Chem. Senses 24: 23-35,
1999
© Oxford University Press
Taste Qualities of Solutions Preferred by Hamsters
Department of BioStructure & Function, School of Dental Medicine, University of Connecticut Health Center, Farmington, CT 06030 1 Department of Biological Science, York College, York, PA 17405, USA
Correspondence to be sent to: M.E. Frank, Department of BioStructure & Function, University of Connecticut Health Center, Farmington, CT 06030-3705, USA. e-mail: mfrank{at}neuron.uchc.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Molecules of diverse chemical structure are sweet to humans and several lines of evidence (genetic, physiological, behavioral) suggest that there may be distinct sweet perceptual qualities. To address how many perceptual categories these molecules elicit in hamsters (Mesocricetus auratus), we studied patterns of generalization of conditioned taste aversions for seven sweeteners: 100 mM sucrose, 320 mM maltose, 32 mM D-phenylalanine, 3.2 mM sodium saccharin, 16 mM calcium cyclamate, 10 mM dulcin and 32 mM sodium m -nitrobenzene sulfonate. Each stimulus was preferred versus water in two-bottle intake tests and stimulated the chorda tympani nerve. For each of seven experimental groups the conditional stimulus (CS) was a sweetener and for the control group the CS was water. Apomorphine·HCl was injected i.p. after a CS was sampled and, after recovery, test stimuli (TS) were presented for 1 h daily. The intake (ml) of each TS consumed by experimental animals was compared with mean TS intake by the control group. Learned aversions for 18/21 stimulus pairs cross-generalized, resulting in a single cluster of generalization patterns for the seven stimuli. Cross-generalization failures (maltose–cyclamate, maltose–sucrose, cyclamate–NaNBS) may be the consequence of particular stimulus features (e.g. salience, cation taste), rather than the absence of a `sucrose-like' quality. The results are consistent with a single hamster perceptual quality for a diverse set of chemical structures that are sweet to humans.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Many compounds with diverse chemical structures are preferred by hamsters (Mesocricetus auratus) in two-bottle intake tests (Rehnberg et al., 1990
).
Most of the compounds have sweet or bitter-sweet tastes to humans (Moncrieff, 1967;Walters et al., 1991). Solutions that have pleasant
odors to humans are not preferred by hamsters unless a sweet taste is added (Rehnberg et al. 1990). Based partly on the diversity of chemical structure
but also on the interpretation of genetic, physiological and behavioral data (Jakinovich
and Sugarman, 1988; Ninomiya and Funakoshi, 1993; Schiffman and Erickson, 1993;Froloff et al., 1996), transduction of sweet
stimuli is thought to involve multiple mechanisms
specific for compounds with distinct structural features (Walters et al., 1991). However, a multi-point attachment (MPA) model proposes for humans a single
sweet-taste receptor that is composed of eight recognition sites and 15 points of interaction
between ligand and receptor (Nofre and Tinti, 1996). It is derived from
the AH,B proton donor-acceptor model of Shallenberger and Acree (1967)
as modified by Kier (1972) to account for stereochemical specificity of
sweet amino acids. Six constant and two variable recognition sites of the MPA receptor account
for species differences among primates (Nofre et al., 1996).
Furthermore, although distinct second messenger pathways for sugar and non-sugar sweeteners
may exist in taste receptor cells, both cAMP and IP3 pathways occur in each cell (Lindemann, 1996). Thus, multiple transduction pathways for structurally
diverse sweeteners may converge at the level of a taste receptor cell. That receptor cell may
generate identical signals within the nervous system, providing no cue for distinct behaviors
toward the sweeteners.
We studied cross-generalization of conditioned taste aversions (CTA) to identify the number of
perceptual categories that hamsters may use for diverse sweeteners. The seven
stimuli—sucrose, sodium m-nitrobenzene sulfonate, sodium saccharin,
phenylalanine, dulcin, maltose and calcium cyclamate—are preferred in two-bottle, 48 h
intake tests (Rehnberg et al., 1990). We chose compounds that
represent different structural categories (sugars, amino acids, synthetic anions), which were
likely
to generate diverse generalization patterns. The mixture `Polycose' (Nissenbaum and Sclafani, 1987), which is also in this category, has been dealt with
separately (Rehnberg et al., 1996;Formaker et
al., 1998).
The chorda tympani nerve (CT) is one of three peripheral nerves that carry information about
taste stimuli to the brain. The other nerves are the greater superficial petrosal nerve (GSP) and
the glossopharyngeal nerve. In hamsters, the CT and GSP are more involved in detecting the
taste
of sucrose than the glossopharyngeal (Smith and Frank, 1993). The CT,
for which the most information is available, contains sucrose-best S neurons, which
carry critical information on the similarity of sucrose and other sweeteners (Smithet al., 1983), as well as NaCl-best (N) and HCl-best (H)
neurons, which respond to ionic stimuli (Frank et al., 1988).
Non-ionic sucrose, maltose, D-Phe and dulcin primarily activate S fibers (Hyman and Frank, 1980b; Tonosaki and Beidler, 1989;Rehnberg et al., 1996). Of the two sugars, sucrose
is a more effective stimulus than maltose for the hamster CT and GSP (Harada and
Smith, 1992). The maltose CT stimulus–response function overlaps the
glucose function but sucrose is effective at one-tenth the concentration (Rehnberget al., 1996). In contrast, the anionic sweeteners Na saccharin, Na nitrobenzene
sulfonate and Ca cyclamate activate S fibers as well as N and/or H
fibers. Na saccharin activates S fibers at preferred concentrations (Ogawaet al., 1969;Frank et al., 1988). S
-fiber responses to Na saccharin peak at 10–30 mM but fall off precipitously above 100
mM, whereas N-fiber responses increase monotonically with concentration (Ogawa et al., 1969). S fibers and N fibers are also
activated by sodium salts of preferred sulfonates (Rehnberg et al., 1990), whereas Ca cyclamate activates H fibers as well (Rehnberg et al., 1990).
Thus, we studied cross-generalization patterns in hamsters for seven preferred stimuli of diverse chemical structure that activate different patterns of activity across CT nerve fibers. Six of the seven stimuli generalized to sucrose. Maltose, which generalized to others of the preferred compounds, failed to generalize to sucrose or calcium cyclamate.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Subjects
Male golden hamsters (Mesocricetus auratus) were the subjects. Sixty-seven animals participated in the conditioned taste aversion study. At the time of the first conditioning their mean (± SD) weight was 111 ± 8 g. On 6–10 pretraining days, the animals were accustomed to drinking water from single 15 ml drinking tubes for two 1 h periods, one in the morning (test) and one in the evening (rehydration). During pretraining, daily water intake (ml) for the 1 h morning session was monitored. Individual intake values were used to develop groups of animals with similar water intake (4.05 ± 0.23 ml). Sixteen animals of weight 147 ± 12 g participated in the two-bottle intake studies. Ten of the hamsters were also used for electrophysiological recordings from the CT nerve.
Stimuli (Figure 1)
Preferred stimuli in three structural categories (sugars, amino acids and synthetic anions) were used which, based on current knowledge, were likely to have distinct perceptual properties. A summary of the evidence for this is as follows.
|
Sugars
We chose sucrose and maltose. In contrast to the furanose fructose and the pyranose glucose, which both readily cross-generalize with sucrose (Nowlis et al., 1980
; Frank and Nowlis, 1989), the two disaccharides maltose
and sucrose do not readily cross-generalize and likely have distinct qualitative components in
hamsters (Rehnberg et al., 1996) and rats (Spector
and Grill, 1988;Spector et al., 1997). Fructose,
glucose, sucrose and maltose are `monogeusic' to humans (Breslinet al., 1996) but species differences are common in taste, even among rodents (Beidler et al., 1955; Carpenter, 1956; Nowlis et
al., 1980; Tonosaki and Beidler, 1989). Hamster
preference functions for maltose and sucrose overlap (Rehnberg et al., 1996).
Amino acids
As representative of preferred zwitterionic amino acids, we chose D-phenylalanine (D-Phe), which is generalized to sucrose by hamsters. Hamsters generalize the L-isomer to quinine (Nowlis et al.,
1980) and other
aversive stimuli (Yamamoto et al.,1988). D-Phe is
sweet but L-Phe is bitter to humans (Shallenberger, 1993).
Glycine, D- and L-alanine and L-proline are generalized to
sucrose by hamsters (Nowlis et al., 1980; Yamamoto et al.,1988). As a counterpoint to D-Phe, we chose
nonionizable dulcin, an aryl urea which is a hydrophobic synthetic sweetener with structural
similarity to D-Phe (Shallenberger, 1993). We know of no
data on the generalization of dulcin in hamsters.
Synthetic anions
We chose three synthetic anions: the aromatic saccharin and m-nitrobenzene sulfonate
(NBS), as well as the aliphatic sulfamate cyclamate. Each of these anions is amphiphilic.
Saccharin and cyclamate have been used as non-nutritive sweetening agents for humans. NBS is
bitter to humans (Moncrieff, 1967) but hamsters generalize 10 mM
sodium NBS (Herness and Pfaffmann, 1986) as well as 10 mM sodium
mercaptoethane sulfonate (Frank et al., 1987) to sucrose, not to
quinine. Saccharin has a significant bitter component to humans at 1 mM (Shallenberger, 1993;DuBois et al., 1991), whereas
little bitterness is reported for cyclamate (DuBois et al., 1991).
We presented saccharin and NBS as sodium salts and cyclamate as a calcium salt; hamsters
strongly generalize sodium cyclamate to NaCl (Nowlis et al., 1980).
The taste of saccharin to hamsters is well studied. Sodium saccharin is strongly preferred
between 1 and 30 mM, but solutions stronger than 100 mM are rejected (Herness and
Pfaffmann, 1986; Rehnberg et al., 1990). Within the
range of preferred concentrations, Na saccharin cross-generalizes with sucrose and fructose and
hardly generalizes to NaCl (Frank and Nowlis, 1989; Nowlis et al., 1980). The tastes of m-nitrobenzene sulfonate and
cyclamate to hamsters are less well studied. Na NBS is moderately preferred between 3 and 30
mM (Herness and Pfaffmann, 1986; Rehnberg et al.,
1990), and a learned aversion to 10 mM Na NBS generalizes to sucrose and NaCl (Herness and Pfaffmann, 1986). Generalization of aversions to Ca
cyclamate have not been studied but aversions to 25 mM sodium cyclamate generalize more
strongly to NaCl than sucrose (Nowlis et al., 1980).
Stimulus concentrations
The stimulus solutions for the behavioral studies (and abbreviations used in text and figures) are
listed in Table 1. Each stimulus served as both a conditional (CS) and a
test stimulus (TS). Initially, we chose solutions that were preferred versus water in two-bottle
intake tests at a 70% level (mean ± SD for the seven compounds, 70.2 ±
2.7%) (Rehnberg et al., 1990, 1996).
We conditioned 24 hamsters with the 70%-preferred concentrations, but the
mean (± SE) aversion (relative to preconditioning mean water intake) expressed to the
CS was weak (22.92 ± 0.04% suppression). We increased the concentration of
each solution except dulcin by one-half log10 step. Solubility limits allowed only
doubling of the dulcin concentration. Conditioning to these higher concentrations (Table 1) produced, 1 week later, workable aversions (56.01 ±
0.04%) across the seven stimuli. The aversions expressed to stimuli listed in Table 1 were significantly stronger than those expressed to the
70%-preferred stimuli [t (23) = 7.91, P < 0.0001]
However,
the aversion expressed following the second weekly conditioning to the stronger stimuli (54.63
± 0.04%) did not differ from the previous week's value [t
(23) = 0.43]
|
The stimulus solutions listed in Table 1
were also used in two-bottle
intake tests. One group of six animals (each with a different random order) was tested
sequentially with all listed stimuli except NaNBS. In the context of other experiments (Rehnberg et al., 1990), a different group of 10 animals was
tested with NaNBS and subsequently with CaCyc2.
Stimuli applied to the anterior portion of a hamster's tongue during CT recording
included those listed in Table 1 and the lower levels noted above. Also,
NaCl and CaCl2 were tested to compare their effects with the anionic sweeteners
presented as sodium (NaNBS, NaSac) and calcium (CaCyc2) salts. Two
concentrations of each sweetener–chloride pair were tested; i.e. 3.2 mM and 1 mM NaSac
and NaCl, 32 mM and 10 mM NaNBS and NaCl, and 16 and 5 M CaCyc2 and CaCl2. Stimulus solutions were applied at flow rates of ~.5 ml/s from overhead
funnels into a glass flow chamber, which allows rapid fluid exchange (Harper, 1979). A stimulus solution was applied for 20–30 s, and followed by a
30–60 s deionized water rinse; a subsequent stimulus solution was applied 30–60 s
following the rinse.
Behavioral procedures
The procedure used for conditioning taste aversions is described in Frank and Nowlis (1989). Briefly, each of the seven groups of eight experimental animals
was conditioned against one stimulus (Table 1). A single control group of
11 animals was conditioned against water. The animals were conditioned against the same
stimulus (on Friday) at 1 week intervals to offset extinction of the learned response. Conditioning
involved pairing i.p. injection of apomorphine·Cl (30 mg/kg body wt) with ingestion of
a stimulus (CS) and was followed by a 3 day recovery period. On the conditioning day, animals
had access to the CS for 1 h; if intake was <1.0 ml, a small amount of the CS was forced into
the mouth before the apomorphine injection. Ingestion of a test stimulus (TS) was measured for 1
h on one of the four possible test mornings (Monday to Thursday) following the weekend
recovery period; on the day following the fourth test day (1 week after the first conditioning), the
animals were reconditioned and testing continued following recovery as in the first week. TS
were randomly presented over the test days for each animal to balance for effects of extinction (Kraemer and Spear, 1992; Rosas and Bouton, 1996),
which occurred with presentation of CS and/or TS that generalized to the CS. An
indication of the amount of extinction that occurred is the drop [t (43) =
5.34, P < 0.0001]in aversion from 59.06 ± 0.04% for the first
two test days post-conditioning to 38.46 ± 0.05% for the third and fourth test
days. Extinction was offset by weekly conditioning trials (see above).
The procedure used for establishing two-bottle preferences is described in Rehnberg et al., (1990). Briefly, a group of 6–10 animals was allowed to
sample from two 15 ml drinking tubes (one containing a stimulus and the other water) for two 24
h periods. After the first period the relative position of the stimulus and water tube was reversed.
Ingestion from each bottle was measured at the end of each 24 h period and summed for the two
periods. Mean stimulus and water intake were calculated for each stimulus.
Electrophysiological recordings
Recordings were taken from the right CT nerve following standard procedures (Frank, 1995). The animal was deeply anesthetized with sodium pentobarbital
(Nembutal, Abbott Labs, Chicago, IL, 100 mg/kg body wt i.p.) and maintained under deep
anesthesia with periodic supplemental injections (25–50 mg/kg). The nerve was exposed
using the mandibular approach and cut near its exit from the tympanic bulla. The cut nerve was
lifted with a Nichrome wire recording electrode above exposed muscles, into which an
indifferent electrode was inserted. The electrophysiological signals were amplified 104–105 times. The amplified signals approximated 1 V and were used to
drive a chart recorder. Examples of the integrated (rectified and averaged through a low-pass
filter with a time constant of 200 ms) and squared (Harper and Knight, 1987) recordings are shown in Figure 5. From these
`whole-nerve' recordings, which monitor the neural spiking of many nerve fibers
at once, the peak responses (minus baseline level) (Rehnberg et al., 1989) to test stimuli (Rs) and the standard (30 mM NaCl, RNaCl)
were measured. The standard stimulus was presented frequently during the recording session.
With this preparation, water or low concentrations of stimulus compounds following water rinses
(Rehnberg et al., 1996) do not elicit above-baseline responses.
At the end of the experiment, the animal was terminated with an intracardiac injection of
Nembutal.
|
Derived measures
The strength of a learned aversion was quantified as the percent suppression ([ 1 – (TSe/TSC)] x 100) based on the ratio of the 1 h intake (ml) of a TS by each experimental animal (TSe) and the mean intake of that TS by the control group (TSC). The mean percent suppression of each TS was calculated for each experimental group.
In addition, mean CS cross-generalization for each of the seven CS and reciprocal cross-generalization for each of the 21 stimulus pairs were calculated. A CS cross-generalization was calculated from data of an experimental group; it was the mean percent suppression for the six TS, excluding the value for suppression of CS drinking. Reciprocal cross-generalization was calculated for each stimulus pair from data of two experimental groups. The mean of the 16 values of percent suppression for individual animals was calculated for each stimulus pair. For example, the value for sucrose–dulcin included the eight values for TS dulcin from the experimental group with sucrose as CS and the eight values for TS sucrose from the experimental group with dulcin as CS. Reported values were in the form of percent suppression, calculated as described above.
Percent preference was calculated for each animal as [(ml stimulus)/(ml total fluid)] x 100, and a mean preference calculated for each stimulus.
The percent that a neural response (Rs) was of the standard response (RNaCl) was calculated for every stimulus for each preparation as: R% = (Rs/RNaCl) x 100, i.e. the percent of the standard response. Mean standardized neural responses were calculated for each test stimulus.
Data analysis
In general, significance of differences (
= 0.05) was addressed with analysis of
variance (ANOVA) followed by post-hoc Neuman–Keuls (NK) tests.
Conditioned taste aversion
One-way ANOVAs addressed water intake during the test period, intake of TS by the control
group during the test period, CS and water intake in the experimental groups, percent suppression
of CS and TS in each experimental group, and percent suppression of TS in experimental groups.
Two-way ANOVAs addressed intake of TS by each experimental group and the control group,
and the patterns of aversions for each TS pair across the CS.
Two-bottle preference
Significance of preferences and differences in preference levels were addressed with
repeated-measures two-way ANOVA comparing stimulus versus water intake across stimulus
solutions. Significance of differences in preference levels by different groups of animals were
addressed with t-tests for independent samples.
Chorda tympani recording
Significance of differences in standardized whole-nerve responses (R%) were
addressed with ANOVA.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Conditioned taste aversions
Mean (± SE) intake values during the 1 h test period for the control and experimental
groups are given in Table 2. Water intake by the control and experimental
groups did not differ significantly [top row of Table 2, F
(7,59) = 1.21] water intake across groups was 3.59 ± 0.14 ml. TS intake by
the control (H2O) group varied [first column in Table 2, F (7,70) = 3.82, P < 0.01] specifically, NaNBS intake was
significantly lower (P < 0.05) than intake of all test stimuli but water (P <
0.08) and cyclamate (P < 0.06). Conditioning against each CS affected intake across
the 7 test stimuli including the CS [F (1,17), P < 0.01] but TS
were not uniformly affected [F (6,102), P < 0.01] In all cases
CS
intake was strongly affected (P < 0.001), although, among CSs, NaNBS intake was
more strongly affected than sugar intake (P < 0.05).
|
Cross-generalization
Cross-generalization occurs for two stimuli when an aversion learned to one stimulus generalizes to the other stimulus and vice versa. Five stimuli—sucrose, D-Phe, saccharin, dulcin and NaNBS—cross-generalized. Learning an aversion to each stimulus affected intake of the other four stimuli (P < 0.02; Table 2
). Aversions learned to sucrose, D-Phe and saccharin cross-generalized
perfectly; that is, the expressed aversions were equivalent. However, learning an aversion to
dulcin affected intake of dulcin more than intake of saccharin and NaNBS (P < 0.05),
and learning of an aversion to NaNBS affected intake of NaNBS more than intake of the four
other stimuli (P < 0.01). Aversions to maltose did not cross-generalize as readily. Maltose did cross-generalize with saccharin, dulcin and NaNBS (P < 0.02); there was also an asymmetrical D-Phe to maltose (P < 0.05) generalization but no generalizations between maltose and sucrose or maltose and cyclamate. Cyclamate cross-generalized only with D-Phe (P < 0.01). However, there were also asymmetrical dulcin to cyclamate (P < 0.001), cyclamate to sucrose (P < 0.01) and cyclamate to saccharin (P < 0.01) generalizations. There were no generalizations between cyclamate and NaNBS or, as noted above, cyclamate and maltose.
TS intake patterns
The pattern of suppression of intake of each TS is presented in Figures 2
and 3, in which the pattern for the standard `sweet'
stimulus sucrose is repeated. The amount of suppression of sucrose intake (Figure 2A) by the sucrose-CS group did not significantly differ from the suppression of
sucrose intake in any other CS group, yet the suppression after maltose conditioning was not
significant (Table 2). The results were similar for D-Phe
intake suppression (Figure 2B) and saccharin intake suppression (Figure
2C), but all suppressions of saccharin intake were significant.
|
CS but suppression does not differ from
when TS = CS; coarse hatch bar: TS
|
Dulcin intake (Figure 2
D) was suppressed more in the dulcin-CS group
than several of the other groups, including the sucrose-, NaNBS- and maltose-CS groups (P < 0.05), which showed significant intake suppression (Table 2).
NaNBS intake (Figure 3B) was suppressed more in the NaNBS-CS group
than all other CS groups, five of which showed significant suppression (Table 2). Although dulcin and NaNBS intake suppression show this specificity for the CS,
their intake was significantly affected by conditioning in all but the cyclamate-CS group. CS
intake suppression differed across groups [F (1,6) = 3.18, P <
0.02] It was greatest for the NaNBS CS group, with a mean ± SE of 93.1
± 2.3% which was significantly (P < 0.05) greater than for the sucrose
(60.2 ± 7.9%), saccharin (63.9 ± 5.0%) or maltose (61.5 ±
10.4%) groups.
Maltose intake (Figure 3C) was suppressed more in the maltose-CS
group than either the sucrose-CS or cyclamate-CS groups. Cyclamate intake (Figure 3D) was suppressed more in the cyclamate-CS group than in all but the D-Phe-CS and dulcin-CS groups (P < 0.05). Furthermore,
maltose intake was not
affected in the cyclamate-CS group and cyclamate intake was not affected in the maltose-CS
group (Table 2). There was no generalization of aversions between these
two stimuli. Intake suppression (mean ± SE) was less for maltose (32.9 ±
6.6%) and cyclamate (31.7 ± 9.0%) than for other TS (which ranged from
48.0 ± 4.3% for saccharin to 55.2 ± 7.2% for NaNBS) across the
seven CS groups [F (6,294) = 8.79, P < 0.00001]
reflecting the specificity of the aversions to maltose and cyclamate.
Two-bottle preferences
Hamsters preferred each of the sweeteners over water (Figure 4). The
group of six hamsters showed significant preferences [solution versus water intake:
F (1,4) = 578.0, P < 0.0001] for five stimuli: sucrose,
phenylalanine, Na saccharin, dulcin and maltose. The preference values for these stimuli did not
differ significantly [F(4,20) = 1.01, ns] and had a mean ±
SE of 86.5 ± 3.1% This group of six hamsters
did not show a significant preference for cyclamate, likely due to the wide variation in their
cyclamate intake (mean = 16.1 ml, SD = 10.0 ml). However, cyclamate was
significantly preferred (60.7 ± 3.6%) over water by a different group of 10
hamsters, who also preferred NaNBS (72.5 ± 4.3%) [F(1,9)
= 23.02, P < 0.001] in these animals, cyclamate intake was more
consistent from animal to animal (mean = 21.2 ml, SD = 5.9 ml). NaNBS was
more highly preferred than CaCyc2 [F(1,9) = 9.02, P
< 0.02] but both NaNBS [t(14) = 2.31, P <
0.05] and CaCyc2 [t (14) = 4.90, P <
0.001] were preferred less than the other five TS.
|
Chorda tympani recordings
As seen in Figure 5, the CT nerve responded to each TS and 30 mM
NaCl (the standard stimulus). The 95% confidence intervals for mean values of R% did not include zero for any TS. The mean (±SE) CT response was 55.9
± 11.7% for the TS and 31.5 ± 7.8% for 70%-preferred,
lower concentrations (see Materials and methods). The 95% confidence intervals for the
means of these weaker solutions also did not include zero,
indicating they elicited reliable CT responses that were lower than responses to the TS [t (6) = 4.13, P < 0.01]. Figure 6
presents mean (± SE) responses for the TS, as well as sodium chloride and calcium
chloride solutions matched in molarity to the anionic sweetener solutions. Although equally
preferred (see above), 100 mM sucrose was a more effective CT stimulus than 320 mM maltose
[F (1,3) = 56.14, P < 0.01]. Also, although equally
preferred, 3.2 mM sodium saccharin was a more effective CT stimulus than 10 mM dulcin or 32
mM D-phenylalanine [F (2,6) = 5.35, P <
0.05]. Further, 32 mM sodium nitrobenzenesulfonate was a more effective CT stimulus
than 16 mM calcium cyclamate [ F(1,3) = 35.91, P <
0.01], 100 mM sucrose and 3.2 mM sodium saccharin [F(2,9) =
16.94, P < 0.001, NK P < 0.003]. At TS concentrations, the anionic
sweeteners NaNBS (+25%) and NaSac (+ 15.3%) were more effective
CT stimuli than NaCl (P < 0.01) but the lower concentrations of 1 mM Na saccharin
and 10 mM NaNBS were not [F(1,6) = 22.88, P < 0.01].
Calcium cyclamate (–12.2%) was not more effective than calcium chloride.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Review of CTA findings
Seven stimuli that were preferred and elicited responses of the CT nerve in hamsters were
studied. Learned aversions to D-Phe, Na saccharin, dulcin and Na m-nitrobenzene sulfonate cross-generalized with each other and with the sweet prototype sucrose.
Although maltose and Ca cyclamate also generalized with five of the
seven stimuli, maltose–sucrose, maltose–cyclamate and cyclamate–NBS
generalization failed reciprocally. Do these data suggest that hamsters prefer several distinct taste
qualities? Perhaps a `sweet1' and `sweet2 ',
as has been suggested for bitters (Herness and Pfaffmann, 1986) or
carbohydrates (Nissenbaum and Sclafani, 1987;Sclafaniet al., 1998)? Or, were features of maltose, calcium cyclamate and NaNBS
obscuring their sucrose-like perceptual components? Below we address two possible factors:
stimulus salience and cation taste, which may have affected the generalization patterns. But first
a context will be provided for the differences among generalization patterns observed.
Diversity of CTA generalization patterns
Figure 7 presents the results of two hierarchical cluster analyses for
comparison. Both are for seven stimuli, and used 
(
2) as
the metric and centroid amalgamation (Frank and Nowlis, 1989; Bieber and Smith, 1986). The upper dendrogram (solid lines) is for the
present data: TS generalization patterns (rows in Table 2). The lower
dendrogram (broken lines) is for a subset of the data presented in Table 1 of Frank
and Nowlis (1989): TS generalization patterns for sucrose and saccharin,
and five salts: 300 mM NH4Cl, 300 mM KCl, 100 mM MgSO4, 100 mM
NaCl and 410 mM NaSO4. The reciprocal cross-generalization of sucrose and
saccharin yielded similar small `distances' in
the two studies. One cluster accommodated the seven preferred stimuli of the present study,
whereas three clusters were required for seven stimuli in the other study that included
non-preferred salts. In the Frank and Nowlis (1989) study, neither
sucrose
nor saccharin
cross-generalized with any of the salts, and the sodium salts did not cross-generalize with the
non-sodium salts. Calcium cyclamate, the most distant member of the cluster of preferred
stimuli, joins the other six preferred stimuli at a distance of 8.7, a much smaller distance than the
point of amalgamation of the two salt clusters at 14.1. This implies that cyclamate is not an
`outlier' (Bieber and Smith, 1986) and that sodium salts
and non-sodium salts are less similar to each other than are any of the seven preferred stimuli.
Significant (P < 0.01) reciprocal cross-generalization resulted for 18 of the 21
stimulus pairs. Only Malt–CaCyc2 , with 10.2 ± 7.5%
suppression, Malt–Sucr, with 18.2 ± 6.5% and CaCyc2– NaNBS, with 20.9 ± 8.9% did not cross-generalize.
|
2 values for pairs of TS patterns.Stimulus concentration and salience
The measured strength of a learned aversion (CTA) is affected by the concentration of both the
conditional stimulus (CS), to which the aversion is trained, and the test timulus (TS), to which it
is expressed (Nowlis, 1974; Rescorla and Cunningham, 1978; Spector and Grill, 1988; Frank and Nowlis,
1989; Formaker and Hill, 1990). For example, increases in
concentration of stimuli by one-half log step resulted in an average increase in expressed
aversions of 33% suppression in our study (see Methods). In the context of CTA
experiments, we use stimulus salience to mean the combined effects of concentration on
establishing and eliciting an aversion. Salience may be best estimated by the strength of the
aversion (quantified by percent suppression) expressed to the CS itself, for which there is no
generalization decrements due to quality and intensity differences. So defined, the stimuli are
ordered for increasing salience as follows: sucrose (60%), maltose (61%), saccharin
(64%), D-Phe (71%), cyclamate (76%), dulcin (84%),
NaNBS (93%). NaNBS was by far the most salient stimulus and the most effective CT
stimulus.
Weaker, less salient TS would elicit weaker expression of learned aversions even if qualitatively
identical to stronger stimuli (Nowlis, 1974; Frank and
Nowlis,
1989). Perhaps part of the generalization decrements from NaNBS [F
(4,28) = 8.98, P < 0.001]or dulcin [F (4,28) =
5.99, P < 0.01]to sucrose, saccharin, maltose and D-Phe could be
attributed to differences in stimulus salience. On
the other hand, sucrose and maltose were least salient. Their combined stimulus salience made
sucrose–maltose the weakest CS–TS pair, which may have contributed to the
failure of sucrose– maltose reciprocal cross-generalization. Maltose–sucrose
cross-generalization occurs in rats trained with two conditioning trials and with the TS stronger
than the CS (Spector and Grill, 1988). Also, in hamsters, a CTA can be
established to the glucose–polymer mixture Polycose, at high but not low concentrations (Rehnberg et al., 1996), that generalizes to sucrose
(Formaker et al., 1998).
Cation tastes of anionic sweeteners
Cations and anions may contribute tastes of different perceptual qualities to a salt. For example,
a learned aversion to 25 mM sodium cyclamate is generalized by hamsters to both 100 mM
sucrose and 100 mM NaCl (Nowlis et al., 1980), a
generalization pattern similar to the one seen for an aversion learned to a mixture of sucrose and
NaCl (Frank, 1989). Neural recordings from single fibers of the hamster
CT nerve suggest that cation and anion activate distinct taste transduction cascades. For example,
20 mM sodium saccharin activates a set of CT fibers responsive to sucrose and a second set of
CT fibers responsive to NaCl (Ogawa et al., 1969). With regard
to our behavioral measurements, cations of salts may contribute their
characteristic tastes if levels are above those present in hamster saliva: 6.6 ± 0.5 mM Na+ and 1.5 ± 0.1 mM Ca2+ (means ± SE
reported byRehnberg et al., 1992). Thus, cation tastes for 32
mM NaNBS and 16 mM CaCyc2, stimuli preferred less (72 and 61%
respectively) than the other five stimuli (87%, will be addressed. We can gain an
appreciation of taste effects of sodium and calcium ions per se by looking at responses
to CaCl2 and NaCl. The chloride ion may contribute little to tastes of salts to
hamsters (Frank and Nowlis, 1989;Rehnberg et al.,
1993).
Calcium ion
Maltose–CaCyc2, one of three CS–TS pairs for which reciprocal
cross-generalization failed, showed the lowest average
percent suppression: 11% A rationale for this failure considers the pairing of the weak
maltose and an anionic sweetener presented as a calcium salt. An `off taste' may
be associated with the calcium cation.
Calcium cyclamate is preferred by hamsters between 5 and 50 mM; they are indifferent to 0.5
mM but find 160 mM CaCyc2 aversive [mean ± SE = 25.8
± 1.5% preference, t (5) = 10.8, P < 0.001 (Rehnberg et al., 1990; unpublished data)] The preference
for CaCyc2 is never high; the peak mean (± SE) preference we
observed was 67.8 ± 2.8% for 5 mM CaCyc2 (Rehnberg et al., 1990). For comparison, hamsters are indifferent to 5 mM CaCl2 and lower concentrations but show increasing levels of aversion for
16 mM CaCl2 [mean ± SE = 34 ± 5%
two-bottle preference, t (5) = 3.29, P < 0.03]and higher
concentrations [e.g. 160
mM: mean ± SE = 17.8 ± 3.2% t (5) = 15.4, P < 0.001] Thus, CaCyc2 may be considered a cation–anion
mixture, comparable to a mixture of sucrose and quinine·HCl, which is preferred less
than sucrose (Rehnberg et al., 1990). The anionic component of
the taste of 16 mM CaCyc2 is `like sucrose' (Nowlis et
al., 1980) and the cationic component is aversive and may be `like
quinine' to hamsters, as are other non-sodium salts (Frank and Nowlis, 1989) which are not `like NaCl' (Yamamoto et al.,
1988); CaCl2 is bitter to humans (Tordorf, 1996).
The aversive taste of Ca2+ may contribute to CaCyc2 showing the
weakest average CS cross-generalization (26% and lowest preference (61% of the
seven stimuli studied.
At 16 mM CaCl2 is as an effective CT stimulus as CaCyc2.
Interpretation of this result must take into consideration not only the substitution of chloride and
cyclamate, but also the effect of that substitution on the effect of the cation (Rehnberg et al., 1993). A number of lines of reasoning point toward a lesser effect
for the cyclamate salt than the chloride salt on ion-sensitive N and H fibers of
the CT. These include: (i) the saliva-adapted state of the behaving hamster's tongue
(Rehnberg et al., 1992); (ii) the negative effect of anion size
(cyclamate > chloride) on the paracellular pathway of salt-taste stimulation
(Ye et al., 1991, 1994); and (iii) the
blocking of this pathway by Ca2+ at decimolar levels (Kloubet al., 1998). The suggestion is that the sucrose-sensitive S fibers
contribute to the recorded CT response to 16 mM CaCyc2, as they do to 5 mM (Rehnberg et al., 1990).
Sodium ion
NaNBS–CaCyc2 was another of the three CS–TS pairs for which
reciprocal cross-generalization failed. A rationale for this failure
considers the distinct tastes of the cations, which may contribute different `off
tastes' to the two salts. CaCl2 and NaCl do not cross-generalize in hamsters (Yamamoto et al., 1988). NaCl cross-generalizes with other
sodium salts (Nowlis et al., 1980; Frank and Nowlis,
1989). It is likely that, like Na cyclamate (Nowlis et al., 1980), the learned aversion to NaNBS would have generalized to NaCl as well as to
sucrose, the generalization pattern seen for sucrose– NaCl mixtures (Frank,
1989). It is also likely that the NaNBS aversion would have generalized to odorous
compounds (Frank et al., 1987). During two-bottle intake
testing, we noted that 10–100 mM NaNBS had odors of increasing strength. With regard
to the failure of the NaNBS–CaCyc2 cross-generalization, hamsters have
great
difficulty in identifying a taste mixture component in another mixture (Frank, 1989). This difficulty may have cognitive as well as sensory components. Thus, as
sodium
and calcium add different `off-tastes' to NBS and cyclamate, the
cross-generalization of the two cation–anion mixtures would be expected to be weak even
if the anions were equally `like-sucrose'.
In 48 h intake tests, hamsters are indifferent to 1 mM NaNBS and lower concentrations but
equally prefer 3, 10, 30 and 100 mM NaNBS at a mean ± SE of 72.1
± 1.6% (Rehnberg et al., 1990;
unpublished data). The flat preference function reflects, perhaps, the negative effect of the
increasing concentration of Na+ on intake that offsets the positive effect of
increasing concentration of NBS on intake. Like CaCl2, NaCl is aversive to
hamsters in 48 h intake tests (Carpenter, 1956; Rehnberget al., 1990); the mean preference ± SE for 10 mM is 37.7
± 3.9% (t = 3.1, P < 0.02, unpublished data) and
for 100 mM, 28.2 ± 2.5% (Frank, 1989).
In our CTA study, 32 mM NaNBS was the most salient stimulus (93% suppression); 30
mM NaCl is less salient (35% suppression) to hamsters (unpublished data). The NBS
anion adds salience to this salt and, although both 32 mM sodium salts were quite good CT
stimuli, NaNBS was more effective than NaCl. The component that the NBS anion adds to the
CT response above that attributable to Na+ likely occurs in S fibers
of the CT and other taste nerves, particularly the greater superficial petrosal (GSP) (Smith and Frank, 1993). Anions larger than Cl– result
in reduced responses to sodium salts in CT H fibers and anion size hardly affects
responses of N fibers (Rehnberg et al., 1993). The CT
is critical in NaCl CTA expression (Barry et al., 1993) and NaCl
long-term preference (Barry et al., 1996) in hamsters. In the
present study, the mean (± SE) CS cross-generalization for NaNBS with the other six
sweeteners was a high 44 ± 5% compared with the maximum 49 ±
5%for dulcin. This is consistent with the possibility that theN- andH-fiber, Na+ portions of the CT response to NaNBS, which were recorded with
a water-adapted tongue, were exaggerated, given the saliva-adapted tongue, in the behaving
animal (Rehnberg et al., 1992). In addition, the
`off-taste' attributable to Na+ is likely not aversive. Intake of
100 mM NaCl does not differ from water intake in a 1 h test following 17.5 h fluid
deprivation (Frank and Nowlis, 1989;Formaker et al., 1998), the schedule used in the present CTA experiments. The
aversiveness of 100 mM NaCl seen in 48 h intake tests is likely based on post-ingestional effects
rather than taste per se (Pfaffmann, 1960; see also below).
Preference and taste
The taste of a solution may influence its intake versus water in two-bottle tests measured for 48
h. Many preferred substances are `sweet' in the sense they are `like
sucrose' (Nowlis et al., 1980) and, in a
variety of mammalian species, they also activate sucrose-best fibers of the CT nerve (Ninomiya et al., 1984a,b, 1987; Ninomiya and Funakoshi, 1993; Frank and
Nowlis, 1989; Rehnberg et al., 1990;Danilova et al., 1997;Hellekant et al., 1997). However, it as been long
documented that other variables, sensory and non-sensory,
influence long-term intake in rats (Pfaffmann, 1960). With regard to
caloric sources, possible post-ingestional factors include enzymatic hydrolysis in the mouth (Ramirez, 1991;Rehnberg et al., 1996),
absorption by the gut and build-up of osmotic pressure. The latter is thought to effect a decrease
in 48 h preference for sucrose at concentrations higher than 300 mM, a decrease that does not
occur for tests in which little sucrose intake occurs (Pfaffmann, 1960). In
hamsters, there is no turn down in the overlapping two-bottle preference curves for sucrose or
maltose up to 320 mM (Rehnberg et al., 1996).
The CT response to 320 mM maltose is much weaker than the CT response to 100 mM sucrose,
although they are equally preferred. This mismatch in preference and CT response level for
sucrose and maltose, is also evident for sucrose versus D-Phe or dulcin. With regard
to the two sugars, it is known that the hamster GSP nerve, which innervates taste buds on the
palate, is relatively more responsive to a series of 500 mM sugars than the CT (Harada and Smith, 1992). Both GSP and CT sucrose responses appear larger than
maltose responses, however. Besides undocumented offsetting contributions of other taste nerves
to the behavior, differences in preferences and CT response levels for sucrose and the salts
NaNBS and CaCyc2 could be attributed to CT effects of cations that are not preferred
in 48 h intake tests. Also, during two-bottle preference testing, we noted that NaNBS had an odor
and that CaCyc2 precipitated on the sipper tubes. These factors, as well as the
water-adapted (CT recordings) versus saliva-adapted (behavior) tongue (Rehnberg et al., 1992), may also have affected the match between preference and CT
response to salts. However, besides the possibility of offsetting effects of other taste nerves, what
may account for the relatively weak CT effects of maltose, D-Phe and dulcin?
As we have reasoned previously with regard to Polycose (Rehnberg et al.,
1996), 320 mM maltose, a glucose dimer, may be hydrolysed in the
mouth by salivary amylase. If hydrolysis were complete, the yield would be 640 mM glucose, a
concentration that would elicit a CT response that approaches the size of the response to 100 mM
sucrose. Concerning D-Phe, we noted the presence of precipitate on the sipper tubes
during the two-bottle tests with 32 mM D-Phe. The hamsters may have been
sampling a higher concentration of D-Phe by mixing the crystals with the 32 mM
solution. At 100 mM, sucrose and D-Phe are equally effective CT stimuli (Hyman and Frank, 1980a). Similarly, 10 mM dulcin is at the limit of
solubility and the animals may have been sampling crystals with the solution from the sipper
tubes during two-bottle intake testing. Such increases above the presented concentrations would
not occur during CT recordings.
Conclusion
The stimuli listed in Table 1 were all preferred and elicited reliable CT
responses. Patterns of generalization of conditioned taste aversions for the stimuli cluster at
distances consistent with stimuli that belong to one `sucrose-like' perceptual
quality.
|
Acknowledgments |
|---|
This research was supported by NIDCD grant R01 DC00058 and the University of Connecticut Health Center Research Advisory Committee.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Barry, M.A., Larson, D.C. and Frank, M.E. (1993) Loss and recovery of sodium-salt taste following bilateral chorda tympani crush. Physiol. Behav., 53,75 –80.[Medline]
Barry, M.A., Larson, D.C. and Frank, M.E. (1996) Effects of chorda tympani transection on long-term salt preference in hamsters.Physiol. Behav. , 60, 347–352.[Medline]
Beidler, L.M., Fishman, I.Y. and Hardiman, C.W. (1955) Species differences in taste responses. Am. J. Physiol., 181, 235–239.
Bieber, S.L. and Smith, D.V. (1986) Multivariate analysis of sensory data: a comparison of methods. Chem. Senses, 11, 19–47.
Breslin, P.A.S., Beauchamp, G.K. and Pugh, E.N., Jr (1996) Monogeusia for fructose, glucose, sucrose and maltose. Percept. Psychophys., 58, 327–341.[ISI][Medline]
Carpenter, J.A. (1956) Species differences in taste preferences. J. Comp. Physiol. Psychol., 49, 139–144.[ISI][Medline]
Danilova, V., Hellekant, G., Roberts, T., Tinti, J.-M. and Nofre, C. (1997) Behavioral and single chorda tympani taste fibers responses in common marmoset, Callithrix jacchus jacchus. Chem. Senses, 22, 667.
DuBois, G.E., Walters, D.E., Schiffman, S.S., Warwick, Z.S., Booth, B.J., Pecore, S.D., Gibes, K., Carr, B.T. and Brands, L.M. (1991) Concentration–response relationships of sweeteners: a systematic study. In Walters, D.E., Orthoefer, F.T. and DuBois, G.E. (eds), Sweeteners: Discovery, Molecular Design and Chemoreception. ACS, Washington, DC, pp. 261–276.
Formaker, B.K. and Hill, D.L. (1990) Alterations in salt taste perception in the developing rat. Behav. Neurosci., 104, 356–364.[ISI][Medline]
Formaker, B.K., Kearns, C.E. and Frank, M.E. (1998) The taste of Polycose in hamsters. Chem. Senses, 23, 675–682.[Abstract]
Frank, M.E. (1989) Processing of mixtures of stimuli with different tastes by primary mammalian taste neurons. In Laing, D.G., Cain, W.S., McBride, R.L. and Ache B.W. (eds), Perception of Complex Smells and Tastes. Academic Press, Sydney, pp. 127–147.
Frank, M.E. (1995) Taste nerve recordings in rodents. In Spielman, A.I and Brand, J.G. (eds), Experimental Cell Biology of Taste and Olfaction, Current Techniques and Protocols. CRC Press, Boca Raton, FL, pp. 263–270.
Frank, M.E. and Nowlis, G.H. (1989) Learned aversions and taste qualities in hamsters. Chem. Senses, 14, 379–394.
Frank, M.E., Patel, S.J. and Hettinger, T.P. (1987) Analysis of tastes in mixtures by hamsters. In Roper, S.D. and Atema, J. (eds), Olfaction and Taste IV. Ann. NY Acad. Sci., 510, 293–295.
Frank, M.E., Bieber, S.L. and Smith, D.V. (1988) The organization of taste sensibilities in hamster chorda tympani nerve fibers. J. Gen. Physiol., 91,861
–896.
Froloff, N., Faurion, A. and MacLeod, P. (1996) Multiple human taste receptor sites: a molecular modeling approach. Chem. Senses, 21, 425–445.
Harada, S. and Smith, D.V. (1992) Gustatory sensitivities of the hamster's soft palate. Chem. Senses, 17, 37–51.
Harper, H.W. (1979) Improved flow-chamber for taste stimulation. Physiol. Behav., 22, 1033–1035.[Medline]
Harper, H.W. and Knight, B.W. (1987) The correct method for signal processing for whole-nerve recordings. Chem. Senses, 12, 664.
Hellekant, G., Danilova, V., Ninomiya, Y. (1997) Primate sense of taste: behavioral and single chorda tympani and glossopharyngeal nerve fiber
recordings in the rhesus monkey, Macaca mulatta. J. Neurophysiol., 77, 978–993.
Herness, M.S. and Pfaffmann, C. (1986) Generalization of conditioned taste aversions in hamsters: evidence for multiple bitter receptor
sites. Chem. Senses, 11, 347–360.
Hyman, A.M. and Frank, M.E. (1980a) Effects of binary taste stimuli on the neural activity of the hamster chorda tympani. J.
Gen. Physiol., 76, 125–142.
Hyman, A.M. and Frank, M.E. (1980b) Sensitivities of single nerve fibers of the hamster chorda tympani to mixtures of taste stimuli. J. Gen. Physiol., 76,143
–173.
Jakinovich, W., Jr and Sugarman, D. (1988) Sugar taste reception in mammals. Chem. Senses, 13, 13–31.
Kier, L.B. (1972) A molecular theory of sweet taste. J. Pharmacol. Sci., 61, 1394–1397.[ISI][Medline]
Kloub, M.A., Heck, G.L. and DeSimone, J.A. (1998) Self-inhibition of Ca2+-evoked taste
responses: a novel tool for functional dissection of salt taste transduction mechanisms. J. Neurophysiol., 79, 911–921.
Kraemer, P.J. and Spear, N.E. (1992) The effect of nonreinforced stimulus exposure on the strength of a conditioned taste aversion as a function of retention interval: do latent inhibition and extinction involve a shared process? Anim. Learn. Behav., 20, 1–7.
Lindemann, B. (1996) Taste reception. Physiol. Rev., 76, 719–766.
Moncrieff, R.W. (1967) The Chemical Senses, 3rd edn. Leonard Hill, London.
Ninomiya, Y. and Funakoshi, M. (1993) Genetic and neurobehavioral approaches to the taste receptor mechanism in mammals. In Simon, S.A. and Roper, S.D. (eds), Mechanisms of Taste Transduction. CRC Press, Boca Raton, FL, pp. 253–272.
Ninomiya, Y., Higashi, T., Katsukawa, H., Mizukoshi, T. andFunakoshi, M. (1984a) Qualitative discrimination of gustatory stimuli in three different strains of mice. Brain Res., 322, 83–92.[ISI][Medline]
Ninomiya, Y., Mizukoshi, T., Higashi, T., Katsukawa, H. andFunakoshi, M. (1984b) Gustatory neural responses in three different strains of mice. Brain Res., 302, 305–314.[ISI][Medline]
Ninomiya, Y., Higashi, T., Mizukoshi, T. and Funakoshi, M. (1987) Genetics of the ability to perceive sweetness of D-phenylalanine in mice. Ann. NY Acad. Sci., 510,527 –529.
Nissenbaum, J.W. and Sclafani, A. (1987) Qualitative differences in polysaccharide and sugar tastes in the rat. A two-carbohydrate taste model. Neurosci. Biobehav. Rev., 11, 187–196.[ISI][Medline]
Nofre, C. and Tinti, J.M. (1996) Sweetness reception in man: the multi-point attachment theory. Food Chem., 56, 263–274.
Nofre, C., Tinti, J.M. and Glaser, D. (1996) Evolution of sweetness receptor in primates. II. Gustatory responses of non-human primates
to nine compounds known to be sweet to man. Chem. Senses, 21, 747–762.
Nowlis, G.H. (1974) Conditioned stimulus intensity and acquired alimentary aversions in the rat. J. Comp. Physiol. Psychol., 86, 1173–1184.
Nowlis, G.H., Frank, M.E. and Pfaffmann, C. (1980) Specificity of acquired aversions to taste qualities in hamsters and rats. J. Comp. Physiol. Psychol., 94,932 –942.[ISI][Medline]
Ogawa, H., Sato, M. and Yamashita, S. (1969) Gustatory impulse discharges in response to saccharin in rats and hamsters. J.
Physiol. (Lond.), 204, 311–329.
Pfaffmann, C. (1960) The pleasures of sensation. Psychol. Rev., 67, 253–268.[ISI][Medline]
Ramirez, I. (1991) Chemoreception for an insoluble
nonvolatile substance: starch taste? Am. J. Physiol., 260, R192–R199.
Rehnberg, B.G., Hettinger, T.P. and Frank, M.E. (1989) Asymmetrical neural cross-adaptation of hamster chorda tympani responses to
sodium and chloride salts. Chem. Senses, 14, 663–672.
Rehnberg, B.G., Hettinger, T.P. and Frank, M.E. (1990) The role of sucrose-sensitive neurons in ingestion of sweet stimuli by hamsters. Physiol. Behav., 48,459 –466.[Medline]
Rehnberg, B.G., Hettinger, T.P. and Frank, M.E. (1992) Salivary ions and neural taste responses in the hamster. Chem. Senses, 17,179
–190.
Rehnberg, B.G., MacKinnon, B.I., Hettinger, T.P. and Frank,
M.E., (1993) Anion modulation of taste responses in sodium-sensitive
neurons of the hamster chorda tympani nerve. J. Gen. Physiol., 101, 453–465.
Rehnberg, B.G., MacKinnon, B.I., Hettinger, T.P. and Frank, M.E. (1996) Analysis of polysaccharide taste in hamsters: behavioral and neural studies. Physiol. Behav., 59, 505–516.[Medline]
Rescorla, R.A. and Cunningham, C.L. (1978) Within-compound flavor associations. J. Exp. Psychol., 4, 267–275.
Rosas, J.M. and Bouton, M.E. (1996) Spontaneous recovery after extinction of a conditioned taste aversion. Anim. Learn. Behav., 24, 341–348.
Schiffman, S.S. and Erickson, R.P. (1993) Psychophysics: insights into transduction mechanisms and neural coding. In Simon, S.A. and Roper, S.D. (eds), Mechanisms of Taste Transduction. CRC Press, Boca Raton, FL, pp. 253–272.