
Concentration		F 9,63 = 62.26	F 7,49 = 65.56	F 10,70 = 52.41
		P < 0.01	P < 0.01	P < 0.01
Phase × Conc		F 9,63 = 5.11	F 7,49 = 1.87	F 10,70 = 5.85
		P < 0.01	P = 0.09	P < 0.01
NaCl in 0.6 M sucrose
Phase			F 1,7 = 0.11	F 1,7 = 68.88
			P = 0.75	P < 0.01
Concentration			F 7,49 = 47.06	F 9,63 = 48.31
			P < 0.01	P < 0.01
Phase × Conc			F 7,49 = 4.07	F 9,63 = 7.81
			P < 0.01	P < 0.01
NaCl in 1.0 M sucrose
Phase				F 1,7 = 5.49
				P = 0.05
Concentration				F 7,49 = 35.12
				P < 0.01
Phase × Conc				F 7,49 = 8.45
				P < 0.01

These statistics only include data from rats tested in all phases of NaCl masked by sucrose.

Bolded values indicate statistical significance (P < 0.05).

While the use of a background decreased NaCl detectability, increasing the sucrose concentration used as a masker did not necessarily compromise performance to all concentrations of the target stimulus similarly. ANOVAs revealed no significant main effect of sucrose mask concentration between any of the 3 masking phases (Table 3). However, there were significant Phase × Concentration interactions between 0.6 M sucrose and each of the other sucrose maskers, indicating that there was some difference at some NaCl concentrations (Table 3). These results demonstrate that the masking stimulus can impair detectability of the target stimulus in a concentration-dependent manner—as the masker concentration increases, the target detectability decreases.

The rats tested with sucrose masked by NaCl were initially tested with NaCl in water, then NaCl in 0.3 M sucrose. The performance of these rats detecting NaCl, like that of those tested with NaCl throughout the experiment, was significantly affected by the 0.3 M sucrose mask (F_1,7 = 24.81; P < 0.01), target concentration (F_10,70 = 100.52; P < 0.01), and displayed a significant interaction (F_10,70 = 3.462; P < 0.01). Because these rats were only tested with 1 masking phase with NaCl as the target, their data are not included in any further analyses for NaCl masked by sucrose. Their curve parameters are presented in Table 2.

Sucrose masked by NaCl

The first NaCl masking concentration chosen (0.4 M) was very effective at reducing sucrose detectability (Figure 1). Thus, the concentration was decreased across phases.

For all masking phases, rats were able to learn to perform the task, with asymptotes similar to or higher than those derived for sucrose testing without a mask (Sucrose in Water; Table 2). However, for the 2 higher NaCl masking concentrations (0.2 and 0.4 M NaCl), sensitivity to sucrose was significantly lower than that seen when sucrose was dissolved in water (Table 2 and Figure 1). The lowest masking concentration (0.04 M NaCl) did not appear to significantly affect sucrose sensitivity (Table 2). As with NaCl masked by sucrose, the slope of the psychometric functions did not differ from those derived without a mask present (Table 2).

Overall, ANOVAs demonstrated that the concentrations of NaCl chosen were able to significantly reduce detectability of sucrose, relative to performance during the Sucrose in Water phase as well as to each other (Table 4). The only exception is that there was no effect of phase between Sucrose in Water and Sucrose in 0.04 M NaCl, although there was a significant interaction suggesting that the performance was impaired on some concentrations (Table 4). As with NaCl in sucrose, it would appear that sucrose detectability decreases as a function of NaCl masking concentration.

Table 4.

ANOVAs comparing performance across phases: Sucrose masked by NaCl

	In 0.04 M NaCl	In 0.2 M NaCl	In 0.4 M NaCl
Sucrose in water
Phase	F 1,7 = 0.03	F 1,7 = 8.00	F 1,7 = 88.49
	P = 0.88	P = 0.03	P < 0.01
Concentration	F 6,42 = 102.75	F 6,42 = 109.60	F 5,35 = 77.51
	P < 0.01	P < 0.01	P < 0.01
Phase × Conc	F 6,42 = 4.64	F 6,42 = 5.65	F 5,35 = 9.03
	P < 0.01	P < 0.01	P < 0.01
Sucrose in 0.04 M NaCl
Phase		F 1,7 = 12.70	F 1,7 = 226.40
		P < 0.01	P < 0.01
Concentration		F 6,42 = 141.60	F 5,35 = 126.79
		P < 0.01	P < 0.01
Phase × Conc		F 6,42 = 10.64	F 5,35 = 5.59
		P < 0.01	P < 0.01
Sucrose in 0.2 M NaCl
Phase			F 1,7 = 35.59
			P < 0.01
Concentration			F 5,35 = 129.66
			P < 0.01
Phase × Conc			F 5,35 = 3.98
			P < 0.01

Bolded values indicate statistical significance (P < 0.05).

Comparison of masker effectiveness

Based on the performance curves, it is tempting to interpret the data to mean that NaCl is more effective at masking sucrose, than sucrose is at masking NaCl. However, it is important to remember that sensitivity to these 2 stimuli are different in rats, with the threshold (c-value) for NaCl being reported as nearly an order of magnitude lower than that for sucrose (Table 2). Thus, equimolar concentrations of NaCl and sucrose might have very different perceived intensities, with NaCl being stronger. Indeed, such a difference in suprathreshold intensity ratings has been demonstrated in humans (Beebe-Center et al. 1959). To effectively compare NaCl and sucrose as masking stimuli, then, we plotted the shift in EC₅₀ values for the target against the change in masking concentration relative to threshold for the masker. The mean baseline EC₅₀ value for NaCl and sucrose (from NaCl in Water and Sucrose in Water) was compared with the masking concentrations used for each chemical, respectively. Thus, the masking concentration was represented as a difference from the threshold concentration of the mask, and these values were then used to plot the shift in EC₅₀ for the target in each mask relative to no mask (Figure 2). The relative shift in masking concentrations (compared with threshold for the masker) was not perfectly matched for NaCl and sucrose, but there was significant overlap across the 2 stimuli, though direct statistical comparison is not possible. Nevertheless, the 2 stimuli appear similarly effective at masking each other (Figure 2), once the relative sensitivity to the mask is taken into account.

Figure 2.

Change in EC₅₀ values (relative to no mask condition) as function of the concentration of the mask relative to the sensitivity to the mask. Mean (± standard error) change in EC₅₀ value of psychometric functions by phase, relative to the EC₅₀ value calculated for the target in water. Data are plotted against the difference in masker concentration from each phase, relative to the EC₅₀ value (detection threshold) calculated for the masker in water. The sucrose masker concentrations for rats tested with NaCl in sucrose (square symbols) are plotted relative to the EC₅₀ value of sucrose in water derived from performance data of rats being tested with sucrose as the target. The masker concentrations for rats tested with sucrose in NaCl (circle symbols) are plotted relative to the EC₅₀ value of NaCl in water derived from performance data of rats being tested with NaCl as the target.

AUCs

While threshold measurements can depict changes in liminal sensitivity, it has been reported that mixture quality changes dramatically depending on the relative concentrations of components (see Keast and Breslin 2003). Using performance values at each concentration may prove informative by incorporating minor changes across the dynamic range of sensitivity. To graphically depict the changes across all tested concentrations of the target, the AUC was calculated using performance values for each rat in each phase (Figure 3). We tested whether the proportional change in the AUCs for a given mask relative to the AUC when there was no mask was different from zero (i.e., no proportional change using 1-sample t-tests). All masking phases for both NaCl and sucrose demonstrated a significant decrease in AUC relative to performance when the target was dissolved in water (i.e., no mask; Figure 3). Also, with the exception of 0.3 and 0.6 M sucrose as maskers, there was a decrease in the AUC as masker concentration increased, resulting in a larger proportional change in AUC (Table 5).

Figure 3.

Proportional changes in AUC. Mean and individual changes in area of the curve (AUC) are presented for NaCl as the target (left panel) and when sucrose was the target (right panel), calculated relative to AUC from performance data when the target was dissolved in water. AUCs (mean ± standard error) for each target when dissolved in water are reported in the upper left corner of each panel. * represents a significant shift from AUC from Target in Water, as determined by 1-sample t-tests comparing the proportional change to zero (i.e., no change).

Table 5.

Paired-sample t-tests comparing proportional change in AUC

NaCl in sucrose
	In 0.3 M sucrose	In 0.6 M sucrose	In 1.0 M sucrose
In 0.6 M sucrose	t 7 = 0.50
	P = 0.64
In 1.0 M sucrose	t 7 = 2.71	t 7 = 4.00
	P ≤ 0.01	P < 0.01
NaCl in Water II	t 7 = 18.05	t 7 = 24.37	t 7 = 16.60
	P < 0.01	P < 0.01	P < 0.01
Sucrose in NaCl
	In 0.04 M NaCl	In 0.2 M NaCl
In 0.2 M NaCl	t 7 = 11.50
	P < 0.01
In 0.4 M NaCl	t 7 = 16.99	t 7 = 3.14
	P < 0.01	P < 0.01

Bolded values indicate statistical significance (P < 0.05). Bonferroni correction did not change any significant outcomes.

Further, once all changes in performance are assessed together (asymptote, slope, and threshold), it does appear that sucrose sensitivity is more disrupted by NaCl than NaCl sensitivity is disrupted by sucrose (Figure 3). The proportional change in AUC relative to performance without a masker (Sucrose in Water) is larger for all 3 masking phases than the changes in AUC for any phase with NaCl as the target. That is, despite there being little difference in the change in liminal sensitivity between these 2 stimuli when the sensitivity of the mask is accounted for (Figure 2), the overall decrease in performance across the entire concentration range for sucrose masked by NaCl is larger than when NaCl is masked by sucrose (Figure 3).

AUC: trials after reinforcement versus trials after no reinforcement

One interpretive consideration in masking studies is that constant exposure to the background may result in receptor adaptation to the masking chemical (Kroeze 1979, 1982; Gillan 1982; Lawless 1982), thereby reducing its effectiveness across the session. One way to determine the possibility of such a neural mechanism in the task employed here, is to assess performance after a correct response in which a water reinforcer, effectively serving as a rinse, was delivered. In that case, should adaptation be occurring, performance on trials following reinforcer delivery should decrease because the masker should be more effective. Similarly, the proportional change in AUC should be larger for trials following reinforcement than when a rinse was not delivered.

Baseline phase AUCs (from NaCl in Water and Sucrose in Water) are reported in Figure 4. There was a significant difference between trials after no reinforcement and trials after reinforcement, in both the NaCl in Water I (t = 4.37, P < 0.01) and Sucrose in Water (t = 3.94; P < 0.01) phases. For both sucrose and NaCl, the AUC for trials following reinforcement was slightly larger (i.e., performance was better).

Figure 4.

Proportional changes in AUC for trials after reinforcement and for trials after no reinforcement. Mean (− standard error) proportional change in AUC for each phase with sucrose as the target (top) and NaCl as the target (bottom). Both the AUC from trials with a response after reinforcement (black bars) and from trials with a response following no reinforcement (gray bars) are calculated relative to performance in the comparable trials when the target was dissolved in water. Mean (− standard error) AUC when the target was dissolved in water is reported in the lower left corner for each target stimulus.

In masking phases, there was almost no difference in proportional change in AUC, relative to no mask, between trials after reinforcement and trials following no reinforcement (Figure 4). Paired-sample t-tests indicated no significant difference between the proportional change in AUC on trials after reinforcement versus trials after no reinforcement, regardless of whether a Bonferroni correction was used.

Discussion

Overall, this methodology demonstrates a systematic effect of masking concentration on the detectability of the target stimulus. Increasing the concentration of the masker incrementally decreased detectability of the target along the concentration ranges tested here, even if changes across masking concentration were only slight. The degree of effectiveness of the masker depended on whether the analysis focused on the relative shifts of the psychometric function as assessed by EC₅₀ or whether the performance across the entire concentration series was considered as assessed by the AUC. In either case, however, it is clear that the detection of a target in water is easier than when the target is dissolved in another chemical solution (i.e., the mask).

Changes in liminal sensitivity to the target appear related to the liminal sensitivity for the masker

With only the exception of the lowest concentration of NaCl, all masker concentrations used in this study were effective at increasing the detection threshold for the target, as evidenced by a rightward shift of the EC₅₀ (c-parameter) of the psychometric function, relative to the curve for the target when dissolved in water (Figure 1 and Table 2). The EC₅₀ lies at the inflection point in the derived curve and is in a prime location to describe lateral shifts in the overall curve. The effectiveness of each compound to compromise detectability of the other compound is, at first comparison, not the same (Figure 1). It certainly would appear that the detectability of NaCl is less impacted by the addition of sucrose, than is the detectability of sucrose by the addition of NaCl. However, it is important to remember that the liminal sensitivity to NaCl and sucrose are very different. The detection threshold for NaCl, for the rats tested in this study, was on average 2.5 mM, while the average EC₅₀ for sucrose was approximately 6× higher (Table 2). As such, equimolar concentrations of the 2 stimuli may be perceived to have very different intensities. It is true that the first concentrations chosen for maskers (0.4 M NaCl and 0.3 M sucrose) are similar to those that create a mixture reported by rats to be both “sodium-like” and “sucrose-like” in generalization tasks (Grobe and Spector 2008). However, the generalization profiles in that study only indicate that both stimuli contribute to the quality of the stimulus, without being informative as to the relative intensity of either component in the mixture. Indeed, it is clear that the chosen starting masker concentrations (i.e., 0.4 M NaCl and 0.3 M sucrose) affected the EC₅₀ of the other stimulus to different extents (Figure 1). For this reason, the masking concentrations used for NaCl were decreased across phases of sucrose detection, while the masking concentrations used for sucrose were increased across phases of NaCl detection (Table 1).

The concentration range used for the maskers was then compared with detectability of the masker in water (i.e., the EC₅₀ of the masker when tested as the target stimulus). The difference between masker threshold and each masker concentration was used to plot the shift in EC₅₀s (i.e., c-values) for the target when tested in the masker (Figure 2). Interestingly, the liminal sensitivity for either target was similarly affected across the overlapping range of masking concentrations when considered relative to the masker threshold. While direct comparison was not possible since the masker concentrations were not perfectly matched, it appears that these 2 stimuli are similarly capable of masking the barely perceptible presence of the other chemical.

That either stimulus used in this study would similarly affect the liminal sensitivity of the other is perhaps not surprising. These 2 stimuli were chosen based on their clear discriminability. Not only do they evoke clearly different qualities in humans, but in generalization studies, rodents do not treat NaCl like sucrose or vice versa (Morrison 1967; Nowlis et al. 1980; Grobe and Spector 2008; Gautam et al. 2012). It is also known that, in numerous electrophysiological studies in rodents, sucrose and NaCl evoke different activity in nerve recordings and generate very different profiles of responding across taste neurons along the entire gustatory neuraxis (e.g., (Frank et al. 1983; Travers and Smith 1984; Frank 1991; Nakamura and Norgren 1991; Vogt and Smith 1993; Nishijo and Norgren 1997; Spector and Travers 2005).

Mixture suppression, defined as a response to the mixture that is less than the sum of responses to the individual components, can be displayed in neuronal activity (Hyman and Frank 1980a; Chen and Di Lorenzo 2008). However, some single-fiber studies have demonstrated that the responses to sucrose + NaCl mixtures mimic the responses to the “best” stimulus for the nerve fiber (Hyman and Frank 1980b; but see Travers and Smith 1984; Formaker and Frank 1996). This result would suggest that, at least in the periphery and hindbrain gustatory nuclei, the signals from each component are still recognizable while in mixture. While little work has been conducted with NaCl/sucrose mixtures in higher-order gustatory nuclei in rodents, what has been done describes similar results (Maier and Katz 2013). Given that these 2 chemicals do not seem to directly interact in solution to interfere with the signals arising from the oral cavity, mixtures of NaCl and sucrose likely cause mixture suppression to similar extents because each is being influenced by general processes, not currently understood, related to mixtures.

That is not to say, though, that the presence of potentially discriminable signals from the periphery representing different compounds negates the role of peripheral processes in mixture suppression. When NaCl and sucrose are presented on the tongue at the same time but to different areas (referred to as “split tongue” methodology), mixture suppression of quality intensity ratings for NaCl and sucrose is reported, but is reduced. As distance between placements of the components is reduced, mixture suppression increases (Gillan 1982). These results, while not ruling out a more central origin, implicate peripheral processes in the mixture suppression seen with NaCl and sucrose mixtures, though not necessarily for other stimulus pairings (e.g., Kroeze and Bartoshuk 1985).

Overall changes in performance were asymmetrical

While the detection threshold is a useful tool in determining sensitivity to the stimulus, it is only one point along the range of stimulus intensity. It describes the inflection point in detectability, but it does not reflect the overall changes in target intensity. In general, the other curve parameters, describing maximal performance (a-value) and the rate of change in performance (b-value) were not significantly shifted by the addition of any of the masking concentrations chosen in this study. The a-value, indicating asymptotic performance in the task, only significantly changed for sucrose mixed in 0.04 M NaCl (Table 2); in that case, the a-value was significantly higher than baseline (Sucrose in Water). This result may be due to an experience effect, given that the lowest concentration of NaCl masker was the last phase of testing for those animals. As for the slope of the functions, or b-value, changes during masking phases only approached significance in many cases (Table 2). However, while fitted psychometric functions are very useful in describing fluctuations in behavior, there are certain limitations inherent in their use. First, the a-value provides support that the task was not too difficult, but there is no information as to when asymptotic performance is reached along the concentration series. For instance, the rats detecting sucrose in NaCl reached a similar asymptote of performance in every phase (Table 2), but the concentration at which asymptote was reached is clearly different when examining their performance values (Figure 1). Some of that difference could be explained by a slower increase in performance with concentration (i.e., a shallow slope), although the b-values never significantly changed (Table 2). As such, small changes in each curve parameter may not reach significance, but fluctuations taken together may demonstrate a large effect across the entire concentration range.

Indeed, the results of the ANOVA tests suggest that there are differences across the entire concentration range when compared with the target dissolved in water, and often from one masking phase to the next (Tables 3 and 4). One way to visualize the overall changes in target intensity is by converting performance values into an AUC. As employed here, the AUC is calculated using performance values, not the psychometric functions, to assess the degree of difficulty of the task once a masker is introduced. For each phase, the addition of a masker significantly decreased the AUC for the target (i.e., a shift of zero; Figure 3). These results correspond to the results from the ANOVAs for each masking concentration relative to the no mask test condition (Tables 3 and 4). Comparisons of the AUCs suggest an asymmetry in the effectiveness of NaCl and sucrose to mask each other, akin to some suprathreshold quality intensity data from human studies (Kamen et al. 1961), and some theories have been proposed (see Wilkie and Capaldi Phillips 2014).

Adaptation seems to have little impact in this task

It is known that constant stimulation with a tastant results in sensory adaptation, reducing the intensity of the chemical over time (e.g., Meiselman 1968; Gent and McBurney 1978). With mixture studies, adaptation can be a useful tool for determining the effectiveness of one stimulus to mask another. It has been shown with a number of stimulus pairings that adapting a participant to the masker concentration results in a higher rating of target quality intensity than when not adapted to the masker. In other words, adaptation decreases mixture suppression on the target’s quality (Kroeze 1979, 1982; Gillan 1982; Lawless 1982). In most cases, cross-adapting to the mixture demonstrates a similar reduction in mixture suppression.

For this design, adaptation could be occurring with the masker because it is presented in all stimuli, either alone (in water) or in combination with the target. Should that constant presentation result in adaptation, it should decrease masker intensity and therefore increase the relative intensity of the target, making the task easier for the animal. Complete adaptation is unlikely to be occurring in the task, given that detection thresholds were shifted for nearly all masker phases (Figure 1 and Table 2). However, incomplete adaptation would reduce masker intensity without completely eliminating mixture suppression upon the target. Should adaptation to the masker have been a factor in performance for this task, the shift in AUC for trials after no reinforcement should have been smaller than the shift in AUC for trials following reinforcement, because the water reinforcement following correct responses would have acted as a rinse that would mitigate adaptation. However, upon analysis, there was no difference between the 2 trial sets for either target stimulus (Figure 4). Thus, it is unlikely that adaptation to the masker occurred in this mixture study, based on the lack of reduction in performance following a water rinse. It may be that the time between trials (~7 s) for the intertrial interval is sufficient to preclude any significant adaptation caused by the masker.

Psychophysical assessment of taste detection in chemical backgrounds

In general, the data presented here demonstrate a systematic effect of the concentration of a chemical background on the detectability of a target taste stimulus. As such, further use of this methodology could provide significant information with respect to the interactions of components of complex stimuli. Additional study of binary mixtures using different chemical combinations would indicate whether other stimulus pairings similarly shift liminal sensitivity for each other, when masking concentrations are matched relative to the masker detection threshold as appears to be the case with NaCl and sucrose (Figure 2). One expectation would be that other stimulus pairings shift sensitivity to each other at a similar rate, but perhaps at a different rate than NaCl and sucrose. For instance, it has been shown that not all sodium salts are equally effective at masking the “sweetness” of sucrose, and that KCl is less effective than most sodium salts (Keast et al. 2001). Perhaps, then, the shift in sucrose sensitivity would be less affected using other salts. On the other hand, sucrose is effective at reducing the “bitterness” of many bitter stimuli in humans, but to a different extent for each one tested (Mennella et al. 2015), suggesting that changing the masking concentration of sucrose would shift detection thresholds for bitter stimuli at a faster rate than seen here with NaCl. However, it is possible that sucrose would shift liminal sensitivity to bitter stimuli to different extents across different bitter chemicals, but to a similar extent that a given bitter stimulus impacts sucrose.

Psychophysical mixture studies have not been previously conducted in rodents. Instead, methodologies that rely on the suprathreshold recognition of the stimuli have typically been used (e.g., Nowlis et al. 1980; Smith and Theodore 1984; Grobe and Spector 2008; Gautam et al. 2012). While these tasks that rely on suprathreshold concentrations are useful in certain applications, they lack a quantitative assessment of the interaction of the mixture components on intensity processing. However, the particular methodology used here does not require recognition of NaCl or sucrose as such—it requires discriminability of the target (in the masker) from the background masker alone, regardless of the quality of the target stimulus.

Ultimately, this method is an extension of that used to study the detection of isolated chemicals (in water), a crucial part of understanding gustatory function. As in those studies, this design provides a systematic method for interrogating the gustatory system; here, with regard to stimulus detectability in a chemical background. It provides opportunities to manipulate different portions of the gustatory system to further reveal its functional organization. It also allows for determining functional parallels between detectability, discriminability, and recognition of stimuli. Other interesting phenomena described with human mixture work could also be studied with this methodology. For instance, individual differences are sometimes reported (e.g., Pangborn 1962; Gregson and McCowen 1963; Stevens 1996). That being said, the results of this study using NaCl and sucrose suggest a potential relationship between the components of mixtures—that is, the effectiveness of one chemical to interfere with detectability of another is based on its own detectability. By adding to the stimulus pairings used here, and moving to even more complex matrices, it could provide crucial information to understand the functional ability of the gustatory system to process complex stimuli as found naturally in our environments.

Funding

Research reported in this publication was supported by the National Institute on Deafness and Other Communication Disorders of the National Institutes of Health under award number R01-DC009821 (ACS). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Conflict of interest

The authors declare no conflicts of interest.